United States of America v. State of California et al

Filing 104

AMICI CURIAE BRIEF by National Health Law Program (and associated parties). (Attachments: # 1 Appendix Studies cited in brief)(Coursolle, Abigail) Modified on 5/18/2018 (Benson, A.).

1
2
3
4
5
6
7
8
9
10

ABIGAIL K. COURSOLLE (Bar # 266646)
[e-mail]
Counsel of Record
MARTHA JANE PERKINS (Bar # 104784)
[e-mail]
SARAH GRUSIN
[e-mail]
JOSEPH MCLEAN
[e-mail]
NATIONAL HEALTH LAW PROGRAM
3701 Wilshire Blvd, Suite 750
Los Angeles, CA 90010
Telephone: [phone]
Facsimile: [phone]
Attorneys for Amici Curiae
National Health Law
Program et al.
UNITED STATES DISTRICT COURT
EASTERN DISTRICT OF CALIFORNIA,

11
12
13
14

Case No: 2:18-cv-00490-JAM-KJN
THE UNITED STATES OF AMERICA,
Plaintiff,

15
16
17
18
19
20
21
22
23
24
25
26
27
28

V.
THE STATE OF CALIFORNIA;
EDMUND GERALD BROWN JR.,
Governor of California, in his Official
Capacity; and XAVIER BECERRA,
Attorney General of California, in his
Official Capacity,
Defendants.

ADDENDUM TO BRIEF OF AMICI
CURIAE NATIONAL HEALTH
LAW PROGRAM ET AL. IN
SUPPORT OF THE DEFENDANT’S
OPPOSITION TO PLAINTIFF’S
MOTION FOR A PRELIMINARY
INJUNCTION

1
Exhibit List

2
3

Exhibit A:

Kimberly Gittings & Kelly L. Matson, Establishing Herd Immunity Against Ebola
Through Vaccination, 34 Vaccine 2664 (2016).

Exhibit B:

Jared M. Baeten, Amplifying the Population Health Benefits of PrEP for HIV Prevention, 217
J. of Infectious Diseases 1509 (2018).

Exhibit C:

A. Elizabeth Iten et al., Undocumented Immigration Status and Diabetes Care Among
Mexican Immigrants in Two Immigration “Sanctuary” Areas, 16 J. Immigr. Minority
Health 229 (2014).

Exhibit D:

Helen B. Marrow, The Power Of Local Autonomy: Expanding Health Care To Unauthorized
Immigrants In San Francisco, 35 Ethnic & Racial Studies 72 (2012).

Exhibit E:

Nicole L Novak et al., Change in Birth Outcomes Among Infants Born to Latina Mothers
After a Major Immigration Raid, 17 Int’l J. Epidemiology 1 (2017)

Exhibit F:

Ronald Glaser & Janice K. Kiecolt-Glaser, Stress-Induced Immune Dysfunction:
Implications for Health, 5 IMMUNOLOGY 243 (2005)

Exhibit G:

Lisseth Rojas-Flores et al., Trauma and Psychological Distress in Latino Citizen Children
Following Parental Detention and Deportation, 9 Psych. Trauma 352 (2017).

Exhibit H:

Kenneth E. Miller & Andrew Rasmussen, War Exposure, Daily Stressors, and Mental
Health in Conflict and Post-Conflict Settings, 70 Soc. Sci. Med. 7 (2010).

Exhibit I:

Allen S. Keller et al., Mental Health of Detained Asylum Seekers, 362 Lancet 1721 (2003).

4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

ADDENDUM TO MPAS IN SUPPORT OF AMICUS CURIAE
2:18-cv-00490-JAM-KJN

1
2
3
4
5
6
7

Exhibit A

8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

ADDENDUM TO MPAS IN SUPPORT OF AMICUS CURIAE
2:18-cv-00490-JAM-KJN

Vaccine 34 (2016) 2644–2647

Contents lists available at ScienceDirect

Vaccine
journal homepage: www.elsevier.com/locate/vaccine

Establishing herd immunity against Ebola through vaccination
Kimberly Gittings, Kelly L. Matson ∗
University of Rhode Island, College of Pharmacy, 7 Greenhouse Road, Kingston, RI 02881, United States

a r t i c l e

i n f o

Article history:
Received 29 October 2015
Received in revised form 15 April 2016
Accepted 18 April 2016
Available online 25 April 2016
Keywords:
Ebola
Vaccine
Herd immunity
Ring vaccination

a b s t r a c t
Objectives: In response to recent concern regarding Ebola outbreaks, this study aims to (1) determine the
relationship between vaccination coverage and herd immunity, (2) determine the vaccination coverage
necessary to establish herd immunity for previous Ebola viruses, and (3) recommend vaccination coverage
thresholds for future Ebola viruses.
Methods: Herd immunity thresholds needed to block transmission of Ebola virus were determined using
vaccine efficacy and number of secondary cases per infected case during an entire infectious period.
Results: In past Ebola outbreaks 42.2–63.0% of the population would need to be vaccinated in order to
prevent transmission and outbreaks. Assuming 80% vaccine efficacy as reported by phase I clinical trials,
52.7–78.7% of the population would require vaccination coverage in order to establish herd immunity. In
recent ring vaccination trials which considered the vaccine to be 100% effective after 10 days, 42.2–63.0%
of the population would need to be vaccinated.
Conclusion: For future Ebola outbreaks, the spread of the virus can be prevented by vaccinating certain
percentages of the population depending on vaccine efficacy and number of secondary cases per infected
case.
© 2016 Elsevier Ltd. All rights reserved.

1. Introduction
In 2014 the largest Ebola epidemic in history spread through
West Africa with additional cases reaching the United States [1].
The first case of Ebola was recognized in 1976 in the Democratic
Republic of Congo as a rare and severe illness with fatal potential.
The virus is transmitted from wild animals to people and spreads
through the population from human-to-human transmission. Due
to the dangerous nature of the virus, it is important to prevent its
transmission through vaccination. Vaccination can reduce the risk
of Ebola virus contraction and its related complications, physician
visits, hospitalizations and death. By vaccinating a certain proportion of the population against the virus, transmission of Ebola in
the community can be blocked through the establishment of herd
immunity.
Vaccines can affect more than just the individual who is vaccinated; vaccines can also protect people who have not been

Abbreviations: Vc , critical vaccination coverage; Ic , herd immunity threshold; R0 ,
basic reproductive number; E, vaccine efficacy.
∗ Corresponding author at: Department of Pharmacy Practice, University of Rhode
Island, College of Pharmacy, 7 Greenhouse Road, Room 244M, Kingston, RI 02881,
United States. Tel.: [phone].
E-mail addresses: [e-mail] (K. Gittings), [e-mail]
(K.L. Matson).
http://dx.doi.org/10.1016/j.vaccine.2016.04.047
0264-410X/© 2016 Elsevier Ltd. All rights reserved.

immunized. The concept of “herd immunity” refers to an indirect
protection of an entire community from disease by immunizing
a critical proportion of the populace. Herd immunity breaks the
chain of an infection’s transmission so that outbreak does not
occur [2]. For example, transmission of measles can be blocked
by vaccinating 92–95% of a given community [3]. The remaining
5–8% of the community who are unvaccinated and susceptible to
measles receive “conferred immunity” from the vaccinated individuals. Given the proportion of vaccinated individuals, in terms
of vaccination coverage, above a pre-determined herd immunity
threshold, transmission of measles is blocked within the community.
The threshold for herd immunity needed to block transmission
of Ebola virus in the population is currently unknown. Herd immunity is established when the prevalence of protected persons (I) is
higher than the herd immunity threshold (I > Ic ) [3,4]. When this
occurs, Ebola virus transmission is blocked within the given population. However when prevalence is lower than the threshold, the
number of infections is able to grow exponentially, thus spreading the virus within the population. Recent early phase trials of
Ebola vaccinations report the efficacy of the vaccines, which can be
used to determine the percentage of the population that requires
vaccination in order to reduce community outbreaks and prevent
transmission [5,6].
The objectives of this study are to determine the relationship between Ebola vaccination coverage and herd immunity,

K. Gittings, K.L. Matson / Vaccine 34 (2016) 2644–2647

2645

Table 1
Average number of secondary cases per infected case (R0 ) and prevalence of protected persons necessary to establish herd immunity (Ic ) for studied Ebola virus
outbreaks.
Ebola virus

R0 , CI

Ic (%), CI

1995 Democratic Republic of
the Congo epidemic [5]
2000 Uganda epidemic [5]
2014 Liberia epidemic [6]

2.7 (1.9–2.8)

63.0% (47.37–64.3%)

2.7 (2.5–4.1)
1.73 (1.66–1.83)

63.0% (60.0–75.6%)
42.2% (39.8–45.4%)

Where Ic = [1 − (1/R0 )].
Ic = herd immunity threshold and R0 = basic reproductive number.

determine the vaccination coverage necessary to establish herd
immunity for previous Ebola viruses, and provide suggestions for
vaccine coverage needed for future Ebola viruses.
2. Methods
This study mathematically determined the herd immunity
threshold required to prevent transmission of Ebola. It was determined by accounting for the number of secondary cases per infected
case (R0 ) during the entire infectious period in a completely susceptible population, or basic reproductive number, in past outbreaks
and the vaccine effectiveness. When R0 > 1, outbreaks and resulting
epidemics occur.
When vaccinations are administered within a specified population or community, the vaccine protects only a proportion (E) of
the vaccinated individuals. The proportion of protected individuals
who were vaccinated represents the effectiveness of the vaccine
against infection transmission.
Using the mentioned variables, the critical proportion of
protected individuals needed to establish herd immunity in a
completely susceptible community can be determined from the
equation Ic = 1 − (1/R0 ) [2–4]. The critical vaccination coverage (Vc )
needed to establish herd immunity can next be determined by
dividing the herd immunity threshold (Ic ) by the level of vaccine
effectiveness (E): Vc = Ic /E = [1 − (1/R0 )]/E [2,4].
Citing R0 values from past Ebola epidemics (Table 1), it is possible to mathematically derive the herd immunity threshold, and the
number of protected persons required to establish herd immunity
in a completely susceptible population.
After determining herd immunity thresholds for previous epidemics, data was pulled from phase I and III clinical trials in order
to determine the critical vaccination coverage needed to establish
herd immunity in past outbreaks given vaccine efficacy. Vaccine
efficacy in early phase I clinical trials was measured by percentage of subjects with positive enzyme-linked immunosorbent assay
results at week 12 after vaccination. Antibodies directed against
specific antigens were measured throughout the trial and an end
point titer with a background-corrected optical density reading of
≥30 was considered a positive result [7]. A more recent phase 3
trial of Ebola ring vaccination determined efficacy of ring vaccination based on zero cases of Ebola virus disease at 10 days or more
post-randomization and vaccination [8].

Fig. 1. Critical vaccination coverage (%) needed to provide herd immunity against
varying Ebola viruses and variable vaccine efficacy. R0 = basic reproductive number.

3. Results
In past Ebola virus epidemics, the prevalence of protected persons needed to establish herd immunity ranged from 42.2% in the
most recent epidemic to 63.0% in earlier epidemics (Table 1).
The required vaccination coverage to establish herd immunity
for past Ebola epidemics varied depending on vaccine efficacy. The
required vaccination coverage to establish herd immunity for these
past Ebola epidemics ranges from 52.7% to 78.7% assuming the vaccine is 80% effective as reported by a phase I clinical trial [7]. A
2015 phase-3 ring vaccination cluster-randomized trial reports the
efficacy of the vaccine in different scenarios. In individuals who randomly received the ring vaccination, the vaccine was considered to
be 100% efficacious after 10 days [8] which requires 42.2–63.0% of
the population to be vaccinated in order to provide herd immunity. The 2015 study reports an estimated 75.1% and 76.3% overall
vaccine efficacy in all eligible participants, which equates to a
critical vaccination coverage of 56.2–83.9% and 55.2–82.6% respectively.
To account for real-world human error and varying degrees
of efficacy, Table 2 reports the vaccination coverage that would
have been required to establish herd immunity against past
epidemic Ebola viruses for different levels of vaccine effectiveness.
The vaccination coverage required to establish herd immunity
against future Ebola viruses for varying levels of vaccine effectiveness and differing R0 values is demonstrated in Fig. 1. For example,
when the number of secondary cases per infected cases, R0 , is equal
to 1.1 and the vaccine is approximately 90% effective, only about
10% of the given population will have to be vaccinated in order to
provide herd immunity against the virus.

Table 2
Vaccine coverage (Vc ) required to establish herd immunity against past Ebola viruses for varying levels of vaccine effectiveness.
Vaccine effectiveness (E)

1995 Democratic Republic
of the Congo epidemic

2000 Uganda epidemic

2014 Liberia epidemic

40%
60%
80%
90%
100%

100%
100%
78.7%
70.0%
63.0%

100%
100%
78.7%
70.0%
63.0%

100%
70.3%
52.7%
46.9%
42.2%

Where Vc = Ic /E.
Vc = vaccine coverage and Ic = herd immunity threshold.

2646

K. Gittings, K.L. Matson / Vaccine 34 (2016) 2644–2647

4. Discussion
This study explored the relationship between Ebola vaccination coverage and herd immunity while assessing the vaccination
coverage necessary to establish herd immunity for previous Ebola
epidemics. The results suggest that higher levels of vaccination coverage would have been needed in order to provide herd immunity
against previous Ebola epidemics. Given that the recent 2014 Ebola
epidemic had a lower R0 value, it may suggest that less vaccine coverage than previous epidemics would be needed in order to prevent
transmission.
Additionally, results of this study suggest that future Ebola virus
transmission can be blocked by vaccinating susceptible populations. Fig. 1 demonstrates vaccination coverage needed in order
to prevent future Ebola outbreaks in a 100% susceptible population given varying R0 values and vaccine effectiveness. As seen in
Fig. 1, it is easiest to prevent an outbreak when a virus has a low R0
value and high vaccine efficacy. This study suggests that for Ebola
viruses with R0 ≥ 3 and vaccine effectiveness of 70%, nearly all of
the population would need to be vaccinated in order to establish
herd immunity. With a lower R0 value and a vaccine efficacy of 70%,
a smaller percentage of the population would need to receive the
vaccine in order to prevent outbreak.
It is important to note that there are several limitations to
this study. The study utilizes a simple threshold theorem which
makes several assumptions: (1) random vaccination within the
population, (2) homogenous mixing of persons within the population, (3) homogeneous distribution of vaccine-induced protected
and infected persons within the population, and (4) fully susceptible population [2]. It can be speculated that because this
model assumes a fully susceptible population, vaccination coverage
needed to establish herd immunity in a population with individuals already protected due to natural infections would be less than
the proposed thresholds within this study. Additionally, this study
assumes that vaccine efficacy is equal to the percentage of patients
from previous trials with positive enzyme-linked immunosorbent
assays when this value may not truly correlate to the vaccine efficacy and immunity.
This study relies heavily on the accuracy of the basic reproductive values (R0 ) reported for past Ebola outbreaks that are included
in Table 1 and used to calculate the results in Table 2. As seen
in Table 1 of this study, the R0 value of the 1995 DRC and 2000
Uganda epidemics are 2.7 whereas the most recent 2014 Liberia
epidemic’s value was significantly lower at 1.73. Although the
strains for the outbreaks were the same, the varying R0 values can
vary due to several factors because it is a property that combines
the process of contagion within a population and the patterns of
contact within the population. Some variables that affect R0 may
include the number of susceptible people in the population that
the affected patients are in contact with, containment and control measures for the virus, and stages of outbreak [9]. In 2014,
the World Health Organization (WHO) redesigned frame work and
created intensive public health containing measures to be implemented from both local and international levels [10]. The R0 value
for the 2014 epidemic was calculated from data collected during
July and September 2014, after said measures were initiated. It is
possible that containment measures were improved in the most
recent outbreak thus resulting in a decreased R0 value.
Currently, there are known contraindications to Ebola vaccination for specific age ranges. Of note, the phase 1 trial by Sarwar,
which this study utilized for vaccine efficacy data, was limited to
adults aged 18–60 years [7]. Applying the data from our study we
can speculate that if 52.7–78.7% of a mixed age population of individuals was vaccinated this would provide community protection
to the young and the older adults who were ineligible to receive the
phase I trial vaccine in that population. Due to the restrictions in

age, a larger percentage of the adult population needs to be vaccinated in order to protect those who cannot be vaccinated. If children
and older adults were also eligible to be vaccinated, it would be easier to reach the herd immunity threshold by being able to vaccinate
more individuals.
The most recent phase 3 trial of Ebola vaccination in Guinea
involved ring vaccination in a cluster-randomized style. Ring vaccination involves administering vaccination only to a cluster of high
risk individuals who are in close contact with a confirmed isolated
infected person [8,11]. Although different from traditional methods
of vaccination this method demonstrates notable efficacy. Eligible
individuals in the clusters were given either immediate or delayed
Ebola vaccination. Of the immediately vaccinated individuals the
vaccine was 100% effective as determined by no symptoms of Ebola
virus disease 10 days after vaccination. Of all eligible individuals
who received immediate or delayed vaccination (21 days after randomization) the vaccine was 75.1% effective and 76.3% effective
depending on cluster [8]. Because this was a cluster trial of individuals in close contact with isolated infected person(s) the R0 value
of the virus is likely higher than in populations where individuals are not in close proximity. Therefore, it is possible that critical
vaccination coverage values may be lower in other communities.
Importantly, the basic reproductive value of the outbreak in Guinea
was not reported so it is difficult to conclude exactly how the critical
vaccination coverage would be affected.
In future outbreaks with a 100% efficacious vaccine, as reported
by the Guinea ring vaccination trial, 42–63% of the population
would have to be vaccinated to prevent transmission. There are
several stressors that could affect the ability to vaccinate individuals against Ebola in the future. When outbreak occurs, access
to adequate supply of vaccine to vaccinate sufficient number of
individuals is not always possible within desired time range to
control transmission. Thus, it is important that future Ebola outbreak protocols include intensive measures for containment to
prevent transmission. If supply is a potential issue it would be wise
to distribute vaccine supply in levels of priority. Similar to vaccination protocols within hospitals an important target population
to vaccinate would be healthcare workers [12]. The Guinea trial
demonstrates the value of ring vaccination in Ebola and the importance of vaccinating individuals who have been or will be in close
contact with the virus. This can include contacts of sick individuals
such as family and friends, caregivers, and contacts of contacts.
5. Conclusions
The novel data within this study can have future impacts in
preventing outbreaks and transmission of the rare but dangerous
Ebola virus. As clinical trials are still underway for development of
Ebola vaccines, the information from this study can be utilized in
the future with novel vaccine efficacy data to determine vaccine
coverage needed to prevent outbreaks.
Conflict of interest: None declared.
References
[1] Ebola (Ebola Virus Disease). Centers for Disease Control and Prevention; 2015.
http://www.cdc.gov/vhf/ebola/ [updated August 4, 2015; accessed August 5].
[2] Fine P, Eames K, Heymann DL. “Herd immunity”: a rough guide. Clin Infect Dis
2011;52(7):911–6.
[3] Anderson RM. The concept of herd immunity and the design of community
based immunization programs. Vaccine 1992;10:928–35.
[4] Plans-Rubio P. The vaccination coverage required to establish herd immunity
against influenza viruses. Prev Med 2012;55:72–7.
[5] Legrand J, Grais RF, Boelle PY, Valleron AJ, Flahault A. Understanding the dynamics of Ebola epidemics. Epidemiol Infect 2007;135(4):610–21.
[6] Yamin D, Gertler S, Ndeffo-Mbah ML, Galvani AP. Effect of Ebola progression
on transmission and control in Liberia. Ann Intern Med 2015;162(1):11–7.

K. Gittings, K.L. Matson / Vaccine 34 (2016) 2644–2647
[7] Sarwar UN. Safety and immunogenicity of DNA vaccines encoding Ebolavirus
and Marburgvirus wild-type glycoproteins in a phase I clinical trial. J Infect Dis
2015;211(4):549–57.
[8] Henao-Restrepo AM. Efficacy and effectiveness of an rVSV-vectored
vaccine expressing Ebola surface glycoprotein: interim results from
the Guinea ring vaccination cluster-randomised trial. Lancet 2015,
http://dx.doi.org/10.1016/S0140-6736(15)61117-5.
[9] Holme P, Masuda N. The basic reproductive number as a predictor for epidemic outbreaks in temporal networks. PLoS ONE 2015,
http://dx.doi.org/10.1371/journal.pone.0120567.

2647

[10] Barriers to rapid containment of the Ebola outbreak: Ebola situation assessment. World Health Organization; 2014 http://www.who.int/csr/disease/
ebola/overview-august-2014/en/.
¸
[11] Ebola ca Suffit Ring Vaccination Trial Consortium. The ring vaccination trial:
a novel cluster randomized controlled trial design to evaluate vaccine efficacy and effectiveness during outbreaks, with special reference to Ebola. BMJ
2015;351:h3740.
[12] Centers for Disease Control and Prevention. In: Hamborsky J, Kroger A, Wolfe
S, editors. Epidemiology and prevention of vaccine-preventable diseases. 13th
ed. Washington, DC: Public Health Foundation; 2015.

1
2
3
4
5
6
7

Exhibit B

8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

ADDENDUM TO MPAS IN SUPPORT OF AMICUS CURIAE
2:18-cv-00490-JAM-KJN

The Journal of Infectious Diseases
E D I T O R I A L C O M M E N TA R Y

Amplifying the Population Health Benefits of PrEP for
HIV Prevention
Jared M. Baeten
Departments of Global Health, Medicine, and Epidemiology, University of Washington, Seattle

(See the Major Article by Volz et al, on pages 1522–9.)
Pre-exposure prophylaxis (PrEP) prevents HIV infection. Robust evidence—
first from pivotal clinical trials [1, 2],
then rigorous studies testing best-practice models of implementation [3–6], and
most recently real-world case studies of
clinical delivery [7]—demonstrate that
PrEP is highly effective, indeed incredibly so, at providing individual protection against HIV acquisition for men
and women from diverse populations
worldwide. Regulatory review and normative guidance from the US Food and
Drug Administration in 2012 and from
the US Centers for Disease Control and
Prevention (CDC) in 2014 have been
followed by World Health Organization
recommendations in 2015, and thereafter by endorsements by more than two
dozen countries (www.prepwatch.org).
PrEP is thus a central part of global, evidence-based, gold-standard HIV prevention for at-risk individuals.
Nevertheless, for PrEP to have an
important impact on the HIV epidemic,
it needs to have not just benefits for
individuals but at the population level,
synergizing with other prevention and
treatment interventions to achieve substantive reductions in new infections and
resultant morbidity and mortality from
Received 19 January 2018; editorial decision 19 January 2018;
accepted 20 January 2018; published online February 26, 2018
Correspondence: J.  M. Baeten, MD, PhD,  Department of
Global Health, University of Washington, Box 359927, 325
Ninth Ave, Seattle, WA 98104 ([e-mail]).
The Journal of Infectious Diseases®  2018;217:1509–11
© The Author(s) 2018. Published by Oxford University Press for
the Infectious Diseases Society of America. All rights reserved.
For permissions, e-mail: [e-mail].
DOI: 10.1093/infdis/jiy045

HIV. By 2020, the US National HIV/
AIDS Strategy calls for a 25% reduction
in new diagnoses [8] and UNAIDS has
set a global goal of <500 
000 new infections, a 75% reduction compared to 2010
[9]. These are ambitious targets, for any
setting, and achieving them will require
the best strategy and science to guide
interventions and measure impact.
In this issue of the Journal of Infectious
Diseases [in press], Volz and colleagues
present an innovative set of analyses,
using molecular sequence data from HIV
infections among men who have sex with
men (MSM) in the United Kingdom,
combined with mathematical modeling, which together provide one potential strategy to maximize population
health benefits of PrEP. Specifically, the
authors analyzed publicly collected HIV
sequence data from nearly 7000 MSM
(linked to basic demographic data), used
phylogenetic and phylodynamic methods to characterize transmission-associated subgroups with greater risk for HIV
acquisition and onward virus spread,
and then estimated the population-level
benefits that would occur from different
scenarios under which a limited amount
of PrEP (enough for 15 
000 individuals)
would be available. Their principal finding (ie, that young MSM, aged < 25 years,
had greater HIV risk) seems on first pass
to be not particularly surprising, as adolescent and young adult men and women
in countries around the world have been
repeatedly called out as a high-risk population. However, the Volz and colleagues’
results go further than simply documenting younger age as a risk factor for HIV,

demonstrating that young MSM are also
more likely to transmit to other young
MSM if they do acquire the virus (75%
of infections in this group attributed to
other young MSM), multiplying the HIV
risk within this age group. As a result, prioritized introduction of PrEP to younger
MSM would have multiplied benefits,
blocking both first-generation infections
as well as large numbers of onward transmissions in this group. Thus, high levels
of transmission, combined with high
connectivity among similarly aged men,
in the words of the authors “amplifies
incidence … and PrEP effectiveness” in
young MSM.
Amplified effectiveness is what is
needed at this time for PrEP and for the
total toolbox of effective HIV prevention
strategies. The findings of Volz and colleagues remind us that HIV prevention is
about a best prevention plan for an individual, but must also be about prevention
for his or her current and future partners
and, ultimately, about the population
more generally. HIV, like other infectious
diseases, can benefit from interventions
that limit transmission cascades, as seen
most dramatically in the last few years
in the public health response to the outbreak in Scott County, Indiana linked to
injection drug use [10]. As done by Volz
and colleagues and by others [11, 12],
phylodynamic analyses of viral sequences
from that outbreak have been used to
characterize opportunities for strategic
prevention efforts [13]. This innovative use of molecular epidemiology and
mathematical modeling offers exciting
new opportunities to use complex science

EDITORIAL COMMENTARY • JID 2018:217 (15 May) • 1509

Downloaded from https://academic.oup.com/jid/article-abstract/217/10/1509/4915988
by University of California, San Diego Library user
on 08 May 2018

to guide public health decision making in
real time. The approach is arguably a continued evolution of the “know your epidemic, know your response” rallying cry
from UNAIDS a decade ago, backed up
by cutting-edge data analytics.
For PrEP, population health benefits
can only occur with enough coverage of
the population accessing it to result in
those key infections prevented to avert
subsequent generations of transmissions. We are not there yet. In the United
States, PrEP uptake has skyrocketed,
with >120 
000 individuals estimated to
have initiated PrEP since 2012 [14]; still,
CDC estimates that 1.2 million persons,
including MSM, people who inject drugs,
and heterosexual adults, have indications
for considering PrEP use [15]. Some
geographies are beginning to see PrEP
use at levels that should result in population benefits—for example, Seattle
and King County, Washington recently
reported that one-third of high-risk
MSM there are estimated to be on PrEP
currently and there are now more people
on PrEP in that locality than are taking
antiretroviral treatment (and that is in the
context of that locality also achieving the
UNAIDS goal of 90% of individuals diagnosed, linked to care, and on suppressive
antiretroviral therapy) [16]. New initiatives related to access to PrEP (as well
as treatment and other prevention interventions) in Florida and elsewhere in the
southern United States offer encouragement that population benefits may be
on the horizon for a part of this country
that faces substantial disparity in new
HIV infections [17]. The United States
accounts for most PrEP use in the world
at this time, although the vast majority of
new infections occur in low and middle
income countries; UNAIDS has called
for 3 million persons to be on PrEP by
2020 [9]. Ambitious programs like the
DREAMS initiative (https://www.pepfar.
gov/partnerships/ppp/dreams/),
supported by the US President’s Emergency
Plan for AIDS Relief (PEPFAR), is prioritizing prevention for adolescent girls and
young women in 10 African countries

with a multifaceted prevention program,
which includes PrEP. In Kenya, a national
roll-out program for PrEP is prioritizing delivery to highest-burden counties,
bringing services to populations most at
risk, and using messaging that is simultaneously entertaining, appealing, and
informative [18]; this is a new way to do
public health.
The results from Volz and colleagues
touch on an important, but potentially
sensitive, topic for prioritizing prevention interventions. The idea of “targeting”
prevention, to an individual or to a group,
may be off-putting, both to those targeted
(who wants to have a target applied to
them?) and to others (who may have
the impression that PrEP is otherwise
denied to them). The best available evidence to date suggests that open access
to PrEP results in high uptake and likely
diminished stigma and discouragement.
Moreover, individuals seeking PrEP
appear to be commonly self-selecting for
being at risk, with high prevalence and
ongoing incidence of curable sexually
transmitted infections (harbingers of HIV
exposure) and corresponding behavioral
risks [3]. Conversely, new HIV infections
in PrEP-accessing persons have sometimes been concurrent with loss of PrEP
access (eg, because of loss of insurance
coverage) [7]. Public health agencies and
others working on PrEP outreach may be
able to address these areas of challenge
by prioritizing access to PrEP for all who
want it, with directed messaging and marketing to those for whom PrEP needs and
potential for benefit are greatest. Indeed,
in the United Kingdom, the setting for
the work by Volz and colleagues, privately-sourced PrEP, guided by websites providing instructions on how to access the
medication prior to its incorporation into
the National Health Service but without
public health targeting at all, has been reported to have resulted in a 40% decline
in new HIV infections in the last year in
London [19].
PrEP is an intervention for an individual, but with enough individuals taking PrEP, particularly if infections among

1510 • JID 2018:217 (15 May) • EDITORIAL COMMENTARY

Downloaded from https://academic.oup.com/jid/article-abstract/217/10/1509/4915988
by University of California, San Diego Library user
on 08 May 2018

those most likely to acquire and pass on
the virus are blocked, population-level
benefits will follow. Bringing together
robust public health, cutting-edge molecular science, quality clinical care and behavioral science, innovative community
engagement and messaging, and strong
political will amplify the prevention benefits of PrEP.
Notes

Financial support.  This work was
supported by the National Institute
of Mental Health of the US National
Institutes of Health (grant number R01
MH095587).
Potential conflict of interest. The author has led studies of pre-exposure
prophylaxis in which study medication
was donated by Gilead Sciences and has
served on an advisory committee for
Gilead Sciences. The author has submitted the ICMJE Form for Disclosure of
Potential Conflicts of Interest. Conflicts
that the editors consider relevant to the
content of the manuscript have been
disclosed.
References

Grant RM, Lama JR, Anderson PL,
1.
et al. iPrEx Study Team. Preexposure
chemoprophylaxis for HIV prevention in men who have sex with men.
N Engl J Med 2010; 363:2587–99.

2. Baeten JM, Donnell D, Ndase P,
et  al. Partners PrEP Study Team.
Antiretroviral prophylaxis for HIV
prevention in heterosexual men
and women. N Engl J Med 2012;
367:399–410.
McCormack S, Dunn DT, Desai M,
3.
et  al. Pre-exposure prophylaxis to
prevent the acquisition of HIV-1
infection (PROUD): effectiveness
results from the pilot phase of a pragmatic open-label randomised trial.
Lancet 2016; 387:53–60.

4. Molina JM, Charreau I, Spire B,
et  al. ANRS IPERGAY Study Group.
Efficacy, safety, and effect on sexual
behaviour of on-demand pre-exposure
prophylaxis for HIV in men who have

sex with men: an observational cohort
study. Lancet HIV 2017; 4:e402–10.
Liu AY, Grant RM, Buchbinder SP.
5.
Preexposure prophylaxis for HIV:
unproven promise and potential pitfalls. JAMA 2006; 296:863–5.
6. Baeten JM, Heffron R, Kidoguchi L,
et  al. Integrated delivery of antiretroviral treatment and pre-exposure
prophylaxis to HIV-1-serodiscordant
couples: a prospective implementation study in Kenya and Uganda.
PLoS Med 2016; 13:e1002099.

7. Marcus JL, Hurley LB, Hare CB,
et  al. Preexposure prophylaxis for
HIV prevention in a large integrated
health care system: Adherence, renal
safety, and discontinuation. J Acquir
Immune Defic Syndr 2016; 73:540–6.

8. Office of National AIDS Policy.
National HIV/AIDS strategy for
the United States: updated to 2020.
Washington, DC:  US government,
2015.

9. Joint United Nations Programme
on HIV/AIDS. HIV prevention
2020 road map—accelerating HIV
prevention to reduce new infections by 75%. http://www.unaids.
org/en/resources/documents/2017/

hiv-prevention-2020-road-map.
Accessed 29 January 2018.
1
0. Peters PJ, Pontones P, Hoover KW,
et  al. HIV infection linked to injection use of oxymorphone in Indiana,
2014–2015. N Engl J Med 2016;
375:229–39.
1
1. Poon AF, Gustafson R, Daly P, et  al.
Near real-time monitoring of HIV
transmission hotspots from routine
HIV genotyping: an implementation case study. Lancet HIV 2016;
3:e231–8.
12. Little SJ, Kosakovsky Pond SL, Anderson
CM, et  al. Using HIV networks to
inform real time prevention interventions. PLoS One 2014; 9:e98443.
1
3. Campbell EM, Jia H, Shankar A, et al.
Detailed transmission network analysis of a large opiate-driven outbreak
of HIV infection in the United States.
J Infect Dis 2017; 216:1053–62.
1 Mera R, Magnuson D, Trevor H,
4.
Bush S, Rawlings K, McCallister S.
Changes in Truvada for HIV pre-exposure prophylaxis utilization in the
USA: 2012–2016. 9th International
AIDS Society Conference on HIV
Science. Paris, France, 2017: Abstract
WEPEC0919.

1
5. Smith DK, Van Handel M, Wolitski
RJ, et  al. Vital signs: Estimated percentages and numbers of adults with
indications for preexposure prophylaxis to prevent HIV acquisition–
United States, 2015. MMWR Morb
Mortal Wkly Rep 2015; 64:1291–5.
16. HIV/AIDS Epidemiology Unit, Public
Health–Seattle and King County and
the Infectious Disease Assessment
Unit, Washington State Department
of Health. HIV/AIDS epidemiology
report 2017, Volume 86. https://
www.kingcounty.gov/depts/health/
communicable-diseases/hiv-std/
patients/epidemiology.aspx. Accessed
29 January 2018.
1
7. Straube T. Florida to roll out free
PrEP in 2018. https://www.poz.com/
article/florida-roll-free-prep-2018.
Accessed 29 January 2018.
18. Jipende JiPrEP. https://twitter.com/
NimejiPrEP. Accessed 29 January
2018.
1
9. Wilson C. Massive drop in London
HIV rates may be due to internet
drugs. https://www.newscientist.com/
article/2117426-massive-drop-in-london-hiv-rates-may-be-due-to-internetdrugs/. Accessed 29 January 2018.

EDITORIAL COMMENTARY • JID 2018:217 (15 May) • 1511

Downloaded from https://academic.oup.com/jid/article-abstract/217/10/1509/4915988
by University of California, San Diego Library user
on 08 May 2018

1
2
3
4
5
6
7

Exhibit C

8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

ADDENDUM TO MPAS IN SUPPORT OF AMICUS CURIAE
2:18-cv-00490-JAM-KJN

J Immigrant Minority Health (2014) 16:229–238
DOI 10.1007/s10903-012-9741-0

ORIGINAL PAPER

Undocumented Immigration Status and Diabetes Care Among
Mexican Immigrants In Two Immigration ‘‘Sanctuary’’ Areas
A. Elizabeth Iten • Elizabeth A. Jacobs
´
Maureen Lahiff • Alicia Fernandez

•

Published online: 2 November 2012
Ó Springer Science+Business Media New York 2012

Abstract The objective of this study is to investigate the
relationship between immigration status and the patient
experience of health care, diabetes self-management, and
clinical outcomes among Mexican immigrants with diabetes receiving health care in two immigration sanctuary
cities. We used data from the Immigration, Culture and
Health Care study, a cross-sectional survey and medical
record study of low-income patients with diabetes recruited
from public hospitals and community clinics in the San
Francisco Bay Area and Chicago. Undocumented Mexican,
documented Mexican immigrants, and US-born Mexican–
Americans’ health care experiences, diabetes self-management, and clinical outcomes were compared using
multivariate linear and logistic regressions. We found no
significant differences in reports of physician

communication, or in measures of diabetes management
between undocumented and documented immigrants. All
three groups had similar clinical outcomes in glycemic,
systolic blood pressure, and lipid control. These results
indicate that, at least in some settings, undocumented
Mexican immigrants with diabetes can achieve similar
clinical outcomes and report similar health care experiences as documented immigrants and US-born Mexican–
Americans.
Keywords Immigrants Á Diabetes Á Hispanic/Latino Á
Mexican–American Á Sanctuary Á Disparities Á
Undocumented immigrants

Introduction
We recognize that some Mexican immigrants, regardless of
immigration status, consider themselves Mexican–American and
others do not. For sake of clarity we will use the terms documented
and undocumented immigrants to refer to this heterogeneous group
[1].
A. E. Iten
UC Berkeley-UCSF Joint Medical Program, Berkeley, CA, USA
E. A. Jacobs
Departments of Medicine & Population Health Sciences and
Health Innovation Program, University of Wisconsin School of
Medicine & Public Health, Madison, WI, USA
M. Lahiff
School of Public Health, University of California, Berkeley,
Berkeley, CA, USA
´
A. Fernandez (&)
Division of General Internal Medicine, SFGH, University
of California, San Francisco, Box 1364, San Francisco,
CA 94143, USA
e-mail: [e-mail]

Nearly 60 % of the of the estimated 10.3 million undocumented immigrants living in the United States are Mexican
[2, 3] and nearly half of the Mexican population living in
the United States is undocumented [4–6]. Undocumented
immigrants are less likely to have health insurance [7, 8], to
access and use health care services [9–13], and to report
having a primary care provider [14] or a regular source of
care [15]. Studies have found that those with concerns
about deportation are at heightened risk of emotional health
problems and report poorer health status [14, 16]. While
disparities in access to health care and multiple social
stressors may translate into poorer clinical outcomes for
undocumented immigrants, few studies have examined
how immigration status relates to clinical health outcomes
or to patient experiences of health care.
An examination of the impact of immigration status on
outcomes is important to population health. Despite having
lower prevalence of many medical conditions compared to

123

230

the general population in the United States, Mexican
Americans are more likely to be at greater risk for morbidity and mortality related to chronic illness, particularly
diabetes [17, 18]. Mexican Hispanics in the US are 1.7
times as likely to have diabetes as non-Hispanic whites
[19] and an estimated 9.5 % of Latino adults suffer from
diabetes [20]. Diabetes is increasingly common among
foreign-born Latinos as well, likely reflecting changes in
obesity in source countries, such as Mexico [21]. Prior
research indicates that lack of documentation—and the fear
associated with it—are powerful deterrents to seeking
health care [22], however it is not clear how immigration
status impacts the care or the outcomes of diabetes.
We set out to address this research gap by analyzing data
from a survey and medical record abstraction study that
included US-born Mexican Americans and Mexican immigrants with diabetes receiving care in two sanctuary areas in
the US. Sanctuary areas are cities, counties, or states that
have policies and statutes limiting routine reporting of the
documentation status of people seeking police or health
services [23]. Advocates have argued that sanctuary policies
at an institutional level allow for hospital administrators and
health care providers to create trust in the local immigrant
community [24]. It is plausible that a ‘safe’ environment
created by these policies may decrease some of the barriers to
health care access and foster trust with clinicians, thereby
potentially resulting in better adherence to diabetes self-care
practices and medications, and positive clinical outcomes. It
is also plausible that the social burdens associated with
undocumented status, which include stigma and barriers to
employment and financing, might render diabetes self-care
very difficult, or spill over into patient doctor interactions
and thus negatively impact clinical outcomes. We specifically set out to evaluate whether undocumented immigration
status in two sanctuary areas is associated with: (1) poorer
perceived doctor-patient interactions; (2) poorer diabetes
self-management (which include measures of diabetes selfcare, self-care barriers/supports, diabetes self-efficacy, and
medication adherence); and (3) poorer control of clinical
outcomes: blood glucose, blood pressure, and cholesterol
when compared to documented Mexican immigrants and
US-born Mexican Americans.

J Immigrant Minority Health (2014) 16:229–238

nine free-standing or hospital-based safety-net clinics in
the San Francisco Bay Area and Chicago in 2008–2009.
The main purpose of the ICHC was to explore factors that
impact diabetes self-management and health outcomes in
minority populations. To be included in the study, patients
had to have type 2 diabetes, be 18 years of age or older,
and speak English or Spanish. Patients who exhibited
cognitive impairment, active substance abuse, and/or psychosis were excluded. Recruitment was stratified by race/
ethnicity and patient language in order to ensure a diverse
sample. The participation rate among eligible patients was
91 %.
After providing written informed consent in English or
Spanish, participants completed an in-person survey with a
trained bilingual research assistant. Clinical data was
abstracted from participants’ electronic health record. The
values for glycosylated hemoglobin (A1C), low-density
lipid (LDL) cholesterol and systolic blood pressure recorded within 1 year prior and closest to the date of the
interview were abstracted. For this specific study, we
analyzed data for the 401 patients of the ICHC study
sample population who self-identified as Mexican or
Mexican American. All analyses measuring patient perception of culturally competent care were performed with
the additional inclusion criterion of having reported a
consistent primary care provider during the last 12 months
(N = 317). This study was approved by the Committee for
Protection of Human Subjects at the University of California, San Francisco; Cook County Health and Hospital
System; UC Berkeley; and by participating institutions in
Chicago and the San Francisco Bay Area.
Measures
Immigration Status

Methods

Immigration status was measured using responses to
questions about country of birth and US citizenship and
permanent residency with a green card for foreign-born
participants. Participants who reported the United States as
their country of birth were considered US-born Mexican
Americans. Immigrants who reported having US citizenship or legal permanent residency were considered to be
documented. Participants who reported neither status were
categorized as undocumented immigrants by exclusion.

Participants and Data Collection

CAHPS-Cultural Competency

We analyzed data from the Immigration, Culture and
Health Care (ICHC) Study, a cross-sectional study of a
convenience sample of African American, Spanish- and
English-speaking Mexican/Mexican American, and nonHispanic white adults with diabetes who received care in

We used the trust and positive communication subdomains
from the Consumer Assessments of Healthcare Providers
and Systems’ Cultural Competency Item Set (CAHPS-CC)
to assess patient experiences of care. CAHPS-CC is a
26-item set of Likert-response questions designed to

123

J Immigrant Minority Health (2014) 16:229–238

231

measure patients’ overall experience of their physician’s
interpersonal and cultural competence as well as their
experience of their physician’s office. CAHPS-CC has
been validated for use in ethnically diverse low-income
populations in English and Spanish [25]. Self-reports were
heavily skewed towards positive responses. Therefore,
responses were dichotomized into two categories [26]: the
upper 25 % (‘‘optimal’’) and the lower 75 % (‘‘suboptimal’’). Further details about this tool are provided in the
Appendix.

Health Outcomes

Diabetes Self-Management Measures

Covariates

Diabetes demands daily self-management, and people with
this chronic disease generally need to make lifestyle
modifications to achieve successful glycemic control. We
used several instruments to capture different facets of
patient self-management that we believed may be associated with life as an undocumented immigrant: (1) Diabetes
self-care was measured through a brief adapted version of
the Summary of Diabetes Self-Care Activities (SDSCA)
scale [27]. This instrument represents a multidimensional
process that calls for daily engagement in a complex set of
behaviors relating to diet, exercise, foot care, and glucose
monitoring recommendations. Many of these life-style
modifications may be impacted by competing demands
associated with immigrant employment and lifestyle [28].
(2) Diabetes self-care barriers/supports were measured
using a diabetes-related health belief instrument translated
for use with Spanish-speaking Mexican Americans [29].
These questions focus on barriers to implementing lifestyle
changes, beliefs in the benefits of effective self-management, perceptions that one can control the effects of diabetes, beliefs regarding the impact of one’s job on diabetes
therapy, and perceptions of support for their diabetes provided by family and friends. Of these different beliefs, the
perceptions of control over one’s diabetes and perception
of social support are potentially the most critical factors.
Close familial and social relationships create social support
that can improve glucose control [30]. However, immigration status may have an impact on these through changes in social and family configurations common to
undocumented life [31]. 3) Diabetes self-efficacy was
measured through an 8-item scale originally developed and
tested in Spanish for the Diabetes Self-Management study
[32]. Self-efficacy, or the ability to perform self-care tasks,
has been shown to have a positive effect on glycemic
control, quality of life, and adherence to diet, exercise, and
blood glucose testing [33]. (4) Medication adherence was
measured using the Morisky scale [34]. These two last
factors are influenced by trust in one’s physician [35],
which in turn may be affected by immigration status.
Details about these scales are provided in the Appendix.

We included the following variables in our adjusted model:
age, gender, highest level of education achieved, employment status, marital status, diabetes duration, BMI, and
number of comorbidities (past myocardial infarction,
transient ischemic attack/stroke/cerebrovascular accident,
cancer, hypertension, arthritis, and hypercholesterolemia).
Regressions comparing documented and undocumented
Mexican immigrants additionally included English proficiency and number of years in the US. Food insecurity,
which is defined as the risk of going hungry because of an
inability to afford food [37], and is common in this lowincome population [38], was measured using the wellvalidated six-item Food Security Survey Module [39].
Food security is associated with both hypo and hyperglycemia [40].

Measurements of hemoglobin A1C (A1C), systolic blood
pressure (SBP), low-density lipoprotein (LDL) and weight
were obtained by clinical chart review, selecting the last
measurement prior to the survey interview. Categorical
health outcomes were determined by clinical recommendations [36] for people with diabetes and established as
follows: poor A1C control C8 %, high SBP C130 mmHg,
and high LDL C100 mg/dL.

Analysis
Chi-square tests (for differences of proportions) and t tests
(for differences in means) were used to examine the
association between immigration status and socio-demographic characteristics, perceived doctor-patient interactions, diabetes-related behaviors, and clinical outcomes
after testing that the normality assumption was met. Clinical outcomes (A1C, systolic BP, and LDL) were analyzed
both as binary variables determined by clinical cut-off
recommendations and as continuous variables.
Multivariable linear and logistic regression analyses
were conducted to calculate beta coefficients or adjusted
odds ratios, and 95 % confidence intervals for health care
experiences, diabetes self-management, and the three
clinical outcomes. Each model included age, gender,
highest level of education achieved, employment status,
marital status, diabetes duration, BMI, and number of
comorbidities. The adjusted model for the comparison
between undocumented and documented Mexican immigrants additionally included limited English proficiency
and number of years in the US in the regression. If there
was no evidence of association between immigration status
and the outcomes, power calculations were carried out to

123

232

determine if group sizes were large enough to detect clinically important differences across the three groups.

Results
Of the 401 subjects in the study, 124 (31 %) were US-born
Mexican Americans, 166 (41.4 %) documented Mexican
immigrants, and 111 (27.7 %) undocumented Mexican
immigrants. Socio-demographic and clinical characteristics
are shown in Table 1. Gender, annual income, type of
residence, and report of food insecurity did not vary across
the three groups. While similar with respect to education,
marital status, and acculturation, undocumented Mexican
immigrants were more likely than documented immigrants
to be younger (50 vs. 56 years, p \ 0.05), employed (39 %
vs. 24 %, p \ 0.05), have lived in the US for fewer number
of years (15 vs. 32, p \ 0.05) and have limited English
proficiency (77 % vs. 63 %, p \ 0.05). Compared to USborn Mexican Americans, undocumented Mexican immigrants were more likely to be employed (39 % vs. 26 %,
p \ 0.05) and married/living together (56 % vs. 29 %,
p \ 0.05), less likely to have a high school education
(12 % vs. 38 %, p \ 0.05) and much less likely to be
Anglo-oriented on the Acculturation Rating Scale for
Mexican Americans-II [41] (4 % vs. 80 %, p \ 0.05).
Patients in all three groups reported a mean diabetes
duration of about 10 years. Undocumented immigrants
reported somewhat fewer comorbidities than both documented immigrants and US-born Mexican Americans (1.8
vs. 2.1 vs. 2.3, p \ 0.05).
Table 2 examines participant report of patient perception of culturally competent care and diabetes-related
behaviors across the three groups. There were no statistically significant differences between the three groups on
either physician trust or reports of positive physician
communication. Diabetes self-care behaviors (healthy diet,
exercise, blood glucose testing, and foot care) did not differ
between the two immigrant groups or between the undocumented and US-born. These results need to be interpreted
cautiously as power calculations indicate less than 80 %
power to detect statistical differences with this sample size
and data driven standard deviations.
When compared to US-born Mexican Americans,
undocumented immigrants were more likely to, on a 1–5
scale, perceive barriers to following a diabetic diet and
taking medications (3.2 vs. 2.7, p \ 0.001) and report
impact of their job on their therapy (2.8 vs. 2.4,
p = 0.013). Power calculations indicate at least 80 %
power to detect the differences in these two diabetes selfcare barriers.
Results from separate multivariable logistic and linear
regression analyses of patient perception of culturally

123

J Immigrant Minority Health (2014) 16:229–238

competent care and diabetes-related behaviors also showed
no statistically significant differences among the three
groups, with the exception of perceived barriers to following a diabetic diet and taking medications. Documented
and undocumented immigrants are 1.2 and 1.3 times more
likely than US-born Mexican Americans to perceive higher
level of these barriers to diet and medication adherence
(p = 0.003 and p = 0.024, respectively).
Table 3 shows the association of diabetes intermediate
clinical outcomes with immigration status. While the
prevalence of poor glycemic, blood pressure and lipid
control was high, with 40–60 % having poor glycemic
control, US-born Mexican Americans and undocumented
and documented Mexican immigrants did not differ significantly in either mean glycemic, blood pressure, or lipid
control, measured via means or proportion in poor control.
Table 4 provides the results of separate multivariable
logistic and linear regression analyses for each clinical
outcome. Undocumented immigrants, documented immigrants, and US-born Mexican Americans had similar odds
of poor control (glycemic, blood pressure, and lipids). They
also show comparable results in the continuous clinical
outcome variables. Documented Mexican immigrants had
lower A1C values and were at lower odds of having
uncontrolled diabetes (A1C C8.0 %) than US-born Mexican Americans. Other clinical outcomes did not differ.
Power calculations determined that the study sample size
has over 80 % power to detect differences in good versus
poor control for each comparison.

Discussion
We report on the experience of diabetes care among
Mexican immigrants receiving care in two immigration
sanctuary areas [42, 43] in the US where people seeking
health services are not asked about immigration legal status, nor is immigration status reported to immigration
officials. In this setting, we found that undocumented
immigrants achieved comparable clinical outcomes and
reported similar experiences of health care as documented
immigrants and US-born Mexican Americans. Undocumented immigrants did not differ from documented
immigrants in their ability to manage and control their
blood sugar, blood pressure, or blood lipids. While we are
underpowered to detect small yet significant differences in
doctor patient interactions and trust, we note that undocumented immigrants reported similar perceived doctorpatient interactions and diabetes self-management-related
behaviors. Perhaps contributing to the rest of our findings,
we found that these undocumented immigrants were as
likely as documented immigrants and US-born Mexican
Americans to report trust in their primary care physician.

J Immigrant Minority Health (2014) 16:229–238

233

Table 1 Socio-demographic characteristics, acculturation, limited English proficiency and food insecurity by immigration status (N = 401)
US-born Mexicans

Mexican immigrants

p valuea

N = 124
N (%)

Documented
N = 166
N (%)

Undocumented
N = 111
N (%)

Undocumented
versus US-born
Mexicans

San Francisco

38 (30.7)

72 (43.4)

43 (38.7)

Chicago

86 (69.3)

94 (56.6)

68 (61.3)

51.8 ± 13.9

55.7 ± 10.8

49.8 ± 12.5

Clinic Site

0.19

Age (years), mean ± SD

0.44

0.081

<0.001

108 (65.1)

0.072

0.46

44 (35.5)

0.023

0.25

60 (54.0)
78 (70.3)

Educational level
Incomplete high school

0.36

0.094

83 (50.0)

0.51

0.036

64 (51.6)

<0.001

<0.001

Female

0.24
0.71

Gender

Completed high school/GED

47 (37.9)

30 (18.1)

13 (11.7)

Some advanced degree

33 (26.6)

28 (16.9)

20 (18.0)

Full or part-time

32 (25.8)

39 (23.5)

43 (38.7)

Unemployed

67 (54.0)

75 (45.2)

42 (37.8)

25 (20.2)

52 (31.3)

26 (23.4)

\$10,000

42 (33.9)

34 (20.5)

34 (30.6)

$10,000–$24,999

49 (39.5)

73 (44.0)

38 (34.2)

C$25,000

27 (21.8)

44 (26.5)

23 (20.7)

Unsure/declined

6 (4.8)

15 (9.0)

16 (14.4)

Rent

68 (54.8)

79 (47.6)

68 (61.2)

Own/with family

53 (42.7)

78 (47.0)

39 (35.1)

Shelter/homeless

3 (2.4)

9 (5.4)

4 (3.6)

Employment status

Retired/willingly unemployed
Annual household income

Type of residence

Marital status
Married/living together

36 (29.0)

103 (62.1)

Divorced/separated/widowed

49 (39.5)

47 (28.3)

31 (27.9)

Single, never married

39 (31.45)

16 (9.6)

18 (16.2)

–

31.95 ± 11.63

14.60 ± 8.44

Years in US, mean ± SD

62 (55.9)

Acculturation
Anglo oriented

88 (80.0)

15 (9.5)

Limited English proficiency

–

105 (63.2)

86 (77.5)

Food insecurity

54 (43.5)

80 (48.2)

54 (48.6)

N/A
<0.001

<0.001
0.082

4 (3.9)
N/A

0.012

Obese (C 30.0 kg/m2)

0.43

0.94

0.305

BMI

0.088

82 (66.1)

Diabetes duration (years), mean ± SD
b

Number of comorbidities , mean ± SD
a

Undocumented versus
documented immigrants

96 (57.8)

65 (59.6)

11.75 ± 10.35

11.27 ± 9.71

9.24 ± 9.92

0.059

0.094

2.26 ± 1.20

2.13 ± 1.08

1.78 ± 1.03

0.001

0.007

Significant differences (p \ 0.05) indicated in bold type

b

Comorbidities included past myocardial infarction, transient ischemic attack/stroke/cerebrovascular accident, cancer, hypertension, arthritis,
and hypercholesterolemia

Previous studies have shown that undocumented immigrants are at heightened risk of not having health insurance
or access to health care services [7, 8, 12, 13]. This results
in limited use of these services and a decreased likelihood
of having a regular source of care and a primary care
provider [9–11, 14, 15]. Although Cavazos-Rehg et al. [16]

found that Latino immigrants with concerns about deportation reported poorer subjective health status, to our
knowledge our study is the first of its kind that investigates
the association between immigration status and a chronic
disease health outcomes. Contrary to our hypothesis, we
found that undocumented immigrants’ health outcomes

123

234

J Immigrant Minority Health (2014) 16:229–238

Table 2 Participant report of patient perception of culturally competent care and diabetes-related behaviors
Mexican immigrants

p valuea

N = 124
Mean ± SD

Documented
N = 166
Mean ± SD

Undocumented
N = 111
Mean ± SD

Undocumented
versus US-born
Mexicans

Optimal

50 (48.5)

66 (48.9)

33 (41.8)

Suboptimal

53 (51.5)

69 (51.1)

46 (58.2)

Optimal

41 (39.8)

46 (34.1)

23 (29.1)

Suboptimal

62 (60.2)

89 (65.9)

56 (70.9)

4.17 ± 2.32
4.20 ± 2.38

4.24 ± 2.48
3.67 ± 2.37

2. Exercise

4.13 ± 2.55

3. Blood glucose testing

5.11 ± 2.48

4. Foot-care

5.08 ± 2.64

US-born
Mexicans

Undocumented
versus documented
immigrants

CAHPS-CCb,c
1. Trust

0.36

0.31

0.13

0.45

3.93 ± 2.46
3.62 ± 2.46

0.44
0.068

0.30
0.87

3.97 ± 2.62

3.86 ± 2.54

0.43

0.74

4.42 ± 2.58

4.51 ± 2.53

0.062

0.69

5.05 ± 2.71

5.14 ± 2.72

0.87

0.81

2. Positive communication

Diabetes self-care (0–7 days/week)
1. Healthy diet
General diet
Specific diet

Diabetes self-care barriers/supports (0–5, low–high)
1. Barriers to following a diabetic
diet and taking medications

2.74 ± 0.77

3.04 ± 0.76

3.15 ± 0.69

<0.001

0.19

2. Social support for diet

3.58 ± 0.92

3.58 ± 0.97

3.75 ± 0.86

0.15

0.13

3. Impact of job on therapy

2.43 ± 0.77

2.72 ± 0.96

2.76 ± 0.89

0.013

0.75

4. Benefits of therapy

4.14 ± 0.49

4.14 ± 0.37

4.14 ± 0.35

0.90

1.00

5. Control of effects of diabetes

3.16 ± 1.24

3.35 ± 1.20

3.29 ± 1.15

0.42

0.67

7.59 ± 1.40
1.04 ± 1.10

7.47 ± 1.66
1.16 ± 1.13

7.61 ± 1.68
1.02 ± 1.06

0.91
0.87

0.50
0.30

Diabetes self-efficacy (1–10, low–high)
Medication adherence (0–4, good-poor)
a

Significant differences (p \ 0.05) indicated in bold type

b

Dichotomous variables are given as number (percentage)

c

CAHPS-CC, The California Assessment of Healthcare Providers and Systems – Cultural Competency. A validated tool used to measure patient
perception of culturally competent care for those with a steady primary health care provider (N = 317)

were comparable to those of both documented Mexican
immigrants and US-born Mexican Americans in this setting,
despite having several risk factors such as younger age [7]
and limited English proficiency that are known to be associated with poorer glycemic control [44]. Our finding that
reports of trust in physician and positive communication
among doctor-patient interactions were also similar among
the three groups may help explain the comparable clinical
results achieved [45–48], though again we note sample-size
limitations. It is possible that the high levels of physician
trust reported are an indirect consequence of immigration
sanctuary policies that help create a safe environment where
undocumented immigrant populations can seek out health
care and manage their diseases. However, as we did not
study immigrants in non-sanctuary cities, we can only
speculate about the connection between these policies and
physician trust or clinical outcomes we report.

123

With the exception of employment status, the demographic characteristics of the undocumented immigrants
in this study were consistent with socio-demographical
profiles of this population from previous studies [28].
The fact that undocumented immigrants in our study are
more likely to be employed may reflect issues related to
(un)employment in San Francisco and Chicago. Our
survey instrument did not specify employment in the
formal sector. Undocumented immigrants are more likely
to hold informal, low-skilled jobs and less likely to be in
white-collar occupations [49], and the somewhat higher
reported employment rate in our study reported by
undocumented immigrants may represent sporadic informal employment. We found no differences in income
among the three groups, reinforcing this interpretation.
This lack of difference in income may partly account for
the lack of difference in clinical outcomes we report.

J Immigrant Minority Health (2014) 16:229–238

235

Table 3 Clinical outcomes by immigration status (N = 401a)
US-born Mexican
N = 124

Documented
N = 166
Mean ± SD

p valueb

Mexican immigrants

Mean ± SD

Undocumented
N = 111
Mean ± SD

Undocumented
versus US-born
Mexicans

Undocumented
versus documented
immigrants

HbA1c C 8.0 %c

67 (60.9)

63 (41.2)

50 (49.5)

0.096

0.19

HbA1c (%)

8.70 ± 2.22

7.98 ± 1.76

8.45 ± 2.11

0.41

0.063

Systolic BP C 130 mmHgc

54 (43.9)

83 (50.3)

47 (42.3)

0.81

0.19

Systolic BP (mmHg)

130.46 ± 18.07

130.2 ± 17.03

126. 07 ± 18.58

0.069

0.063

LDL C 100 mg/dLc

40 (42.1)

48 (35.6)

30 (33.7)

0.24

0.78

LDL (mg/dL)

94.28 ± 33.26

89.70 ± 33.67

88.29 ± 33.49

0.23

0.76

a

The number of subjects in each clinical outcome is limited to those whose lab values were present during chart review

(HbA1c N = 364, Systolic BP N = 399, LDL N = 319)
b

Significant differences (p \ 0.05) indicated in bold type

c

Dichotomous variables are given as number (percentage)

Table 4 Adjusted and unadjusted odds ratio and beta coefficients (95 % CI) for clinical outcomes by immigration status
Undocumented versus documented
Mexican Immigrants

Undocumented Mexican immigrants
versus US-born Mexicans

Documented Mexican
immigrants versus US-born
Mexicans

Unadjusted b/OR
(CI)

HbA1c, mean (%)

Unadjusted b/OR
(CI)

Adjusteda b/OR
(CI)

Unadjusted
b/OR (CI)

Adjusteda
b/OR (CI)

1.4

1.5

0.63

0.57

0.45

0.51

(0.85,2.32)

HbA1c C 8.0 %

Adjustedb b/OR
(CI)

(0.75,3.04)

(0.36,1.09)

(0.31,1.07)

(0.27,0.74)

(0.29,0.90)

0.47

0.53

-0.24

-0.39

20.72

20.59

(-0.03,0.98)

(-0.09,1.14)

(-0.79,0.30)

(-0.96,0.18)

(21.21,20.22)

(21.10,20.07)

Systolic
BP C 130 mmHg

0.73

0.83

0.94

1.39

1.29

1.56

(0.45,1.18)

(0.43,1.61)

(0.56,1.58)

(0.77,2.50)

(0.81,2.07)

(0.92,2.65)

Systolic BP, mean
(mmHg)

-4.13

-3.57

-4.38

-1.12

-0.26

1.22

(-8.42,0.17)

(-9.02,1.87)

(-8.96,0.20)

(-5.93,3.69)

(-4.42,3.91)

(-3.11,5.55)

LDL C 100 mg/dL

0.92

1.04

0.7

0.69

0.76

0.86

LDL, mean (mg/dL)

(0.53,1.62)
0.84

(0.45,2.37)
2.93

(0.38,1.27)
-5.94

(0.35,1.35)
-6.45

(0.44,1.30)
-6.79

(0.47,1.57)
-4.82

(-7.64,9.33)

(-8.16,14.02)

(-15.11,3.23)

(-15.99,3.10)

(-15.10,1.53)

(-13.43,3.80)

a

Model adjusted for age, gender, highest level of education achieved, employment status, marital status, diabetes duration, BMI, and number of
comorbidities

b

Adjusted model additionally includes limited English proficiency and number of years in the US

Significant differences (p \ 0.05) indicated in bold type

The generalizability of this study is limited. First, all
participants were receiving clinical care at community
clinics in sanctuary areas, where immigration status is not
ascertained or shared with immigration enforcement.
Therefore, the results are only representative of an immigrant population that has access to health care services and
can receive care in a legally safe environment. As importantly, the study enrolled patients in a primary care setting;
undocumented immigrants with diabetes who have great

mistrust of the health system may refuse all but episodic
care and their clinical outcomes would likely differ from
those enrolled in primary care. Third, the cross-sectional
design of the study provides only one snapshot in time. As
national debates on immigration continue, the lack of
association between immigration status and health status or
perception of patient-doctor interaction may change.
Finally, the relatively small number of participants enrolled
cannot allow us to exclude small yet statistically significant

123

236

differences in clinical outcomes. This is particularly true of
our analysis of reports of diabetes self-care behaviors
where we had less than 80 % power to detect statistically
significant differences. The study also has several
strengths. It involves a relatively large number of undocumented immigrants, has detailed patient report on the
physician-patient relationship and on self-care behavior,
and includes measures of diabetes outcomes.
In conclusion, this analysis suggests that undocumented
Mexican immigrants are able to achieve comparable clinical outcomes and diabetes self-management behaviors as
documented immigrants and US-born Mexican Americans,
at least in some environments. While this may reflect the
impact of legally safe access to health care, much more
research, in many different environments, is needed before
that conclusion can be drawn. Future studies should
investigate the association between immigration status and
health outcomes when immigrants do not reside in sanctuary cities and should draw participants from community
settings. Yet, while rates of poor clinical control were high
among all three groups–underscoring the need for
improvement in Latino diabetes outcomes overall—it is
noteworthy that despite the many social burdens associated
with illegal immigration status, undocumented patients and
their clinicians are able to successfully partner in diabetes
care at least in some settings. As the Patient Protection and
Affordable Care Act, which promises to expand health care
coverage to millions of Americans, excludes undocumented immigrants, these patients are likely to continue to
concentrate in relatively few health care delivery settings.
It may be reassuring to clinicians and policy makers to
know that, at least in some settings, undocumented immigration status need not result in worse diabetes outcomes.
Acknowledgments This research was funded by the Russell Sage
Foundation, Arnold P. Gold Foundation, UCSF School of Medicine
Dean’s Summer Research Fellowship, and the UC Berkeley-UCSF
Joint Medical Program Helen Schoeneman Research Fellowship. We
acknowledge Dr. Karen Sokal-Gutierrez and Dr. Sylvia Guendelman
of the UC Berkeley School of Public Health for their valuable comments. We are grateful for the patients and staff of participating
clinics.

Appendix
Details on the different measures discussed in the methods
section.
CAHPS-Cultural Competency
The CAHPS-CC has seven subdomains: positive doctor
communication, negative doctor communication, health
promotion, alternative medicine, shared decision-making,

123

J Immigrant Minority Health (2014) 16:229–238

equitable treatment, and trust. Internal consistency for the
Immigration, Culture and Health Care (ICHC) study population was determined by Cronbach alpha (0.82 for positive communication, 0.77 for trust, 0.72 for preventive care
counseling). Negative communication and equitable treatment were not included in this study due to their low
Cronbach alpha in the Spanish-speaking population.
Shared decision-making and alternative medicine are also
excluded due to their overall low Cronbach alpha. Due to
highly skewed distributions and as a proof of concept,
scores for the two domains included in the study were
dichotomized into two categories [26]: the upper 25 %
(optimal) and the lower 75 % (suboptimal). Scores for each
of the CAHPS-CC subscales range from 0 to 100.
Diabetes Self-Management
(1) Diabetes self-care is measured through a brief version
of the Summary of Diabetes Self-Care Activities (SDSCA)
scale [27]. It is a self-report questionnaire that measures
levels of self-management across different components of
the diabetes regimen: general diet (2 items), specific diet (2
items), exercise (2 items), blood-glucose testing (2 items),
foot care (2 items), and smoking (not included in the ICHC
Study survey). Due to issues of recall, we decided to use
only the item of each measure that asked about behaviors
within the immediate past week. For the foot care measure,
we only used the item that asked about checking feet and
not inspecting the inside of shoes. Responses range from 0
to 7 (days a week) with higher scores indicating better
diabetes self-management. (2) Diabetes self-care barriers/
supports were measured using a diabetes-related health
belief instrument translated for use with Spanish-speaking
Mexican Americans [29]. It is a 25-item health belief
instrument from Starr County, Texas consisting of five
subscales: Social support for diet, Impact of job on therapy,
Benefits of therapy, Control of effect of diabetes, and Total
barriers to diet and taking medications. Responses were
given in a Likert-scale format, ranging 1 (strongly disagree) to 5 (strongly agree). The higher the score on an
item, the stronger the belief. (3) Diabetes self-efficacy was
measured through an 8-item scale originally developed and
tested in Spanish for the Diabetes Self-Management study
[32]. Responses range from 1 (not at all confident) to 10
(completely confident) and the score for this scale is the
average of the eight items. Higher averages indicate higher
self-efficacy. (4) Medication adherence was measured
using the Morisky scale, a four-item self-reported adherence measure (Cronbach alpha = 0.61) that addresses
barriers to medication-taking. Responses are yes/no categories and the score is calculated by assigning 1 point for
each ‘‘yes’’ answer, thus ranging from 0 to 4. Higher scores
indicate poorer medication adherence [34].

J Immigrant Minority Health (2014) 16:229–238

References
1. Taylor P, Lopez MH, Martinez JH, Velasco G. When labels don’t
fit: hispanics and their views of identity. Washington: Pew Hispanic Center; 2012.
2. Passel J. Unauthorized migrants: numbers and characteristics.
Washington: Pew Hispanic Center; 2005.
3. Nandi A, Galea S, Lopez G, Nandi V, Strongarone S, Ompad DC.
Access to and use of health services among undocumented
mexican immigrants in a us urban area. Am J Public Health.
2008;98:2011–20.
4. Annual estimates of the resident population by sex, Race,
and Hispanic Origin for the United States. US Census Bureau;
2009. Accessed 5 July 2010 http://www.census.gov?_jpc/www/
usinterimproj.
5. American community survey. US Census Bureau; 2006. Accessed 5 July 2010 http://factfinder.census.gov/servlet/DatasetMain
PageServlet?_program=ACS&_Submenuld=&_lang=en&_ts.
6. Passel J. Growing share of immigrants choosing naturalization.
Washington: Pew Hispanic Center; 2007.
7. Goldman D, Smith JP, Sood N. Legal status and health insurance
among immigrants. Health Aff. 2005;24:1640–53.
8. Ku L, Matani S. Left out: immigrants’ access to health care and
insurance. Health Aff. 2001;20:247–56.
9. Chavez L, Cornelius W, Jones OW. Mexican immigrants and the
utilization of US health services: the case of San Diego. Soc Sci
Med. 1985;21:93–102.
10. Berk ML, Schur CL, Chavez LR, Frankel M. Health care use
among undocumented latino immigrants. Health Aff. 2000;19:
51–64.
11. Fuentes-Afflick E, Hessol N. Immigration status and use of health
services among Latina women in the San Francisco bay area.
J Women’s Health. 2009;18:1275–80.
12. Hubbell FA, Waitzkin H, Mishra SI, Dombrink J, Chavez LR.
Access to medical care for documented and undocumented
Latinos in a Southern California County. West J Med. 1991;154:
414–7.
13. Ortega AN, Fang H, Perez VH, Rizzo JA, Carter-Pokras O,
Wallace SP, Gelberg L. Health care access, use of services, and
experiences among undocumented Mexicans and other Latinos.
Arch Intern Med. 2007;167:2354–60.
14. Marshall KJ, Urrutia-Rojas X, Mas FS, Coggin C. Health status
and access to health care of documented and undocumented
immigrant Latino women. Health Care Women Int. 2005;26:
916–36.
15. DeVoe JE, Wallace LS, Pandhi N, Solotaroff R, Fryer GE Jr.
Comprehending care in a medical home: a usual source of care
and patient perceptions about healthcare communication. J Am
Board Fam Med. 2008;21:441–50.
16. Cavazos-Rehg P, Zayas L, Spitznagel E. Legal status, emotional
well-being and subjective health status of Latino immigrants.
J Natl Med Assoc. 2007;99:1126–31.
17. Hunt KJ, Williams K, Resendez RG, Hazuda HP, Haffner SM,
Stern MP. All-cause and cardiovascular mortality among diabetic
patients in the San Antonio heart study. Diabetes Care. 2002;
25:1557–63.
18. Hunt KJ, Gonzalez ME, Lopez R, Haffner SM, Stern MP,
Gonzalez-Villalpando C. Diabetes is more lethal in Mexicans and
mexican-Americans compared to non-hispanic whites. Ann Epidemiol. 2011;21:899–906.
19. National Diabetes Fact Sheet: (2005) General information and
national estimates on diabetes in the United States. Atlanta, GA:
US Department of Health and Human Services, Center for Disease Control and Prevention 2005.

237
20. Livingston G, Minushkin S, Cohn D. Hispanics and health care in
the United States: access, information and knowledge. Washington: Pew Hispanic Center and Robert Wood Johnson Foundation; 2008.
21. Wild S, Roglic G, Green A, Sicree R, King H. Global prevalence
of diabetes: estimates for the year 2000 and projections for 2030.
Diabetes Care. 2004;27:1047–53.
22. Berk ML, Schur CL. The effect of fear on access to care among
undocumented latino immigrants. J Immigr Health. 2001;3:151–6.
23. Villazor R. ‘Sanctuary cities’ and local citizenship. Fordham
Urban Law J. 2010;37:573–98.
24. Marrow H: The Power of Local Autonomy: Expanding Health
Care to Unauthorized Immigrants in San Francisco. Ethnic and
Racial Studies 2011; 1-16.
25. Seligman H, Stern RJ, Fernandez A, Jacobs E, Neilands T,
Maldonado RW, Carle A, Quan J: Advances in Measuring Culturally Compentent Care: A Confirmatory Factor Analysis of
CAHPS-CC in a Safety-Net Population. Unpublished manuscript
2010.
26. Zweifler J, Hughes S, Lopez RA. Controlling for race/ethnicity: a
comparison of California commercial health plans CAHPS scores
to NCBD benchmarks. Int J Equity Health. 2010;9:1–9.
27. Toobert D, Hampson S, Glasgow R. The summary of diabetes
self-care activities measure: results from 7 studies and a revised
scale. Diabetes Care. 2000;23:943–50.
28. Passel J. The size and characteristics of the unauthorized migrant
population in the US. Washington: Pew Hispanic Center; 2006.
29. Brown SA, Becker HA, Garcia AA, Barton SA, Hanis CL.
Measuring health beliefs in spanish-speaking Mexican Americans
with type 2 diabetes: adapting an existing instrument. Res Nurs
Health. 2002;25:145–58.
30. Brown SA, Blozis SA, Kouzekanani K, Garcia AA, Winchell M,
Hanis CL. Health beliefs of Mexican Americans with type 2
diabetes: the Starr county border health initiative. The Diabetes
Educator. 2007;33:300–8.
31. Viruell-Fuentes EA, Schulz AJ. Toward a dynamic conceptualization of social ties and context: implications for understanding
immigrant and Latino health. Am J Public Health. 2009;99:
2167–75.
32. Stanford Patient Education Research Center: Diabetes Self-Efficacy Scale. Palo Alto: Stanford University. http://patienteducation.stanford.edu/research/sediabetes.html.
33. Williams KE, Bond MJ. The roles of self-efficacy, outcome
expectancies and social support in the self-care behaviours of
diabetics. Psychology, Health & Medicine. 2002;7:127–41.
34. Morisky D, Green L, Levine D. Concurrent and predictive
validity of a self-reported measure of medication adherence. Med
Care. 1986;24:67–74.
35. Lee Y, Lin J. The effects of trust in physician on self-efficacy,
adherence and diabetes outcomes. Soc Sci Med. 2009;68:1060–8.
36. American Diabetes Association. Standards of medical care for
patients with diabetes mellitus. Diabetes Care. 2003;26:S33–50.
37. Nord M, Andrews M, Carlson S: Household Food Security in the
United States, 2008. United States Department of Agriculture
(Economic Research Service); 2009.
38. Seligman HK, Jacobs EA, Lopez A, Tschann J, Fernandez A.
Food insecurity and glycemic control among low-income patients
with type 2 diabetes. Diabetes Care. 2012;35:233–8.
39. Bickel G, Nord M, Price C: Guide to measuring household food
security. United States Department of Agriculture (Food and
Nutrition Service); Alexandria, VA: 2000.
40. Seligman HK, Jacobs EA, Lopez A, Sarkar U, Tschann J,
Fernandez A. Food Insecurity and hypoglycemia among safety
net patients with diabetes. Arch Intern Med. 2011;171:1204–6.

123

238
´
41. Cuellar I, Arnold B, Maldonado R. Acculturation rating scale for
Mexican Americans-II: a revision of the original ARSMA scale.
Hisp J Behav Sci. 1995;17:275–304.
42. City & County of San Francisco: Sanctuary Ordinance. 2000.
Accessed 6 February 2012 http://sfgsa.org/index.aspx?page=1067.
43. Ohio Jobs and Justice Political Action Committee: Sanctuary
Cities USA. 2012. Accessed 6 February 2012 http://ojjpac.org/
sanctuary.asp.
44. Fernandez A, Schillinger D, Warton EM, Adler N, Moffet HH,
Schenker Y, Salgado MV, Ahmed A, Karter AJ. Language barriers, physician-patient language concordance, and glycemic
control among insured Latinos with diabetes: the diabetes study
of Northern California (DISTANCE). J Gen Intern Med.
2010;26:170–6.
45. Blendon RJ, Buhr T, Cassidy EF, Perez DJ, Hunt KA, Fleischfresser C, Benson JM, Hermann MJ. Disparities in health:

123

J Immigrant Minority Health (2014) 16:229–238

46.

47.

48.

49.

perspectives of a multi-ethnic, Multi-Racial America. Health Aff.
2007;26:1437–47.
Sixma HJ, Kerssens JJ, Campen CV, Peters L. Quality of care
from the patient’s perspective: from theoretical concept to a new
measuring instrument. Health Expect. 1998;1:82–95.
Harmsen JA, Bernsen RM, Brujinzeels MA, Meeuwesen L.
Patient’s evaluation of quality of care in general practice: what
are the cultural and linguistic barriers. Patient Educ Couns.
2008;72:155–62.
´
Rodrıguez M, Bustamante A, Ang A. Perceived quality of care,
receipt of preventive care, and usual source of health care among
undocumented and other Latinos. J Gen Intern Med. 2009;24:
S508–13.
Passel Jeffrey, Cohn D’Vera. A portrait of unathorized immigrants
in the United States. Washington, DC: Pew Hispanic Center; 2009.

1
2
3
4
5
6
7

Exhibit D

8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

ADDENDUM TO MPAS IN SUPPORT OF AMICUS CURIAE
2:18-cv-00490-JAM-KJN

Ethnic and Racial Studies Vol. 35 No. 1 January 2012 pp. 72Á87

The power of local autonomy: expanding
health care to unauthorized immigrants in
San Francisco
Helen B. Marrow
(First submission June 2010; First published August 2011)

Abstract
I analyse the mechanisms through which an inclusive local policy
environment in the integrated city-county of San Francisco operates to
improve unauthorized immigrants’ access to and utilization of health
care. Within Á but not outside Á the bounds of this inclusive local context,
committed providers reported being able to provide attention to
unauthorized immigrants without worrying about the direct costs of
doing so; to buffer, marshal resources, and advocate for individual
unauthorized patients; and to exert substantial autonomy in deciding
how to approach lack of legal status in their patientÁprovider interactions. These results highlight the potential and limitation of sub-national
policies seeking to ameliorate unauthorized immigrants’ health vulnerability in a hostile US federal context.

Keywords: Immigration; incorporation; unauthorized; health care; safety net;
autonomy.

Introduction
The federal and state health care policy context toward the estimated
11.1 million unauthorized immigrants living in the USA today has
been described as so decidedly hostile that it leaves little leeway for
government officials, health care providers, and immigrant advocates
to make the situation more inclusive, even when they want to (Newton
and Adams 2009). With few exceptions, restrictive government policies
have rendered unauthorized immigrants ineligible for most federally
funded public health insurance Á such as Medicare, regular Medicaid,
and State Children’s Health Insurance Program (SCHIP) Á since the
# 2011 Taylor & Francis
ISSN 0141-9870 print/1466-4356 online
http://dx.doi.org/10.1080/01419870.2011.594168

The power of local autonomy 73

early 1970s (Schwartz and Artiga 2007; Fox 2009). All unauthorized
immigrants qualify for select public health and nutrition measures Á
including immunizations, the Special Supplemental Nutrition Program for Women, Infants and Children (WIC), and testing and
treatment for communicable diseases Á but they can only qualify for a
limited form of Emergency Medicaid (which covers labour and
delivery and other designated emergencies) if they fall into certain
categories like low-income children or pregnant women, and they can
only qualify for non-emergency care in a handful of states that use
their own state funds to offer it.
In addition, unauthorized immigrants face a range of indirect
eligibility restrictions. Many are effectively barred or deterred from
seeking care even in federally funded institutions that do not in
theory restrict care based on legal status. This is because they are
employed in informal jobs, move constantly between jobs, and live in
overcrowded housing, so they often have difficulty producing income
tax forms or utility bills that can serve as proof of local residency
and low income Á two bureaucratic criteria that are required for
admission into such institutions (Portes, Light, and Fernandez-Kelly
´
2009; Portes, Fernandez-Kelly, and Light 2011).
´
Together with other barriers like fear, direct and indirect eligibility
restrictions lead to some of the most severe disparities in access to and
utilization of care among comparable populations in national, state,
and local studies (Goldman, Smith and Sood 2005; Ortega et al. 2007).
Moreover, under the Health Care and Education Reconciliation Act of
2010, unauthorized immigrants will not be eligible to receive federal
subsidies to purchase their own private insurance, nor will they be
allowed to purchase health insurance through new state-based health
insurance exchanges, even if they pay with their own money. In fact,
unauthorized immigrants are projected to become a full one-third of
the remaining 23 million uninsured Americans by 2019 (Pear and
Herszenhorn 2010).
If government officials, health care providers, immigrant advocates,
and other actors want to reduce disparities by legal status Á whether to
help prevent the spread of infectious diseases, reduce the cost of
preventable emergency care, or help institutions comply with ethical
stances that support the provision of care to all humans, all residents
of their communities, or all workers Á they must look to other creative
alternatives. One viable alternative is the national network of federally
qualified health centres (FQHCs), which offer a variety of primary,
mental, and dental services to unauthorized immigrants across the
country and which, like public hospitals, do not in theory restrict care
based on legal status. The Health Care and Education Reconciliation
Act of 2010 did increase federal funding to FQHCs, and this will

74 Helen B. Marrow

certainly help to reduce some, but not most, disparities in access to
and utilization of care for unauthorized immigrants.
Other creative alternatives are bi-national, although these too may
be problematic since unauthorized immigrants face increasing restrictions on moving back and forth across international borders.
A third set of creative alternatives consists of inclusive sub-national
policies that may be enacted at the state and local levels in receiving
communities Á especially since new patterns of geographic dispersion
have brought unauthorized immigrants into an unprecedented array of
states and localities, all of which are now struggling to determine how
best to respond to their presence. In this article, I ask: what are the
mechanisms through which inclusive local policy environments can
operate to improve unauthorized immigrants’ access to and utilization
of health care, specifically via the actions of providers and staff
working in public health care safety nets that they govern?
Site selection and methods
To identify such mechanisms, I conducted a case study of thirty-six
safety net health care providers and staff who work in a large,
residency-training, outpatient clinic Á hereafter called Hospital Outpatient Clinic (HOC). HOC is associated with the public safety net
hospital of the integrated city and county of San Francisco, which
exhibits a uniquely inclusive local policy environment toward unauthorized immigrants but which continues to be embedded in the
more restrictive federal and state context described above. Examining
providers and staff in a public safety net clinic such as HOC is valuable
because it is they who constitute the front-line or street-level bureaucratic arms of local governments, and who have some discretion to
interpret, enact, and enforce government policies during the execution
of their work, even while remaining heavily influenced by rules and
bureaucratic processes (Lipsky 1980).1 HOC provides comprehensive
primary care and select specialty services Á most FQHCs do not
provide the latter Á and is one of the city’s Healthy San Francisco
(HSF) medical homes. Like its parent hospital, HOC serves a diverse
patient population that is predominantly low income, uninsured, and
racially, ethnically, or linguistically in the minority.
Between May and September 2009, I sought out a variety of providers
and staff in HOC through a combination of purposive and snowball
sampling. Purposively, I wanted to include a range of types (from
physicians to non-physician staff) who come into contact with unauthorized immigrants in different statuses and roles. Respondents
included five physicians; seven resident physicians-in-training; and
twenty-four non-physician providers and staff members, including eight
registered nurses, three nurse practitioners, seven medical evaluation

The power of local autonomy 75

assistants, four clerical staff, one social worker, and one health worker.
I also conducted interviews with an additional eighteen safety net
providers and staff (including two hospital Medi-Cal eligibility staff)
working in other hospital clinics and departments, a nearby Latinooriented FQHC, and a nearby Latino day labourer-oriented free clinic
in order to uncover their perspectives on how unauthorized immigrants
view and interact with providers and staff at HOC and its parent
hospital.
Interviews lasted between forty-five and ninety minutes and over
two-thirds (thirty-eight of the fifty-four respondents) were conducted
in isolation, although due to their workday time constraints the
remainder (sixteen) were interviewed in small sets of focus groups. I
tape-recorded, transcribed, cleaned, coded, and analysed all interviews
using ATLAS.ti, a qualitative analysis software program. To ensure
anonymity, I have changed all names and identifying characteristics of
individual respondents.
Creating an inclusive policy climate
Local government officials in San Francisco have worked hard to
create an inclusive and less stigmatizing environment for unauthorized
immigrants than exists at the federal level or in most other localities,
one that is consistent with the city’s vanguard reputation for being on
the leading edge of progressive social and political change. They have
allocated relatively generous funds to the city’s public safety net, which
stands at the country’s leading edge of promoting culturally and
linguistically competent care and is anchored by a communityoriented acute care public teaching hospital affiliated with a wellrespected academic medical centre. This public teaching hospital gets
referrals for specialty care from its own internal outpatient clinics, a
system of closed satellite public outpatient clinics, and another system
of affiliated non-profit FQHCs. Providers and staff working within the
infrastructure are paid on public salaries with local Department of
Public Health funds.
Local government officials in San Francisco have also enacted
measures that separate lack of legal status from the provision and
receipt of public services and benefits. First, they have strengthened
their commitment to an official sanctuary policy. Originally passed as
a symbolic resolution in 1985 to declare the city a refuge for, and to
prohibit discrimination against, Salvadoran and Guatemalan refugees,
San Francisco’s sanctuary policy has evolved into an active ordinance
in the city’s Administrative Code. Recently, the ordinance has been
subjected to a federal grand jury investigation (ongoing) to determine
whether or not it violates federal immigration law but, through it, San
Francisco has joined over fifty other American localities to actively

76 Helen B. Marrow

prohibit: (a) the asking or collection of any information on legal status
other than that required by state/federal statute, court decision, or
regulation, or by federal, state, or local public assistance criteria; and
(b) the cooperation of public service providers with federal immigration officials regarding any persons not under investigation or
convicted of felonies (Tramonte 2009).
Second, local government officials recently approved a municipal ID
ordinance (effective 15 January 2009), making San Francisco the
second city in the country after New Haven, Connecticut, to offer a
municipal identification card to all city residents regardless of legal
status. The ordinance’s originators were primarily interested in the
benefits it would bring to the city’s approximately 40,000 unauthorized
immigrants, yet they were also careful to design and frame the
ordinance inclusively to better withstand public criticism and avoid
stigmatizing the card’s future holders (de Graauw 2009). Thus,
although the ID card does not grant any new services or benefits to
unauthorized immigrants, it makes them easier to access. Both the
sanctuary and municipal ID ordinances acknowledge unauthorized
immigrants’ de facto legitimacy to be part of San Francisco’s civic
community, based on what Ridgley (2008) and de Graauw (2009) term
a conception of local ‘inhabitance’ or ‘residence’ (e.g. jus domicili)
rather than birthright, ancestry, or legalistic citizenship.
Third, local government officials enacted and committed substantial
local public funds to San Francisco Healthy Kids (SFHK) (effective
2002) and Healthy San Francisco (HSF) (effective April 2007). SFHK
provides subsidized health care plans to all local resident children aged
up to eighteen who do not qualify for other forms of federal or state
public insurance (including regular Medi-Cal and Healthy Families Á
California’s regular Medicaid and SCHIP programmes) regardless of
legal status. Similarly, HSF provides ‘universal access’ to primary
medical care to all local resident adults aged eighteen to sixty five who
have incomes under 500 per cent of the federal poverty line but do not
qualify for other forms of federal or state public insurance coverage,
regardless of legal status. Participation is free if residents’ incomes fall
below the federal poverty line; otherwise it is based on designated
quarterly participation and point-of-service fees. However, services
covered in the HSF universal access model are not equivalent to
insurance coverage. They are limited to those primary care services
provided by participating health care institutions (to date, almost
exclusively public safety net ones) or otherwise funded by HSF
monies, and a range of specialty and select primary care services are
not covered, including dental, vision, organ transplants, and long-term
care.
Thus, although San Francisco is not immune to the conservative
pressures that have increased class and racial inequality and put severe

The power of local autonomy 77

pressure on the public social safety net nationwide; even though its
two main policy efforts to divorce lack of legal status from the
provision and receipt of local public services and benefits have come
under strong attack; and even though its HSF universal access model
remains ‘categorically unequal’ (Light 2011), the city exemplifies a
much more inclusive and less stigmatizing environment than does
either the state of California or the nation as a whole. HOC
respondents, especially those who have worked in other states and
localities, including some in the surrounding Bay area, noted such
‘exceptionalism’ frequently and emphatically.
Local bureaucratic autonomy: self-selecting into the safety net and
providing primary care
HOC providers feel that they have actively self-selected themselves into
the San Francisco safety net environment. Over the course of their
medical training, all selected: (1) primary care, which is lower paying
and less prestigious than specialty care; (2) the American social safety
net, which is devoted to serving under-served populations; and (3)
living and working in San Francisco, which is one of the most
politically liberal cities in the country. Taken together, self-selection
shapes their positive attitudes toward unauthorized immigrants, whom
many considered to be equally deserving of care either because they
are human beings (human rights perspective), members of a disadvantaged and under-served population (social justice perspective),
members of the local community (public health and community
perspective), positive contributors to the American economy (deserving worker perspective), or simply ‘sick’ and in need of care
(humanitarian perspective). While some variation did exist among
respondents in the degrees to which Á and various rationales for why Á
they supported providing care to unauthorized immigrants, all
exhibited an inclusive attitude, and several reported that public safety
net hospital’s institutional culture imposes sanctions on providers and
staff who openly disagree.
San Francisco’s inclusive policy climate helps HOC providers put
their attitudes into practice in several ways. First, it allows them to
provide care to unauthorized immigrants without having to worry
about direct costs. As physician Charlotte explained, San Francisco’s
public-salaried payment structure insulates them from having to ‘eat’
the direct costs of treating uninsured patients, making them less
reluctant to treat them than many providers working in private
practice. Likewise, San Francisco’s generous funding of public health
care insulates these public safety net providers from the ‘frustration’
of unfunded mandates to treat unauthorized immigrants who are
uninsured, which HOC providers view as a more pressing issue in

78 Helen B. Marrow

other localities with more restrictive policy environments. As Charlotte
stated:
We all take a pay cut to work in a safety net institution. . . We do it
because it frees us from some of the structural problems of private
practice. And there is. . . a whole [San Francisco] city understanding
for investing in this infrastructure. [It is] a place where you don’t
have [that] level of frustration. . . like ‘‘Okay, so we have to deal with
the unfunded patients and do not get any resources from the local
government to do so.’’
Indeed, local HSF investment Á which, as clerical worker Shana
reported ‘kicks in the money’ for many services that federal and state
policies do not fund for poor people and unauthorized immigrants Á
allows HOC providers, in nurse practitioner Sarah’s words, to offer
‘access to better than 90 per cent’ of primary care services without
thinking or asking about patients’ legal statuses. According to Sarah:
There’s just once in a while something you can’t do. And I feel lucky
that I don’t really care [about legal status]. It doesn’t, you know, for
the most part it doesn’t really affect what we can do for people.
HOC providers knew that their ability to disregard patients’ legal
statuses would disappear if local sanctuary or insurance policy were
to become more restrictive, since, according to physician Joseph, it
would:
become much more germane to know if someone’s unauthorized
because you want to know what benefits you can try to get people
access to. If that’s [a] criterion for excluding people from benefits
then, you know, we would probably start asking [patients about
their legal statuses].
Physician Mary agreed that providers:
often don’t know [legal status] because we are very lucky in San
Francisco in having no [legal or financial constraints placed on us]
for anything we can provide on-site [at the public safety net hospital]
to anyone who lacks health insurance.
She went even further than Charlotte or Sarah to explain how
additional local investment even allows providers to link patients to
care at other area institutions through a system of city contracts if the
public safety net hospital does not provide a particular service on site:
‘If we don’t have a whole department that provides it here, the city

The power of local autonomy 79

actually has a contract where they pay for it, usually at [another
nearby academic medical centre], to buy the care for patients there.’ As
Mary showed, San Francisco’s inclusive local policy, which includes a
generous allocation of local funding to the city’s safety net, allows
HOC providers to more effectively marshal resources for individual
uninsured patients.
This inclusive context also allows providers to more effectively
muster resources and advocate for such patients in other ways. For
instance, the city’s sanctuary policy reaffirms many providers’ beliefs
that unauthorized immigrants are often scared to seek health care
services because of the restrictive federal policy toward unauthorized
immigrants. Sanctuary policy supports them in their efforts to engage
in what Horton (2006) and Lamphere (2005) call ‘buffering’ strategies
as they try to smooth and compensate for such fear (Konczal and
Varga 2011). At an institutional level, buffering strategies include
attempts by hospital administrators to create trust in the local
immigrant community (by advertising a ‘safe’ hospital context, in
which data on lack of legal status will not be unlawfully collected or
transmitted to authorities) and the creation of a standard, neutrally
coloured hospital ‘gold card’ (which outpatient clinic providers and
staff can use to view patients’ medical record numbers instead of
requesting their citizenship and legal status information directly).
At the individual level, buffering strategies include attempts by
hospital Medi-Cal eligibility workers to reassure applicants that they
only ask about citizenship and legal status insofar as it is required to
determine plans and payers (e.g. federal or state public insurance
programmes or the two local initiatives, SFHK and HSF), never for
disqualification purposes. As medical evaluation assistant Marta
attested, front-desk clerks and medical evaluation assistants reported
accepting alternative documents (e.g. medical record numbers) instead
of social security numbers when checking in patients for appointments; sometimes they also ‘run after’ patients whom they notice are
visibly fearful and so ‘turn around and leave’, in order to reassure
them that ‘it is safe here’.
In the more insulated back rooms of the clinic, physicians, residents,
registered nurses, and nurse practitioners also reported going to great
lengths to advise and advertise to patients that San Francisco is a ‘safe’
context where it is okay to utilize services and programmes. Likewise,
they reported encouraging patients to collaborate with ‘safe’ social
workers (who understand eligibility rules and can help both providers
and patients ‘work around’ confusing legal status restrictions to access
more resources) and drawing on their professional networks to request
compassionate assistance from external non-safety net providers when
they are unable to provide certain services internally. As nurse
practitioner Lynne described:

80 Helen B. Marrow

I really do encourage people. ‘‘It’s okay. You’re not going to get
arrested. You’re not going to get deported just because you’re
seeking health care. You can use your real name.’’ Or, ‘‘If you’re
really scared, go to the refugee clinic.’’ Or I’ll try to send them to the
social worker to get some referrals to a Spanish-speaking advocacy
agency where they can get reassurance if that’s what they need.
Throughout the clinic, providers and staff all reported engaging in
buffering strategies to reduce unauthorized immigrants’ fear, including
not asking patients about their legal statuses directly, heavily downplaying such requests when they do (e.g. prefacing requests with ‘It
doesn’t matter to me. . .’), never documenting patients’ lack of legal
status directly in patients’ records, and utilizing patients’ social
networks (both inside and outside the clinic) to encourage greater
utilization of resources.
Perhaps most importantly, San Francisco’s inclusive policy grants
these providers substantial autonomy to decide if, how, and why they
will approach lack of legal status in their patientÁprovider interactions
in order to offer unauthorized patients with the most effective medical
care. Some providers choose to actively ‘ignore’ or ‘look beyond’
patients’ legal statuses, to comply not only with the city’s sanctuary
policy but also with their dominant professional norm to ‘suspend
judgument’ and ‘not disenfranchise’ patients. As physician Charlotte
explained, ‘We try to treat people the same no matter what. . . do our
damnedest not to think about [legal status].’ In fact, registered nurse
Jane emphatically described a strategy of never asking patients about
legal status ‘not because are trying to avoid the issue, but rather
because we are trying to get around it to help people and give them
equal care. It just interferes with medical care to bring [legal status]
up.’ For this group of providers, San Francisco’s inclusive context not
only strengthens the dominant professional norm of ‘don’t ask, don’t
know, don’t care’ regarding legal status but also legitimates their
cognitive beliefs that ideally lack of legal status should not matter to
health care delivery. Resident Eduardo even interpreted this as an
additional directive to ‘follow the same algorithm no matter what’ and
treat all patients equally.
Other providers choose different approaches to offer culturally
competent and compensatory care. San Francisco’s inclusive local
policy environment assists them by creating a secure environment in
which they can, according to physicians Mary and Elena, more easily
seek out patients’ ‘social histories of migration’ to help ‘contextualize
their medical conditions’. Most providers in this group continued to
refrain from asking patients directly about their legal statuses, fearing
that doing so would give them a false impression of stigmatization or
service restriction. Instead, they tried to elicit the information from

The power of local autonomy 81

patients indirectly, often making inference to it through ‘related’
characteristics such as their recent arrival, day labourer or caregiver
occupational status, lack of English language skills, separation from
family members abroad, and/or inability to travel internationally.
Context-oriented providers saw such characteristics associated with
conditions of illegality, as causing patients stress and trauma, and
as compromising patients’ health and ability to follow care recommendations.
While San Francisco’s inclusive policy climate does not solve the
internal provider debate as to whether to ‘ignore’ or ‘acknowledge’
lack of legal status, it does set both approaches within a more
protective and enfranchising context.
Where autonomy ends: referring to specialty and ancillary care
The restrictive federal and state context in which San Francisco’s
inclusive local policy is embedded does not necessarily change HOC
providers’ desires to provide care for unauthorized immigrants. They
are, after all, part of the heavily self-selected social safety net.
Nonetheless, restrictive federal and state policy does limit the range
of resources that HOC providers can offer to unauthorized patients,
force providers to directly engage patients’ legal statuses when they
might not otherwise do so, and depress providers’ strategies for
buffering and advocating for individual unauthorized patients.
This important set of limits on providers’ bureaucratic autonomy is
most clearly visible at two critical junctures Á the first between primary
medical care and specialty medical care, and the second between
primary medical care and ancillary social support care. Government
officials have enacted and committed substantial local public funds to
two programs (SFHK and HSF) that in theory expand access to care
to all low-income children and adults who are San Francisco city
residents. Nevertheless, as a universal access model, HSF remains
categorically unequal (Light, 2011) with respect to other forms of
public health insurance Á even Medi-Cal and Healthy Families Á in
that it only includes primary care provided by participating health care
institutions or otherwise funded by HSF monies. HSF does not cover
certain specialty care services (including dental and vision) or other
ancillary services (including public housing, General Assistance (GA),
Supplemental Security Income, food stamps, disability, or hospice).
Unauthorized immigrants’ access to these services lies outside the
domain of local San Francisco policy and continues to be delimited by
more restrictive federal and state polices (WIC is a notable exception).
Consequently, HOC providers’ ways of dealing with patients’ lack of
legal status change dramatically as they cross the line separating
locally covered primary medical care services from other specialty care

82 Helen B. Marrow

and ancillary services. They are suddenly forced into thinking and
asking about lack of legal status. According to health worker Mariana,
clerical workers and medical evaluation assistants learn about patients’
legal statuses not only when they ‘need to know [what insurance is]
going to cover some specific test we are setting up, or if patients need a
pre-authorization to do that’ (specialty services), but also when they
‘need to send patients to the social worker to see if there are any [social
support] resources available.’ Similarly, physician Elena does not
usually have to ask about legal status and is ‘able to provide standard
of care for the majority of my patients who are chronically ill’ without
knowledge of it since ‘the city and county of San Francisco commits
amazing, amazing resources to provide an enormous amount of
things.’ However, for the small group of patients who do become
‘severely ill, or have the wrong thing’, it matters because they ‘just can’t
get [specialty] care’ and ‘it becomes really hard, depending on what the
service is.’ It is rare that Elena has ‘to come flat out and ask a patient,
‘Are you documented?’. Nevertheless, in a ‘clinically exigent situation
when patients need a specialty service that requires they are U.S.
citizens or legal immigrants’, she is forced to ask.
As a result, providers like Elena see clear patterns of ‘blocked access’
emerging for unauthorized patients regarding select high-tech specialty procedures such as organ transplants, open MRIs, nuclear
medicine tests, coronary bypass or bariatric surgeries, endoscopies,
cystoscopies, screening colonoscopies, intervention cardiology procedures, and PET or DEXA scans, because such services are either not
offered on site at the public safety net hospital or not covered by HSF
or other local, state, or federal monies. Coming up against these
barriers, HOC providers reported going into advocacy mode, trying
desperately to ‘twist some arms’ and find ways to link their
unauthorized patients to care. In a few cases their efforts have been
successful, but as resident doctor Laura explained of the time when an
external allergist agreed to see one of her unauthorized patients who
had recurrent anaphylaxis, such success is ‘voluntary’ and ‘discretionary’ rather than systemic, and it declines as the cost of the
specialty procedure rises. In most situations, providers reported that
their ‘hands are tied’ and that their efforts to buffer and advocate for
their unauthorized patients fall short, as happened to physician Mary:
[My patient] is someone who by. . . every criterion would get a liver
transplant. She’s socially stable, she’s married, she’s adherent to
absolutely everything that you ask her to do, there’s like nothing
wrong. And I asked the liver specialist here to see her [but] as soon
as they found out she didn’t have papers it was like very clear [that
she would not be treated]. . . That’s just a devastating conversation
to have [with a patient].

The power of local autonomy 83

Likewise, physician Elena argued that lack of legal status quickly
becomes ‘determinant of the care one receives’ at the point of
transition into high-tech specialty care:
If I advocate hard enough for an African-American patient who
needs a particular service outside the ones we provide [at the public
safety-net hospital], usually I can get it. There’s usually all sorts of
hoops to jump through but I can get it.
However, with unauthorized patients her advocacy strategies prove
ineffective: ‘I just can’t.’ As registered nurse Harriet confirmed, in the
context of high-tech specialty care, ‘When you’re unauthorized,’ providers have even less [ability to] ‘work around.’ At this point providers
described the frustration of watching unauthorized patients unable to
access needed services, or only able to access insufficient stopgap or (in
resident Kate’s words) ‘band-aid’ emergency services that do not
constitute a ‘long-term solution’. Several even began debating the pros
and cons of advising patients to return to their home countries to try
to obtain specialty medical care services there.
The restrictive federal and state policy context also depresses these
public safety net providers’ strategies for buffering and advocating for
individual unauthorized patients in the realm of ancillary services,
where rules governing access are strict and strongly enforced. As
physician Mary said, even in remarkable cases, where the city does fund
certain specialty medical care services for unauthorized immigrants,
HOC providers’ ‘hands get tied’ when accessing critical support
services like unemployment, disability, or public housing that would
allow patients to support themselves and their families as they heal:
When I sent a patient to the social workers, I asked them, ‘‘Is there
any miracle we can pull off here [hooking him up to unemployment
or disability benefits]?’’ And they basically said ‘‘No.’’ And at this
point, you know, the city’s about to pay $100,000 to get an ICD
[implantable cardioverter-defibrillator] implanted in him [for cardiac arrythmia]. So it’s hard. We work to send him to the food bank
and stuff, but he’s basically losing his housing and it’s just a mess.
He wound up having to send his children, who are American-born
and are U.S. citizens, and his wife back to his home country, because
he can’t afford to keep them fed or anything. He’s someone who,
because he can get this procedure, should be able to recover, be a
productive member of our society, and be able to raise two kids who
will be, too. But there’s nothing we can do right now.
In fact, restrictive policy context and the ever-looming threat of budget
cuts keeps many HOC providers’ constantly aware of, in Mary’s words,

84 Helen B. Marrow

the ‘power of the state’. It curtails their willingness to ‘rock the boat too
much’ or to encourage unauthorized patients to apply for services
they ‘know they can’t get’, lest such actions bring about a restrictive
backlash that could jeopardize what access to primary care unauthorized immigrants do have.
Conclusion
I have examined the mechanisms through which inclusive local policy
environments can operate to improve unauthorized immigrants’ access
to and utilization of health care, specifically via the actions of streetlevel bureaucratic providers and staff working in public health care
safety nets that they govern. HOC providers have come to their jobs
both strongly selected and heavily committed to expanding access, and
providing culturally competent care to vulnerable patient populations,
including unauthorized immigrants. By and large they reported that
the inclusive San Francisco policy context allows them to give care to
unauthorized immigrants without worrying about the direct costs of
doing so; to buffer, marshal resources, and advocate for individual
unauthorized patients; and to exert substantial autonomy in deciding
how to approach lack of legal status in their patientÁprovider
interactions Á to the point that some even go beyond the official
bounds of the city’s sanctuary ordinance to extend what they view as
more effective medical care to unauthorized patients in ways they
believe to be consistent with the ordinance’s broader spirit.
Nevertheless, HOC providers also reported that their advocacy
efforts and bureaucratic autonomy break down at two critical
junctures during the transition from an inclusive local to a more
restrictive federal and state policy climate, especially as the cost of
high-tech specialty medical services rises or when federal and state
regulations regarding ancillary services are strictly codified and
enforced. In these realms, care becomes more ‘discretionary’ and
‘voluntary’ Á successful only in a few ‘miracle’ cases and frustratingly
unsuccessful most of the time.
These results carry important practical and theoretical implications
for policymakers, health care providers, and advocates. Most importantly, they demonstrate that sub-national strategies such as San
Francisco’s are imperfect substitutes for including unauthorized
immigrants within the bounds of federal and state health insurance
and social welfare programmes. Even in San Francisco, it is ‘access’
rather than insurance that is the goal, since full insurance is still
deemed to be unaffordable. The continued exclusion of unauthorized
immigrants from national and state programmes means that even in
San Francisco, local providers still face difficulties working around
specialty and ancillary problems in order to care for unauthorized

The power of local autonomy 85

immigrants. ‘San Francisco is not paradise,’ physician Elena reported,
and ‘if these problems exist here,’ resident Carla said, ‘you know
they’re everywhere else.’ To fully overcome these barriers, HOC
respondents noted, the American public must ultimately change its
mindset about unauthorized immigrants to see them as more
deserving of inclusion and financial investment.
Yet the results of my study also highlight the real potential for subnational actors to play a positive role in enacting and implementing
local strategies that can help overcome some of the barriers to access
and utilization. Such strategies may be politically and financially
difficult to enact elsewhere. They will likely require supportive, or at
least neutral, backing from local communities and a fiscal base large
enough to support redistribution. San Francisco is an extremely
wealthy and politically liberal city whose public has proven willing to
support and contribute taxes to progressive local policies, and whose
politicians have committed what physician Elena describes as ‘amazing’ and ‘generous’ resources to its health care infrastructure.
Still, government officials in San Francisco note that even prior to
the implementation of HSF, the city was already paying substantial
amounts to care for the uninsured, including unauthorized immigrants,
and so the programme does not necessarily represent an infusion of
new money into the safety net system. Rather, it was conceived as a way
to integrate, further de-stigmatize, and make more efficient the robust
safety net that the city already had in place (SF DPH and SF OLSE
2010). In this regard, San Francisco can serve as an important model
for other localities throughout the USA as they search for practical
ways to respond to unauthorized immigration. Unless these localities
are willing to let unauthorized immigrants die in the streets, they
already pay for their treatment somehow Á and usually in ways that are
unduly expensive and less efficient than in San Francisco. If local
actors are concerned about reducing disparities by legal status, the
unique San Francisco case demonstrates how creating a protective civic
environment and focusing on expanding and integrating access to
primary care can help. This gives providers greater ability to help
reduce disparities by legal status and, by extension, it allows patients
greater access to and utilization of care at a systemic level.
Acknowledgements
Support for this article was provided by a grant from the Robert Wood
Johnson Foundation. Special thanks to Irene Bloemraad, Daniel
Dohan, Alicia Fernandez, Sylvia Guendelman, Michael Jones-Correa,
´
Ming-Cheng Lo, Mary Waters, Sarah Willen, my fellow RWJ scholars
in Health Policy, the directors of the Health and Immigration Study,
and anonymous reviewers at E&RS for helpful feedback.

86 Helen B. Marrow

References
DE GRAAUW, ELS 2009 ‘Documenting the Unauthorized: Nonproft Organizations and
Intergovernmental Tensions over Municipal ID Cards in San Francisco’, Paper presented at
the Unauthorized Hispanic Migration: On the Margins of a Dream Conference, Connecticut
College, New London, CT, 16Á18 October
FOX, CYBELLE 2009 ‘A New Nativism or an American Tradition: Federal Citizenship and
Legal Status Restrictions for Medicaid and Welfare’, Paper presented at the 2009 Meeting of
the Robert Wood Johnson Foundation Scholars in Health Policy Program, Aspen,
Colorado, 27Á30 May
GOLDMAN, DANA P., SMITH, JAMES P. and SOOD, NEERAJ 2005 ‘Legal status and
health insurance among immigrants’, Health Affairs, vol. 24, no. 6, pp. 1640Á53
HORTON, SARAH 2006 ‘The double burden on safety net providers: placing health
disparities in the context of the privatization of health care in the US’, Social Science &
Medicine, vol. 63, no. 10, pp. 2702Á14
KONCZAL, LISA and VARGA, LEAH 2011 ‘Structural violence and compassionate
compatriots: immigrant health care in South Florida’, Ethnic and Racial Studies, DOI:
10.1080/01419870.2011.594169
LAMPHERE, LOUISE 2005 ‘Providers and staff respond to Medicaid managed care: the
unintended consequences of reform in New Mexico’, Medical Anthropology Quarterly, vol.
19, no. 1, pp. 3Á25
LIGHT, DONALD 2011 ‘Categorical inequality, institutional ambivalence, and permanently failing institutions: the case of immigrants and barriers to health care in America’,
Ethnic and Racial Studies, DOI: 10.1080/01419870.2011.594172
LIPSKY, MICHAEL 1980 Street-Level Bureaucracy: Dilemmas of the Individual in Public
Services, New York: Russell Sage
NEWTON, LINA and ADAMS, BRIAN E. 2009 ‘State immigration policies: innovation,
cooperation, and conflict’, Publius: The Journal of Federalism, vol. 39, no. 3, pp. 408Á31
ORTEGA, ALEXANDER N., et al. 2007 ‘Health care access, use of services, and
experiences among unauthorized Mexicans and other Latinos’, Archives of Internal
Medicine, vol. 267, no. 21, pp. 2354Á60
PEAR, ROBERT and HERSZENHORN, DAVID M. 2010 ‘Obama hails vote on health
care as answering ‘‘the call of history’’’, New York Times, 21 March
´
PORTES, ALEJANDRO, FERNANDEZ-KELLY, PATRICIA and LIGHT, DONALD W.
2011 ‘Life on the edge: immigrants confront the American health system’, Ethnic and Racial
Studies, DOI: 10.1080/01419870.2011.594173
RIDGLEY, JENNIFER 2008 ‘Cities of refuge: immigration enforcement, police, and the
insurgent genealogies of citizenship in U.S. sanctuary cities’, Urban Geography, vol. 29, no. 1,
pp. 53Á77
SAN FRANCISCO DEPARTMENT OF PUBLIC HEALTH (DPH) and SAN FRANCISCO
OFFICE OF LABOR STANDARDS ENFORCEMENT (OLSE) 2010 ‘Status Report on the
Implementation of the San Francisco Health Care Security Ordinance’, presented to the San
Francisco Board of Supervisors, San Francisco, CA, June
SCHWARTZ, KARYN and ARTIGA, SAMANTHA 2007 Health Insurance Coverage and
Access to Care for Low-Income Non-Citizen Children, Washington, DC: Kaiser Commission
on Medicaid and the Uninsured
TRAMONTE, LYNN 2009 Debunking the Myth of ‘‘Sanctuary Cities’’: Community Policing
Policies Protect American Communities, Washington, DC: Immigration Policy Center,
March

The power of local autonomy 87

HELEN B. MARROW is Assistant Professor in the Departments of
Sociology and Latin American Studies at Tufts University.
ADDRESS: Department of Sociology, Tufts University, Eaton Hall
115, 5 the Green, Medford, MA 02155, USA.
Email: [e-mail]

1
2
3
4
5
6
7
8

Exhibit E

9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

ADDENDUM TO MPAS IN SUPPORT OF AMICUS CURIAE
2:18-cv-00490-JAM-KJN

International Journal of Epidemiology, 2017, 1–11
doi: 10.1093/ije/dyw346
Original article

Original article

Change in birth outcomes among infants born
to Latina mothers after a major immigration raid
Nicole L Novak,1* Arline T Geronimus2,3 and
Aresha M Martinez-Cardoso2
1

Department of Epidemiology, University of Michigan School of Public Health, Ann Arbor, MI, USA,
Department of Health Behavior and Health Education, University of Michigan School of Public Health,
Ann Arbor, MI, USA and 3Population Studies Center, Institute for Social Research, University of
Michigan, Ann Arbor, MI, USA
2

*Corresponding author. Department of Epidemiology, University of Michigan School of Public Health, 1415 Washington
Heights, 4667 SPH I, Ann Arbor, MI 48109–2029, USA. E-mail: [e-mail]
Accepted 15 November 2016

Abstract
Background: Growing evidence indicates that immigration policy and enforcement adversely affect the well-being of Latino immigrants, but fewer studies examine ‘spillover
effects’ on USA-born Latinos. Immigration enforcement is often diffuse, covert and difficult to measure. By contrast, the federal immigration raid in Postville, Iowa, in 2008 was,
at the time, the largest single-site federal immigration raid in US history.
Methods: We employed a quasi-experimental design, examining ethnicity-specific patterns in birth outcomes before and after the Postville raid. We analysed Iowa birthcertificate data to compare risk of term and preterm low birthweight (LBW), by ethnicity
and nativity, in the 37 weeks following the raid to the same 37-week period the previous
year (n ¼ 52 344). We model risk of adverse birth outcomes using modified Poisson regression and model distribution of birthweight using quantile regression.
Results: Infants born to Latina mothers had a 24% greater risk of LBW after the raid when
compared with the same period 1 year earlier [risk ratio (95% confidence interval) ¼ 1.24
(0.98, 1.57)]. No such change was observed among infants born to non-Latina White
mothers. Increased risk of LBW was observed for USA-born and immigrant Latina mothers. The association between raid timing and LBW was stronger among term than preterm births. Changes in birthweight after the raid primarily reflected decreased birthweight below the 5th percentile of the distribution, not a shift in mean birthweight.
Conclusions: Our findings highlight the implications of racialized stressors not only for
the health of Latino immigrants, but also for USA-born co-ethnics.
Key words: immigration enforcement, birth outcomes, stress, Latinos/Hispanics, nativity

C
V The Author 2017; all rights reserved. Published by Oxford University Press on behalf of the International Epidemiological Association

1

2

International Journal of Epidemiology, 2017, Vol. 0, No. 0

Key Messages
• We compare risk of adverse birth outcomes before and after a major federal immigration raid in Postville, Iowa.
• Whereas there was no change in risk of low birthweight for infants born to White mothers in Iowa, infants born to

Latina mothers in Iowa had a 24% higher risk of low birthweight in the period following the Postville raid.
• Analyses including gestational age reveal an elevation in risk of moderate-preterm birth (PTB) after the raid among

Latina mothers, rather than an increase in very-PTB.
• These findings are consistent with theories linking immigration enforcement to the health of Latino immigrants and

their USA-born co-ethnics.

Introduction
Investigators theorize that unintended consequences of social policies affecting disadvantaged groups contribute to
entrenched US health disparities.1–5 A growing literature
examines effects of US immigration policy on immigrants
(particularly Latino immigrants),5 documenting links between immigration policy and health care utilization,6–9
Medicaid participation10,11 or food insecurity.12 A smaller,
but growing number of studies examine links between immigration enforcement and psychosocial well-being,13,14
self-rated health15 and enforcement-related distress.15–17
Many existing studies of immigration policy/enforcement and health focus on immigrant (or specifically
undocumented-immigrant) samples18 or examine Latinos
regardless of nativity.6,7,15 Whereas some studies have
documented effects of immigration policy and enforcement
on USA-born children of immigrants,9,11,19 far fewer
examine implications specifically for co-ethnic USA-born
adults.8,12,16 Although USA-born Latinos are not subject
to immigration deportation, many are embedded in communities targeted by immigration enforcement20,21 and
may experience discrimination, ‘othering’20 or chronic
identity-related vigilance22 in response to racialized exclusion.5,23–26
Measuring causal relationships between immigration
policy/enforcement and health outcomes has proved challenging: policy changes usually occur after an extended deliberation period that makes exposure classification
difficult, and enforcement practices are often diffuse and
covert.10,21 In contrast, the 2008 Immigration and
Customs Enforcement (ICE) raid on a meat-processing
plant in Postville, Iowa, was the largest single-site raid yet
seen in the USA, and occurred without warning, allowing a
clear before-and-after comparison.
We compare risk of adverse birth outcomes among
Latina and non-Latina White mothers in the state of Iowa
before and after the Postville raid. Birth outcomes, and
particularly low birthweight (LBW), are well suited as
health outcomes in this study because birth-certificate data

are publicly available and collected for all births regardless
of the mother’s immigration status, and birthweight has little measurement error. Previous studies have documented
increased LBW risk after population-level stressors such as
terrorist attacks or natural disasters27–31 (although some
studies report null or mixed findings).32,33
Economic and demographic commonalities across
Latino population clusters in Iowa (many, like Postville,
centre on meat processing) and social and affective ties between foreign-born and USA-born Latinos lead us to hypothesize effects for Latinos across the entire state. We
hypothesize that, among Iowa births, the association between LBW and birth after the raid will be modified by
mother’s ethnicity, such that foreign-born and USA-born
Latina mothers will have higher LBW rates after the raid,
whereas non-Latina White mothers will have no change in
LBW. We anticipate that this effect modification will be independent of potential socio-economic confounders and
traditional risk factors for LBW.

Exposure: the Postville raid
The ICE raid on a meat-processing plant in Postville, Iowa,
on 12 May 2008 was implemented without advance warning to local or state officials. ICE deployed 900 agents
using military tactics, including armed officers and a UH60 Black Hawk helicopter, to arrest 389 employees, 98%
of whom were Latino.34 Agents used presumed race/ethnicity to identify suspected undocumented immigrants, allegedly handcuffing all employees assumed to be Latino
until their immigration status was verified.35
Male arrestees were detained at the National Cattle
Congress in Waterloo, Iowa (80 miles from Postville),
whereas women were detained in county jails. Mothers of
small children were allowed to return to Postville with
ankle monitors but, barred from working, survived on
charitable aid.34,36 Detainees were chained together and
arraigned in groups of 10 for felony charges of aggravated
identity theft (knowingly working under a false Social
Security Number). A plea bargain led nearly all to plead

International Journal of Epidemiology, 2017, Vol. 0, No. 0

guilty, although few were technically guilty of the crime,
and 297 arrestees were deported after serving a 5-month
prison sentence.34
The raid separated hundreds of families, most often
from their primary breadwinner. Fear of follow-up home
raids kept many Postville families from staying in their
own homes, choosing instead to sleep in church pews or
leave town altogether.36 News of the raid immediately
spread throughout the state. La Prensa, a Spanishlanguage newspaper in western Iowa, published eyewitness
testimony of arrestees detained at a cattle fairground,
cuffed and chained together from the waist to the ankles.37

Methods
We obtained birth-certificate data for all births in Iowa
from 2006 to 2010 (n ¼ 209 389). We classified infants as
‘exposed’ to the post-raid environment if they were born in
the 37 weeks following the Postville raid (12 May 2008–
26 January 2009) and ‘unexposed’ if they were born in
same period one year earlier (12 May 2007–26 January
2008). We chose 37 weeks because it was the minimum
length of a normal gestation.
The primary outcome variable, LBW, was defined as
birthweight <2500 g. We used self-reported race and
Hispanic ethnicity to categorize mothers as Latina or nonLatina White, creating a ‘Latina’ category by restricting to
mothers in any Hispanic subgroup except Hispanic/
Spaniard. Latina mothers were predominantly of Mexican
descent (81%), although 11% were of Central American
origin and 8% were of other Hispanic origin. We used selfreported birthplace to categorize mothers as USA- or
foreign-born. Immigration status is not collected in birthcertificate data. We included data on maternal age (<20,
20–25, 26–30, 31–35, 36–40, 41þ years), education (<8th
grade, 9th–11th grade, high-school diploma/equivalent,
some college/Associate’s degree, college diploma), marital
status (married/unmarried at conception) and parity (first
live birth/second or higher). We also divided maternal education into tertiles within strata of ethnicity/nativity. We
included data on prenatal maternal smoking (no smoking,
<10, 10–19, 20þ cigarettes/day) and prenatal care utilization (Kessner index for adequate, intermediate and inadequate prenatal care).38
We used a data-cleaning algorithm to create a gestational age (GA) variable, which we categorized into preterm birth (PTB) and categories of GA. We took this step
to address previously reported data-quality issues for GA
estimation in vulnerable populations, including immigrant,
Latina or low-English-proficient mothers and mothers
with late prenatal care initiation.39–42 The algorithm used
a LMP-based estimate of GA wherever possible, and the

3

clinical estimate when a LMP-based estimate was unavailable or implausible for the infant’s birthweight; for more
information, see Basso and Wilcox.43 PTB was defined as
GA < 37 weeks. We further categorized GA as verypreterm (<32 weeks), moderate-preterm (32–36 weeks),
early-term (37–38 weeks) and full-term (!39 weeks).44,45
We used the cleaned GA variable to estimate the stage
of gestation at the time of the raid (or the comparison
date), classifying infants as not yet conceived, or in the
first, second or third trimester on the date of interest.
We restricted our analysis to singletons born in the 37
weeks following the raid or the same period one year earlier (n ¼ 57 850), although we include data from the same
period 2 years earlier (n ¼ 26 531) for description. We
excluded 4659 infants born to mothers who were not
Latina or non-Latina White, and those missing data on
birthweight (n ¼ 20), GA (n ¼ 115), maternal nativity
(Latina mothers only, n ¼ 11), age (n ¼ 2), education
(n ¼ 332), marital status (n ¼ 6), parity (n ¼ 256) and prenatal smoking (n ¼ 105). Excluded infants (n ¼ 847) were
more likely to be LBW, both among Latina and non-Latina
White mothers. The final sample included 52 344 infants,
25 979 born in the 37 weeks following the Postville raid
and 26 365 born during the same 37-week period 1 year
earlier.

Statistical methods
We used modified Poisson regression46 to estimate risk
ratios (RRs) comparing risk of LBW among infants born
after Postville to those in the comparison period, and used
Knol and VanderWeele’s recommended methods for presenting analyses of effect modification.47 This involved
presenting: (i) RRs for each stratum of maternal ethnicity
and birth timing with a single reference category; (ii) RRs
for being born after the raid, stratified on maternal ethnicity; and (iii) measures of effect modification on the additive scale (relative excess risk due to interaction: RERI)
and multiplicative scale (ratio of RRs).48,49 We estimated a
second set of models with Latina mothers further stratified
on nativity.
To confirm that findings were not confounded by
changes in the population of mothers, we re-estimated all
models, first with adjustments for maternal risk factors for
LBW (age, education, marital status and parity). We further adjusted for measured health behaviours that could
have mediated changes in LBW after Postville: maternal
smoking and prenatal care utilization.
We conducted additional analyses to better understand
observed changes in LBW. We stratified our initial models
on PTB to evaluate whether shifts in LBW were operating
primarily on term or preterm births. We used conditional

4

quantile regression50,51 to analyse the distribution of birthweight among Latinas by exposure period. Quantile regression models the association of the exposure with the
full range of the birthweight distribution, not merely above
or below the set cut-off of 2500 g for LBW.51,52 Quantiles
were specified to evaluate changes in birthweights lower
than the 2500-g cut-off as well as changes throughout the
full distribution of birthweight: the 2nd, 5th, 10th, 25th,
50th, 75th, 90th, 95th and 98th percentiles were considered. We estimated the association of birth post-raid
with the distribution of birthweight, bootstrapping results
1000 times to estimate standard errors and confidence
intervals.
We also examined changes in categories of GA (verypreterm, moderate-preterm, early-term, full-term) before
and after the raid, among Latina mothers, using multinomial logistic regression with robust standard errors.
To examine differences in LBW risk according to stage
of pregnancy at the time of the raid, we repeated initial
analyses with the sample further stratified by stage of gestation at the time of the raid.30,53,54 We estimated RRs for
LBW after the raid compared with before the raid, by ethnicity and gestational category. To determine whether
changes in risk of LBW varied by social position within
strata of ethnicity/nativity, we also estimated LBW models
stratified on within-group tertiles of education. Analyses
were conducted with STATA 13.

Results
Traditional risk factors for LBW and PTB varied by maternal ethnicity and nativity; however, within ethnicity/nativity groups, the distribution of maternal socio-demographic
characteristics remained consistent before and after the
raid, as did mean birthweight (Table 1). Prior to the raid,
Latina and White mothers had similar prevalence of LBW
(4.7% for both) and PTB (7.5% for both), which is consistent with other reports of ethnicity-specific birth outcomes
in Iowa in this period.55,56
Figure 1 displays temporal trends in LBW, by ethnicity
and nativity, including the two time periods in the study
sample and also extending an additional year earlier (12
May 2006–26 January 2007). Among White mothers, rates
of LBW declined slightly (as has been the trend nationwide
since 2006).57 Among Latina mothers, rates of LBW were
stable in 2006–07, but rose among USA- and foreign-born
Latina mothers after the raid.
As displayed in Table 2, the RRs [95% confidence intervals (CIs)] comparing risk of LBW after the raid to before
the raid were 1.24 (0.98–1.57) among Latina mothers and
0.95 (0.87–1.03) among White mothers. The measure of
effect modification on the additive scale, the RERI, was

International Journal of Epidemiology, 2017, Vol. 0, No. 0

0.30 (95% CI 0.03–0.57) and the measure of effect modification on the multiplicative scale, the ratio of RRs, was
1.31 (1.02–1.68). RRs and effect modification measures
were robust to adjustment for potential confounders and
mediators (Supplementary Tables 1 and 2, available as
Supplementary data at IJE online).
Table 3 displays the same models with Latina mothers
further stratified by nativity. Although confidence intervals
widen because of the smaller sample in each group, the
RRs (95% CIs) for LBW after the raid among foreign-born
(1.25, 0.93–1.67) and USA-born Latina mothers (1.22,
0.83–1.81) were similar in magnitude to the RR from the
pooled model, as were the effect modification measures.
Adjustment for potential confounders or mediators did not
affect these findings (Supplementary Tables 3 and 4, available as Supplementary data at IJE online).
In models stratified by term/preterm births, LBW was
more strongly associated with birth post-raid among term
infants born to Latina mothers (RR, 95% CI ¼ 1.49,
0.95–2.33) than among preterm infants (1.08, 0.88–1.33)
(results not shown).
Multinomial logistic regression comparing categories of
GA, by ethnicity, before and after Postville, reveals an elevation in risk of moderate-PTB after the raid among Latina
mothers (relative risk ratio, 95% CI ¼ 1.11, 0.89–1.38),
but no change in risk of very-PTB (0.81, 0.46–1.41)
(Table 4).
Quantile regression on the distribution of birthweight
indicated that, among Latina mothers, birth post-raid was
associated with reduced birthweight only at the left tail of
the birthweight distribution, where infants below the 5th
percentile of birthweight (corresponding to 2518 g before
the raid) were 88 g lighter after the raid (95% CI –168g to
–8g). Birth post-raid was not associated with differences in
birthweight among infants below the 2nd percentile of
birthweight, which corresponded to 2084 g before the raid
(6 g heavier, 95% CI –224g to –236g) or at any other point
in the birthweight distribution.
In models examining risk of LBW among Latina mothers
stratified by stage of gestation at the time of the raid, we
found the strongest association between LBW and birth
post-raid among mothers in the first trimester at the time of
the raid (RR, 95% CI ¼ 1.39, 0.97–1.98) (Table 5). In
LBW models stratified by within-group tertiles of education,
we observed the strongest association between LBW and
birth post-raid in the lower two tertiles of education for
both immigrant and USA-born Latina mothers (Table 6).

Discussion
We used the Postville raid, a large-scale immigration raid implemented without warning, as a natural experiment to

International Journal of Epidemiology, 2017, Vol. 0, No. 0

5

Table 1. Descriptive statistics by mother’s ethnicity/nativity, during the 37 weeks following the Postville raid (12 May 2008–26
January 2009) and during the same time period 1 year earlier (12 May 2007–26 January 2008) (n ¼ 52 344)
White
Time period

Infant sex
Male
Female
Mother’s age (years)
<20
20–25
26–30
31–35
36–40
41þ
Education
8th grade or less
Some high-school education
High-school diploma/equivalent
Some college
College degreeþ
Maternal marital status
Mother unmarried
Mother married
Parity
0 previous
1 or more previous
Prenatal care (Kessner Index)
Inadequate
Intermediate
Adequate
Smoking in pregnancy
No smoking
<10 cigarettes/day
10–19 cigarettes/day
20þ cigarettes/day
Low birthweight
Mean birthweight (g)
Preterm birth

Before raid
(n ¼ 23 878)
Mean %

Foreign-born Latina
After raid
(n ¼ 23 379)
Mean %

USA-born Latina

Before raid
(n ¼ 1689)
Mean %

Before raid
(n ¼ 798)
Mean %

After raid
(n ¼ 1746)
Mean %

After raid
(n ¼ 854)
Mean %

51.3
48.7

51.5
48.5

50.4
49.6

50.2
49.7

50.4
49.6

48.7
51.3

7.7
32.5
33.2
18.6
6.9
1.1

7.8
31.1
34.1
18.9
6.9
1.1

9.9
33.3
27.5
19.7
8.1
1.5

9.3
33.1
27.7
21.0
7.9
1.0

27.3
38.5
22.1
8.9
2.9
0.4

27.3
39.0
20.3
9.4
3.2
0.9

1.3
8.2
21.4
38.5
30.7

1.2
8.2
20.4
38.8
31.4

34.2
35.5
18.1
8.2
4.1

33.7
36.2
18.4
7.8
3.9

2.6
33.6
31.1
26.2
6.5

2.9
32.4
30.8
26.2
7.6

31.3
68.7

31.7
68.3

44.9
55.1

47.0
53.0

64.5
35.5

56.8
43.2

40.0
60.0

39.8
60.2

28.4
71.5

27.8
72.2

41.1
59.9

40.4
59.6

2.2
7.7
90.1

2.2
7.7
90.1

5.0
19.2
75.7

4.9
17.5
77.6

4.8
13.4
81.8

3.9
15.2
80.9

80.0
10.3
7.4
2.4
4.7
3401
7.5

80.7
10.2
6.9
2.2
4.4
3407
7.5

98.7
1.1
0.2
0.1
4.5
3349
7.8

99.1
0.9
0.1
—
5.6
3339
8.0

82.0
12.7
3.6
1.8
5.3
3336
7.5

87.2
9.1
3.0
0.6
6.4
3315
8.9

investigate the effects of immigration enforcement on birth
outcomes among Latina mothers in a Mid-western state. We
found that rates of LBW were steady among White and
Latina mothers in the 2 years preceding the raid, but that
rates of LBW rose only among Latina mothers after the raid.
The association between birth post-raid and LBW was modified by maternal ethnicity on both the additive (RERI > 0)
and multiplicative scales (ratio of RRs > 1). This association
was evident among both foreign-born and USA-born Latina
mothers and persisted after adjustment for maternal risk factors, maternal smoking and prenatal care utilization.
We found that the increases in LBW were greatest
among term births, but that there was also a higher

prevalence of moderate-preterm (not very-preterm) infants
after the raid. Previous studies of psychosocial stressors
and birth outcomes have found that LBW increased both
through increased PTB30,53 and through intrauterine
growth restriction58,59; it appears that both mechanisms
operated in Iowa, which is plausible given the diversity of
economic and psychosocial pathways by which the raid
may have affected Latina mothers.
Comparing births after Postville to births in the same
period 1 year earlier accounted for seasonality in birth outcomes and avoided the ‘mechanical correlation’ between
pregnancy duration and risk of exposure to stressful
events—a methodological pitfall in studies using the time

6

International Journal of Epidemiology, 2017, Vol. 0, No. 0

period immediately preceding an event as the comparison
period.33
Our examination of births in the entire state of Iowa
makes this a conservative analysis as one might hypothesize stronger effects in or near Postville compared with
Latino communities farther away. Previous studies have
examined dose–response relationships based on geographic
distance from a natural disaster or attack28,33 but, for priv-

Figure 1. Descriptive graph: rates of low birthweight (LBW) in the 37
weeks following the Postville raid compared with the same time period
1 and 2 years earlier.

acy reasons, the Iowa Department of Public Health does
not release microdata with date of birth and geographic information. However, we do not have reason to believe that
stressors resulting from Postville raid would emanate by
geographic distance in as dramatic a way as a geographically confined natural disaster or attack. Many Latino communities in Iowa are economically similar to Postville, and
communication networks between communities make it
plausible that Latinos across the state would feel connected
to an enforcement event targeted at a single workplace.
Lauderdale’s finding of increased LBW among Arabicnamed mothers in California after the attacks of 11
September 2001 that occurred across the country in New
York City lends plausibility to the view that social identity
threats can affect co-ethnics at remote distances.
LBW risk increased most among Latina mothers with
lower educational attainment (less than high school for the
foreign-born and less than college for the USA-born). This
could be because low-educated mothers were more vulnerable to the economic and psychosocial fallout of the raid
or had fewer coping resources. Reports from throughout
Iowa after the Postville raid include evidence of individuals
and families preparing for the possibility of further immigration enforcement,37,60–62 avoiding public space,60,63 restricting spending,63,64 losing income or economic security

Table 2. Modified Poisson regression results for risk of LBW by time period of birth (before/after Postville raid) and mother’s ethnicity (White/Latina) (n ¼ 52 344)
Before raid
N LBW/non
White mother
Latina mother

After raid

RR (95% CI); P

N LBW/non

RR (95% CI); P

RR (95% CI); P for
after raid vs before within
strata of ethnicity

1112/22 766
118/2369

1.0 (Reference)
1.02 (0.84, 1.23); P ¼0.84

1029/22 350
153/2447

0.95 (0.87, 1.03); P ¼ 0.18
1.31 (1.02, 1.68); P ¼ 0.03

0.95 (0.87, 1.03); P ¼ 0.18
1.24 (0.98, 1.57); P ¼ 0.07

Measure of effect modification on additive scale: RERI (95% CI) ¼ 0.30 (0.03, 0.57); P ¼ 0.03.
Measure of effect modification on multiplicative scale: ratio of RRs (95% CI): 1.31 (1.02, 1.68); P ¼ 0.03. RRs and measures of effect modification are
unadjusted.

Table 3. Modified Poisson regression results for risk of LBW by time period of birth (before/after Postville raid) and mother’s ethnicity/nativity (White/foreign-born Latina/USA-born Latina) (n ¼ 52 344)
Before raid
N BW/non

White mother
Foreign-born
Latina mother
USA-born
Latina mother

After raid

RR (95% CI); P

N LBW/non

RR (95% CI); P

RR (95% CI); P for
after raid vs before
within strata of
ethnicity

1112/22 766
76/1613

1.0 (Reference)
0.97 (0.77, 1.21); P ¼ 0.77

1029/22 350
98/1648

0.95 (0.87, 1.03); P ¼ 0.18
1.32 (0.97, 1.79); P ¼ 0.07

0.95 (0.87, 1.03); P ¼ 0.18
1.25 (0.93, 1.67); P ¼ 0.14

42/756

1.13 (0.84, 1.53); P ¼ 0.42

55/799

1.29 (0.87, 1.93); P ¼ 0.20

1.22 (0.83, 1.81); P ¼ 0.31

Measure of effect modification on additive scale: RERI (95% CI) ¼ 0.29 (–0.03, 0.62); P ¼ 0.07 (for Foreign-born Latinas), 0.31 (–0.18, 0.80); P ¼ 0.22 (for
USA-born Latinas).
Measure of effect modification on multiplicative scale: ratio of RRs (95% CI) ¼ 1.32 (0.97, 1.79); P ¼ 0.07 (for Foreign-born Latinas), 1.29 (0.87, 1.93); P ¼
0.20 (for USA-born Latinas). RRs and measures of effect modification are unadjusted.

International Journal of Epidemiology, 2017, Vol. 0, No. 0

7

Table 4. Multinomial logistic regression results, categories of gestational age at birth in 37 weeks following the Postville raid
compared with same period 1 year earlier, by maternal ethnicity (n ¼ 52 344)
Gestational age at birth

White (n ¼ 47 907) RR (95% CI); P

Full-term (reference) (39þ weeks)
Early-term (37–38 weeks)
Moderate-preterm (32–36 weeks)
Very-preterm (<32 weeks)

Latina (n ¼ 5149) RR (95% CI); P

1.00
0.91 (0.87, 0.95); P <0.01
0.97 (0.90, 1.04); P ¼0.39
0.96 (0.79, 1.17); P ¼0.70

1.00
0.92 (0.81, 1.04); P ¼0.20
1.11 (0.89, 1.38); P ¼0.34
0.81 (0.46, 1.41); P ¼0.45

Table 5. Modified Poisson regression results for risk of LBW by time period of birth (before/after Postville raid) and mother’s ethnicity (White/Latina), stratified by stage of gestation at time of Postville raid (n ¼ 52 344)
White mothers

Latina mothers

Measures of effect modification

Stage of gestation
at time of
Postville raid

RR (95% CI); P

RR (95% CI); P

Ratio of RRs (95% CI); P

RERI (95% CI); P

Not yet conceived
(n ¼ 359)
First trimester
(n ¼ 14 302)
Second trimester
(n ¼ 23 355)
Third trimester
(n ¼ 14 687)

1.12 (0.93, 1.35); P ¼0.24

1.03 (0.65, 1.63); P ¼0.89

0.92 (0.56, 1.52); P ¼0.75

–0.08 (–0.70, 0.54); P ¼ 0.81

0.89 (0.78, 1.02); P ¼0.10

1.39 (0.97, 1.98); P ¼0.07

1.55 (1.06, 2.27); P ¼0.02

0.49 (0.07, 0.92); P ¼ 0.02

0.98 (0.87, 1.11); P ¼0.75

1.12 (0.79, 1.59); P ¼0.54

1.14 (0.79, 1.65); P ¼0.49

0.14 (–0.26, 0.55); P ¼0.49

1.00 (0.81, 1.23); P ¼0.98

1.12 (0.58, 2.13); P ¼0.74

1.12 (0.58, 2.16); P ¼0.74

0.12 (–0.58, 0.81); P ¼ 0.74

Table 6. Modified Poisson regression results for risk of LBW by time period of birth (before/after Postville raid) and mother’s ethnicity and nativity, stratified by approximate within-group tertiles of education (n ¼ 52 344)
Non-Latina White mothers

Foreign-born Latina mothers

USA-born Latina mothers

(n ¼ 47 257)

(n ¼ 3435)

(n ¼ 1652)

Within-group tertiles
of education (n)

RR for birth after
Postville raid

Within-group tertiles
of education (n)

RR for birth after
Postville raid

Within-group tertiles
of education (n)

RR for birth after
Postville raid

High-school diploma
(n ¼ 14 334)
Associate’s degree or
some college
(n ¼ 18 261)
College diploma or
higher (n ¼ 14 662)

0.95 (0.83, 1.08)

Less than 8th
grade (n ¼ 1165)
Some high school
(n ¼ 1232)

1.42 (0.90, 2.25)

Less than high
school (n ¼ 591)
High-school
diploma (n ¼ 511)

1.18 (0.68, 2.06)

High school or greater
(n ¼ 1038)

0.89 (0.51, 1.57)

Some college or
greater (n ¼ 550)

0.90 (0.32, 2.54)

0.98 (0.86, 1.12)

0.91 (0.76, 1.09)

due to changing employment practices62,65–68 and experiencing increased discrimination, stereotype threat or
racialized exclusion as public discourse frequently conflated Latino/Hispanic phenotype with undocumented status.35,65 These reports align with findings from a recent
quasi-experimental study in Michigan: after a local immigration raid led to several arrests and deportations, Latinos
were more likely to report that they feared the consequences of deportation, and that their immigration status
impeded social relationships.15 In the wake of the Postville

1.40 (0.84, 2.35)

1.48 (0.78, 2.83)

raid, similar restrictions in social support and increases in
day-to-day fear may have coalesced to increase psychosocial stress and reduce coping resources among Latino immigrants and USA-born co-ethnics. Following the raid,
Latino Iowans likened the experience to a flood or earthquake, reflecting the profound impact of this stressor on
their lives and on their health.69,70
Quantile regression indicated that the higher risk of
LBW among Latina mothers after Postville resulted from
decreased birthweight at the left tail of the distribution,

8

not a shift in mean birthweight. This is similar to
Lauderdale’s findings for ethnicity-specific change in birthweight after 9/11.71 Lower birthweights at the left tail of
the birthweight distribution are more likely to be associated with infant mortality than a leftward shift of the entire distribution.71,72
Post-raid increases in LBW risk were greatest for Latina
mothers in the first trimester of gestation at the time of the
raid. Several other studies have also found stronger effects
among first-trimester exposures,27,29,30,73 but others have
not.31,33,53,54 Our finding in Iowa could suggest that earlygestation infants were more vulnerable, or it may be that
those infants were simply exposed to the post-raid environment for a larger proportion of gestation.
Several complex immune, inflammatory and endocrine
pathways are proposed to link psychosocial stressors and
birthweight. One hypothesis is that maternal psychosocial
stress disrupts the balance between maternal glucocorticoid levels and 11 beta-hydroxysteroid dehydrogenase type
2 (HSD2), an enzyme that metabolizes cortisol into inactive cortisone.74,75 Placental HSD2 typically up-regulates
in tandem with serum glucocorticoid levels during gestation, protecting the fetus from 80–90% of circulating maternal glucocorticoids.76 However, psychosocial stress and
disruption of maternal emotional support have been linked
to both higher prenatal glucocorticoid concentrations and
lower placental HSD2,75,77–79 both of which are linked to
LBW.76,80–84 The psychosocial, economic, communal and
identity-based stressors activated by the Postville raid may
have interfered with Latina mothers’ neuroendocrine balance and coping resources, leaving infants vulnerable to a
dysregulated endocrine environment.

Limitations
If healthy pregnant Latinas left Iowa after Postville, increased
LBW among Latinas might reflect selection effects. However,
analysing Census data,85 we found no evidence that the raid
was associated with a change in the size of Iowa’s Latino
population, overall or among women of reproductive age.
And, as noted, we found no difference in demographic characteristics among Latina mothers before or after the raid.
There is random year-to-year variation in LBW prevalence, especially in small populations, which raises concerns that the observed increase in LBW among Latina
mothers was a chance finding. We do not have access to
birth microdata pre-2006, but we used publicly available
data56 to calculate crude LBW prevalences among singleton births to non-Hispanic White and Hispanic mothers
during May–January for the 5 years preceding and following the raid. LBW prevalence among infants born to
Hispanic mothers is higher from May 2008 to January

International Journal of Epidemiology, 2017, Vol. 0, No. 0

2009 than in any other May–January period from 2003 to
2013 (Supplementary Figure 1, available as Supplementary
data at IJE online).
Birth-certificate data for birthweight, maternal ethnicity
and maternal birthplace have high validity relative to medical records.86–89 However, data on GA are of lower quality,86 particularly for Latina and non-English-proficient
mothers,39,41 which affects our classification of PTB and
stage of gestation at the time of the raid. We used a datacleaning algorithm to mitigate data-quality issues, but this
entails several assumptions90 and, whereas it is unlikely
that remaining misclassification of GA differs by raid timing, even non-differential misclassification may have
biased findings for PTB and stage of gestation towards the
null. Birth-certificate data for our hypothesized mediators,
prenatal smoking and prenatal care are also of lower quality,86–88 which reduces our ability to observe mediation by
changed health behaviours.

Conclusions
Our findings are consistent with studies observing changes
in adverse birth outcomes after a major population-level
stressor28,30,53,71 and contribute to literature on racialized
stressors and ethnicity-specific birth outcomes.32,71 We extend the literature on immigration policy/enforcement and
health by specifically examining a physical outcome and
by examining both immigrant and USA-born Latinos.
The Postville raid was an extreme example of diffuse
and pervasive racialized economic and psychosocial stressors that Latinos face throughout the USA. The scale and
temporality of this event created conditions that lend insight into the pervasive effects of these stressors, which are
often difficult to measure. Exclusive immigration policies
and their militarized enforcement exacerbate the racialized
exclusion of Latinos in the USA, which may contribute to a
cumulative health burden for immigrant and USA-born
Latinos alike.

Supplementary Data
Supplementary data are available at IJE online.

Funding
This work was supported in by the National Institutes of Health
(grant numbers T32 AG000221, T32 HD007339), the Rackham
Graduate School at the University of Michigan (grant number
U028717–185539) and the Center for Advanced Study in the
Behavioral Sciences at Stanford. The funding sources had no involvement in study design; in the collection, analysis and interpretation of data; in the drafting of the manuscript; or in the decision to
submit it for publication. The study protocol was approved by the

International Journal of Epidemiology, 2017, Vol. 0, No. 0
Institutional Review Board of the Iowa Department of Public
Health (Research Agreement 2718) and determined to be ‘not regulated’ by the Health Sciences and Behavioral Sciences Institutional
Review Board of the University of Michigan (HUM00074761).
Conflict of interest: The authors have no conflicts of interest to
declare.

References
1. Richman LS, Hatzenbuehler ML. A multilevel analysis of stigma
and health: implications for research and policy. Policy Insights
Behav Brain Sci 2014;1:213–21.
2. Geronimus AT, Pearson JA, Linnenbringer E et al. Race-ethnicity,
poverty, urban stressors, and telomere length in a Detroit
community-based sample. J Health Soc Behav 2015;56:199–224.
3. Pearson JA. Can’t buy me whiteness. Du Bois Rev Soc Sci Res
Race 2008;5:27–47.
4. Keene DE, Padilla MB. Race, class and the stigma of place:
moving to ‘opportunity’ in Eastern Iowa. Health Place
2010;16:1216–23.
5. Viruell-Fuentes EA, Miranda PY, Abdulrahim S. More than culture: structural racism, intersectionality theory, and immigrant
health. Soc Sci Med 2012;75:2099–106.
6. Beniflah JD, Little WK, Simon HK et al. Effects of immigration
enforcement legislation on Hispanic pediatric patient visits to
the pediatric emergency department. Clin Pediatr (Phila)
2013;52(12):1122–126.
7. Rhodes SD, Mann L, Siman FM et al. The impact of local immigration enforcement policies on the health of immigrant
Hispanics/Latinos in the United States. Am J Public Health
2015;105:329–37.
8. Toomey RB, Umana-Taylor AJ, Williams DR et al. Impact of
Arizona’s SB 1070 immigration law on utilization of health care
and public assistance among Mexican-origin adolescent
mothers and their mother figures. Am J Public Health
2014;104(Suppl 1):S28–34.
9. White K, Yeager VA, Menachemi N et al. Impact of Alabama’s
immigration law on access to health care among Latina immigrants and children: implications for national reform. Am J
Public Health 2014;104:397–405.
10. Vargas ED. Immigration enforcement and mixed-status families:
the effects of risk of deportation on Medicaid use. Child Youth
Serv Rev 2015;57:83–9.
11. Watson T. Inside the refrigerator: immigration enforcement and
chilling effects in Medicaid participation. Am Econ J Econ
Policy 2014;6:313–38.
12. Potochnick S, Chen, J-H Perreira K. Local-level immigration enforcement and food insecurity risk among Hispanic immigrant
families with children: national-level evidence. J Immigr Minor
Health 2016; doi:10.1007/s10903-016-0464-5.
13. Hacker K, Chu J, Leung C et al. The impact of immigration and
customs enforcement on immigrant health: perceptions of immigrants in Everett, Massachusetts, USA. Soc Sci Med 2011;73:
586–94.
14. Hacker K, Chu J, Arsenault L et al. Provider’s perspectives on
the impact of Immigration and Customs Enforcement (ICE) activity on immigrant health. J Health Care Poor Underserved
2012;23:651.

9
15. Lopez WD, Kruger DJ, Delva J et al. Health implications of an
immigration raid: findings from a Latino community in the
midwestern United States. J Immigr Minor Health 2016;
doi:10.1007/s10903-016-0390-6.
16. Quiroga SS, Medina DM, Glick J. In the belly of the beast: effects
of anti-immigration policy on Latino community members. Am
Behav Sci 2014;58:1723–42.
17. Sabo S, Shaw S, Ingram M et al. Everyday violence, structural racism and mistreatment at the US-Mexico border. Soc Sci Med.
2014;109:66–74.
18. Martinez O, Wu E, Sandfort T et al. Evaluating the impact of immigration policies on health status among undocumented immigrants: a systematic review. J Immigr Minor Health 2013;17:
947–70.
19. Dreby J. U.S. immigration policy and family separation: the
consequences for children’s well-being. Soc Sci Med
2015;132:245–51.
20. Viruell-Fuentes EA. Beyond acculturation: immigration, discrimination, and health research among Mexicans in the United
States. Soc Sci Med 2007;65:1524–35.
21. Maldonado MM. Latino incorporation and racialized border
politics in the heartland: interior enforcement and policeability
in an English-only state. Am Behav Sci 2014;58:1927–45.
22. Geronimus AT, James SA, Destin M et al. Jedi public health: cocreating an identity-safe culture to promote health equity.
SSM—Popul Health 2016;2:105–16.
23. Cobas JA, Duany J, Feagin JR. How the United States Racializes
Latinos: White Hegemony and Its Consequences. Routledge,
2009.
24. Golash-Boza T. Dropping the hyphen? Becoming Latino (a)American through racialized assimilation. Soc Forces
2006;85:27–55.
25. Frank R, Akresh IR, Lu B. Latino immigrants and the U.S. racial
order: how and where do they fit in? Am Sociol Rev
2010;75:378–401.
26. Arellano-Morales L, Roesch SC et al. Prevalence and correlates
of perceived ethnic discrimination in the Hispanic Community
Health Study/Study of Latinos Sociocultural Ancillary Study.
J Lat Psychol 2015;3:160–76.
27. Mansour H, Rees DI. Armed conflict and birth weight: evidence
from the al-Aqsa Intifada. J Dev Econ 2012;99:190–9.
28. Eskenazi B, Marks AR, Catalano R et al. Low birthweight in
New York City and upstate New York following the events of
September 11th. Hum Reprod 2007;22:3013–20.
29. Camacho A. Stress and birth weight: evidence from terrorist attacks. Am Econ Rev 2008;98(2):511–15.
30. Torche F. The effect of maternal stress on birth outcomes: exploiting a natural experiment. Demography 2011;48:1473–91.
31. Simeonova E. Out of sight, out of mind? Natural disasters and
pregnancy outcomes in the USA. CESifo Econ Stud
2011;57:403–31.
32. El-Sayed A, Hadley C, Galea S. Birth outcomes among Arab
Americans in Michigan before and after the terrorist attacks of
September 11, 2001. Ethn Dis 2008;18:348–56.
33. Currie J, Rossin-Slater M. Weathering the storm: hurricanes and
birth outcomes. J Health Econ 2013;32:487–503.
34. Rigg RR. The Postville raid: a postmortem. Rutgers Race Rev
2011;12:271.

10
35. Krogstad JM. In Postville, shock but no surprise. WaterlooCedar Falls Courier, 13 May 2008 (cited 21 February
2015). http://wcfcourier.com/news/metro/inpostvilleshockbut
nosurprise/article_205edba8a0e159bea16b13ceae6628d6.html.
36. Camayd-Freixas E. Interpreting after the largest ICE raid in US
history: a personal account. New York Times, 2008 (cited 1
December 2012). http://graphics8.nytimes.com/images/2008/07/
14/opinion/14ed-camayd.pdf.
37. Lopez L. No somos animales. La Prensa, 16 May 2008 (cited 21
February 2015). http://www.laprensaiowa.com/5-16/local.html.
38. Kessner D, Singer J, Kalk C, Shlesinger E. Infant Death: An
Analysis by Maternal Risk and Health Care. Institute of
Medicine and National Academy of Sciences, 1973.
39. Dietz PM, England LJ, Callaghan WM et al. A comparison of
LMP-based and ultrasound-based estimates of gestational age
using linked California livebirth and prenatal screening records.
Paediatr Perinat Epidemiol 2007;21 (Suppl 2):62–71.
40. Bengiamin MI, Capitman JA, Ruwe MB. Disparities in initiation and adherence to prenatal care: impact of insurance,
race-ethnicity and nativity. Matern Child Health J 2010;14:
618–24.
41. Wingate MS, Alexander GR, Buekens P et al. Comparison of gestational age classifications: date of last menstrual period vs. clinical estimate. Ann Epidemiol 2007;17:425–30.
42. Reichman NE, Schwartz-Soicher O. Accuracy of birth certificate
data by risk factors and outcomes: analysis of data from New
Jersey. Am J Obstet Gynecol 2007;197:32.e1–32.e8.
43. Basso O, Wilcox A. Mortality risk among preterm babies: immaturity versus underlying pathology. Epidemiology 2010;21:
521–7.
44. Engle WA. A recommendation for the definition of ‘late preterm’
(near-term) and the birth weight–gestational age classification
system. Optim Care Outcomes Late Preterm–Termin Part 1
2006;30:2–7.
45. Gynecologists AC of O and ACOG Committee Opinion No 579:
definition of term pregnancy. Obstet Gynecol 2013;122:
1139–40.
46. Cummings P. Methods for estimating adjusted risk ratios. Stata J
2009;9:175.
47. Knol MJ, VanderWeele TJ. Recommendations for presenting
analyses of effect modification and interaction. Int J Epidemiol
2012;41:514–20.
48. Altman DG, Bland JM. Interaction revisited: the difference between two estimates. BMJ 2003;326:219.
49. Richardson DB, Kaufman JS. Estimation of the relative excess
risk due to interaction and associated confidence bounds. Am J
Epidemiol 2009;169:756–60.
50. Chernozhukov V, Fernndez-Val I. Inference for extremal condia
tional quantile models, with an application to market and birthweight risks. Rev Econ Stud 2011;78:559–89.
51. Abrevaya J, Dahl CM. The effects of birth inputs on birthweight:
evidence from quantile estimation on panel data. J Bus Econ Stat
2008;26:379–97.
52. Liu SY, Kawachi I, Glymour MM. Education and inequalities in
risk scores for coronary heart disease and body mass index: evidence for a population strategy. Epidemiology 2012;23:657–64.
53. Carlson K. Fear itself: The effects of distressing economic news
on birth outcomes. J Health Econ 2015;41:117–32.

International Journal of Epidemiology, 2017, Vol. 0, No. 0
54. Class QA, Lichtenstein P, Langstrom N et al. Timing of prenatal
maternal exposure to severe life events and adverse pregnancy
outcomes: a population study of 2.6 million pregnancies.
Psychosom Med 2011;73:234–41.
55. France J, Johnson D, Jungling J. Vital Statistics of Iowa. Center
for Health Statistics, Iowa Department of Public Health, 2006.
http://idph.iowa.gov/Portals/1/Files/HealthStatistics/vital_stats_
2006.pdf.
56. Natality public-use data 2003–2006, on CDC Wonder Online
Database, March 2009. United States Department of Health and
Human Services (US DHHS), Centers for Disease Control and
Prevention (CDC), National Center for Health Statistics
(NCHS), Division of Vital Statistics (cited 7 July 2016. http://
wonder.cdc.gov/natality-v2006.html.
57. Martin JA, Hamilton BE, Osterman MJ et al. Births: final data
for 2013. Natl Vital Stat Rep Cent Dis Control Prev Natl Cent
Health Stat Natl Vital Stat Syst 2015;64:1–65.
58. Margerison-Zilko CE, Catalano R, Hubbard A et al. Maternal
exposure to unexpected economic contraction and birth weight
for gestational age. Epidemiology 2011;22:855–8.

59. Eir ottir VH, Asgeirsd ttir TL, Bjarnadttir RI et al. Low
ıksd
o
o
birth weight, small for gestational age and preterm births before
and after the economic collapse in Iceland: a population based
cohort study. PLoS ONE 2013;8:e80499.
60. Jacobs J, Perkins J. Immigration raid: workers take care, take
cover. Des Moines Register, 12 May 2008 (cited 21 February
2015). http://archive.desmoinesregister.com/article/20080513/
NEWS/805130402/.
61. Christensen E. Area Hispanics scurry to get paperwork in line.
Waterloo-Cedar Falls Courier, 12 May 2008 (cited 21 February
2015). http://wcfcourier.com/news/top_story/areahispanicsscur
rytogetpaperworkinline/article_805fc83ff238527e9aae638f
cedbd414.html.
62. Toopes C. Diversity group discusses immigration rumors.
Ottumwa Courier, 29 May 2008 (cited 21 February 2015).
http://www.ottumwacourier.com/news/local_news/diversity
groupdiscussesimmigrationrumors/article_21109416a790557
ba7a40b749c549399.html.
63. 90 miles away, Postville raid’s impact is still being felt. Mason
City Globe Gazette, 25 May 2008 (cited 21 February 2015).
http://globegazette.com/news/local/miles-away-postville-raid-simpact-is-still-being-felt/article_ed576161–6c5c-5543–8d9bac0857480bd3.html.
64. Burns D. Postville raid raises worries in Denison Latino community, activist says. Iowa Independent, 23 May 2008 (cited 21
February 2015). http://iowapoliticalalert.blogspot.com/2008/
05/postvilleraidraisesworriesin.html.
65. Saul T. Eastern Iowa Hispanic community fears more raids.
Quad-City Times, 22 May 2008 (cited 21 February 2015).
http://siouxcityjournal.com/news/eastern-iowa-hispanic-commu
nity-fears-more-raids/article_20a51fb3-24e0-55be-b65b-98b82
089b5ae.html.
66. Larsen D. SL does not foresee immigration raid, but concerns
continue. Storm Lake Pilot Tribune, 28 August 2008 (cited 21
February 2015). http://www.stormlakepilottribune.com/story/
1518320.html.
67. Jackson HC. Town wonders if it’s next to face immigration raid.
USA Today, 26 August 2008 (cited 21 February 2015). http://usa

International Journal of Epidemiology, 2017, Vol. 0, No. 0
today30.usatoday.com/news/nation/200808261720516720_x.
68.
69.

70.
71.

htm.
Jones M. Postville, Iowa, is up for grabs. New York Times
Magazine, July 15, 2012.
Gibbs VG, Hernndez L. Shattered Dreams: The Story of a
a
Historic ICE Raid in the Words of the Detainees. Floricanto
Press, 2014.
Brosnan G. The Postville immigration raid: like a great flood.
Rev Harv Rev Lat Am 2011;Fall 2010/Winter 2011:70–1.
Lauderdale DS. Birth outcomes for Arabic-named women in

California before and after September 11. Demography
2006;43:185–201.
72. Wilcox AJ. On the importance—and the unimportance—of
birthweight. Int J Epidemiol 2001;30:1233–41.
73. Zhu P, Tao F, Hao J et al. Prenatal life events stress: implications
for preterm birth and infant birthweight. Am J Obstet Gynecol
2010;203:34.e1–34.e8.
74. Reynolds RM. Glucocorticoid excess and the developmental origins of disease: two decades of testing the hypothesis—2012
Curt Richter Award Winner. Psychoneuroendocrinology
2013;38:1–11.
75. O’Donnell KJ, Bugge Jensen A, Freeman L et al. Maternal prenatal anxiety and downregulation of placental 11b-HSD2.
Psychoneuroendocrinology 2012;37:818–26.
76. Benediktsson R, Calder AA, Edwards CR et al. Placental 11
beta-hydroxysteroid dehydrogenase: a key regulator of fetal
glucocorticoid exposure. Clin Endocrinol 1997;46:161–6.
77. Reynolds RM, Pesonen A-K, O’Reilly JR et al. Maternal depressive symptoms throughout pregnancy are associated with
increased placental glucocorticoid sensitivity. Psychol Med
2015;45:2023–30.
78. La Marca-Ghaemmaghami P, La Marca R, Dainese SM et al.
The association between perceived emotional support, maternal
mood, salivary cortisol, salivary cortisone, and the ratio between
the two compounds in response to acute stress in second trimester pregnant women. J Psychosom Res 2013;75:314–20.

11
79. Graignic-Philippe R, Dayan J, Chokron S et al. Effects of prenatal stress on fetal and child development: a critical literature
review. Neurosci Biobehav Rev 2014;43:137–62.
80. Bolten MI, Wurmser H, Buske-Kirschbaum A et al. Cortisol levels in pregnancy as a psychobiological predictor for birth weight.
Arch Womens Ment Health 2011;14:33–41.
81. Guardino CM, Schetter CD, Saxbe DE et al. Diurnal salivary
cortisol patterns prior to pregnancy predict infant birth weight.
Health Psychol 2016;35:625–33.
82. Baibazarova E, van de Beek C, Cohen-Kettenis PT et al.
Influence of prenatal maternal stress, maternal plasma cortisol
and cortisol in the amniotic fluid on birth outcomes and child
temperament at 3 months. Psychoneuroendocrinology 2013;38:
907–15.
83. McTernan CL, Draper N, Nicholson H et al. Reduced placental
11beta-hydroxysteroid dehydrogenase type 2 mRNA levels in
human pregnancies complicated by intrauterine growth restriction: an analysis of possible mechanisms. J Clin Endocrinol
Metab 2001;86:4979–83.
84. Cottrell EC, Seckl JR. Prenatal stress, glucocorticoids and the
programming of adult disease. Front Behav Neurosci 2009;3:19.
85. Ruggles S, Genadek K, Goeken R et al. Integrated Public Use
Microdata Series: Version 6.0 [machine-readable database].
University of Minnesota, 2015.
86. DiGiuseppe DL, Aron DC, Ranbom L et al. Reliability of birth
certificate data: a multi-hospital comparison to medical records
information. Matern Child Health J 2002;6:169–79.
87. Northam S, Knapp TR. The reliability and validity of birth certificates. J Obstet Gynecol Neonatal Nurs 2006;35:3–12.
88. Buescher PA, Taylor KP, Davis MH et al. The quality of the new
birth certificate data: a validation study in North Carolina. Am J
Public Health 1993;83:1163–5.
89. Baumeister L, Marchi K, Pearl M et al. The validity of information on ‘race’ and ‘Hispanic ethnicity’ in California birth certificate data. Health Serv Res 2000;35:869–83.
90. Parker JD, Schoendorf KC. Implications of cleaning gestational
age data. Paediatr Perinat Epidemiol 2002;16:181–7.

1
2
3
4
5
6
7
8

Exhibit F

9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

ADDENDUM TO MPAS IN SUPPORT OF AMICUS CURIAE
2:18-cv-00490-JAM-KJN

PERSPECTIVES

SCIENCE AND SOCIETY

Stress-induced immune dysfunction:
implications for health
Ronald Glaser and Janice K. Kiecolt-Glaser
Abstract | Folk wisdom has long suggested
that stressful events take a toll on health.
The field of psychoneuroimmunology (PNI)
is now providing key mechanistic evidence
about the ways in which stressors — and
the negative emotions that they generate
— can be translated into physiological
changes. PNI researchers have used
animal and human models to learn how
the immune system communicates
bidirectionally with the central nervous
and endocrine systems and how these
interactions impact on health.

The central nervous system (CNS), the
endocrine system and the immune system are
complex systems that interact with each other.
Various stressors — from parachute jumping
to academic examinations to bereavement —
can dysregulate the immune response by
affecting the interplay of these systems.
Psychoneuroimmunology (PNI) is the broad
interdisciplinary research field that addresses
the interactions of these three systems1,2.
Studies undertaken during the past two
decades have provided evidence that immune
alterations that are stimulated by stressful
events, ranging from commonplace daily
hassles to chronic calamities, can provoke
health changes. One definition of a stressor is
a stimulus that activates the hypothalamic–
pituitary–adrenal (HPA) axis and/or the sympathetic nervous system (SNS) to help an
organism to adapt physiologically to deal
with a threat3. More broadly, psychological
stress ensues when events or environmental
demands exceed an individual’s perceived

ability to cope4. Researchers often categorize
stressors by their duration and course (discrete
versus continuous)5 (BOX 1). For example,
chronic stressors, such as suffering a traumatic
injury that leads to physical disability, can
force people to restructure key aspects of
their daily lives. Whereas chronic stressors are
deleterious to immune function, some investigators have suggested that very brief stressors, lasting less than 2 hours, might enhance
some aspects of immune function, such as
trafficking of cells from lymphoid organs to
the peripheral blood and the skin (BOX 2).

Stressors can increase susceptibility to infectious agents, influence the severity of infectious
disease, diminish the strength of immune
responses to vaccines, reactivate latent herpesviruses and slow wound healing. Moreover,
stressful events and the distress that they evoke
can also substantially increase the production
of pro-inflammatory cytokines that are associated with a spectrum of age-related diseases.
Accordingly, stress-related immune dysregulation might be one core mechanism behind a
diverse set of health risks1,3.
CNS–immune–endocrine interactions

Modulation of the immune response by the
CNS is mediated by a complex network of
bidirectional signals between the nervous,
endocrine and immune systems (FIG. 1). The
HPA axis and the autonomic nervous system provide two key pathways for immunesystem dysregulation: stressors can activate
the sympathetic–adrenal–medullary (SAM)
axis, as well as the HPA axis, and thereby
provoke the release of pituitary and adrenal

Box 1 | How is stress assessed?
When events or environmental demands exceed an individual’s ability to cope, the ensuing
psychological stress response typically includes negative thoughts and emotions4. Studies of
stress and immunity often use measures of negative mood that assess symptoms of general
distress, anxiety or depression. Researchers might also assess the number and type of recent
significant stressful life changes, or they might ask participants to rate their perceptions of stress
on a scale by answering certain questions, such as how frequently in the past week did you feel
you could not control important things in your life, or how often did you feel that things were
piling up so high that you could not overcome them4.
In addition, researchers often study the psychological and immunological responses of
individuals who are experiencing a distress-generating event (for example, students taking an
examination or spouses going through a divorce) or a more chronic stressor (such as caring for
a husband or wife who has Alzheimer’s disease)5. Other longer-term stressors that are associated
with immune alterations have included ‘burnout’ at work, job strain, unemployment, and
isolation and exposure to the hostile climate of Antarctica81. Adverse immunological changes
have also been documented for weeks or months following such natural disasters as earthquakes
and hurricanes, with more persistent immune dysregulation among those who suffered greater
personal losses82. Stressors that are perceived as unpredictable and/or uncontrollable might
continue to be associated with increased levels of stress hormones, even after repeated
exposures83. The ability to ‘unwind’ after stressful events — that is, to return to one’s
neuroendocrine baseline in a relatively short time — is thought to influence the total burden
that stressors place on an individual84.

NATURE REVIEWS | IMMUNOLOGY

VOLUME 5 | MARCH 2005 | 2 4 3

© 2005 Nature Publishing Group

PERSPECTIVES
hormones. For example, the catecholamines
(adrenaline and noradrenaline), adrenocorticotropic hormone (ACTH), cortisol,
growth hormone and prolactin are all influenced by negative events and negative emotions (BOX 1), and each of these hormones
can induce quantitative and qualitative
changes in immune function. Furthermore,
depression can substantially boost cortisol levels, and increases in cortisol levels can provoke
multiple adverse immunological changes.
Almost all immune cells have receptors for
one or more of the hormones that are associated with the HPA and SAM axes; these are
called ‘stress’ hormones (TABLE 1). Immune
modulation by these hormones might proceed through two pathways: directly, through
binding of the hormone to its cognate receptor at the surface of a cell; or indirectly — for
example, by inducing dysregulation of the
production of cytokines, such as interferon-γ
(IFN-γ), interleukin-1 (IL-1), IL-2, IL-6 and
tumour-necrosis factor (TNF). Cytokines
such as IFN-γ have many functions and affect
different target cells. Therefore, there are secondary effects of many stress hormones on
the immune response6,7.
Moreover, communication between the
CNS and the immune system is bidirectional.
For example, IL-1 influences the production
of corticotropin-releasing hormone (CRH)
by the hypothalamus. In turn, CRH can affect
the HPA axis and thereby trigger increases in
stress hormone levels, which results in dysregulation of immune function (FIG. 1). In addition,
lymphocytes can synthesize hormones such
as ACTH, prolactin and growth hormone8.
The role of lymphocyte-derived hormones in
immune responses is not well understood,
although they might have a role in modulating cell function within the microenvironment of lymphoid organs. Furthermore,
studies that show nerve fibres in the spleen
and thymus provide evidence of direct connections or ‘hard-wiring’ between the SNS
and lymphoid organs9. Therefore, there are
many pathways through which stressors might
influence immune function1,6. Moreover,
many individuals working in the field of PNI
are now focusing their efforts on immunesystem-to-brain communication and how the
activation of inflammatory-cytokine networks might shape mood, cognition and
behaviour10,11.
In addition to the direct influences of psychological states on endocrine and immune
function, stressed individuals are more likely
to have health habits that put them at greater
risk, including poorer sleep patterns, poorer
nutrition, less exercising and a greater propensity for abuse of alcohol, cigarettes and other

244

Box 2 | Can stress be beneficial?
The best evidence that stress might be good for the immune system comes from studies of mice
that are exposed to very brief stressors. Delayed-type-hypersensitivity skin responses following
either primary or secondary cutaneous antigen exposure were augmented following stressors
lasting 2 hours, compared with the response of non-stressed control animals. These effects seem
to be mediated by glucocorticoid- and adrenaline-induced stress responses85. It has been argued
that such immunoenhancement would be beneficial to survival, because skin wounding and
infection can result from brief aggressive encounters in nature85. In humans, short-term
stressors, such as public speaking, briefly increase natural-killer-cell activity5,86 and increase the
numbers of some types of leukocyte5. The latter change probably reflects transient alterations
in lymphocyte migration from lymphoid organs and peripheral blood, which is mediated by
receptors at the cell surface of lymphocytes (TABLE 1) or through sympathetic-nervous-system
innervation of lymphoid organs such as the spleen9. However, the same short-term stressors also
produce transitory changes in humans that would generally be seen as maladaptive: they reduce
lymphocyte proliferation5, increase pro-inflammatory cytokine production86 and impair the
ability of the skin to repair abrasions86. Further studies need to be carried out to help to clarify
health outcomes that are associated with short-term acute stressors.

drugs. Although these health behaviours have
immune and endocrine consequences, these
indirect effects of stress are not addressed
here; we focus on immune dysregulation by
stressors and the health consequences of
these changes.
Stress and infectious-disease risks

Stressors can enhance the risk of developing
infectious disease, and they can also prolong
infectious illness episodes. For the mouse
models used to explore this relationship,
restraint is a commonly used stressor. Mice are
placed in tubes such that they can move forwards or backwards but cannot turn around;
holes in the tubes ensure that the mice do not
overheat. Restraint is often applied overnight,
because this is the most active time for mice.
One example of data obtained using a mouse
model of influenza-virus infection shows that
restraint stress altered the immune response
to the virus, including the kinetics of the
antibody response and suppression of both
pro-inflammatory and anti-inflammatory
cytokine responses12,13. Mononuclear-cell
trafficking to virus-infected lungs was significantly reduced in stressed animals, as
was the size of the draining lymph nodes.
Virus-specific cytokine responses of T cells
in restraint-stressed mice were restored in
the draining lymph nodes by pharmacological blockade of the glucocorticoid receptor
with the glucocorticoid receptor antagonist
RU486. These and related studies have
shown that the HPA axis and the SNS are
the main immunoregulatory pathways that
can influence the pathophysiology of a viral
infection12,13.
Consistent with the mouse data on stress
and influenza-virus infections, influenza-virus
vaccine studies with human participants show
that stress can influence infectious-disease

| MARCH 2005 | VOLUME 5

risks. For example, men and women who were
chronically stressed by caring for a spouse with
dementia showed clear deficits in both their
cellular and humoral immune responses to an
influenza-virus vaccine compared with wellmatched control individuals who were not
carers14,15. The protective capacity of antiviral
vaccines depends on their ability to induce
both humoral and cell-mediated immune
responses16, both of which were poorer in
the stressed carers compared with control
individuals. Stress-associated impairments
in antibody responses after vaccination with
influenza virus have also been shown in
younger adults17.
Further studies have confirmed the finding
that stressful events and the negative emotions, such as anxiety and depression, that
accompany them can modulate the antibody
and T-cell responses to other antiviral vaccines, including the vaccines against infection
with hepatitis B virus and rubella virus18,19.
Moreover, antibody responses to antibacterial
vaccines are also influenced by stress: for
example, following vaccination, antibody
titres to a pneumococcal vaccine decreased
during a 6-month period in the carers of
spouses with dementia, whereas antibody
titres were stable in non-carers20. Similarly,
undergraduates who had received a meningitis C conjugate vaccine and who reported
greater stress had a poorer antibody response
1–12 months after vaccination21.
Responses to vaccines show clinically relevant alterations in immunological responses
to challenge under well-controlled conditions; accordingly, they function as a proxy for
a response to an infectious agent. Individuals
who were more distressed and more anxious
had immune responses to vaccines that were
delayed, substantially weaker and/or shorterlived. As a consequence, it is reasonable to

www.nature.com/reviews/immunol

© 2005 Nature Publishing Group

PERSPECTIVES

Stressor

Hypothalamus
Corticotropinreleasing hormone
Pituitary gland
Brain

'Hard-wiring'
sympathetic
innervation

Adrenocorticotropic
hormone

Adrenal
gland

Cortex

Prolactin
and growth
hormone
Lymph node

Medulla
Peripheral blood
Glucocorticoid
hormones

Noradrenaline
and adrenaline

NK cell
B cell

T cell

Cytokines,
such as IL-1

Monocyte
APC

Figure 1 | Stress-associated modulation of the hormone response by the central nervous
system. Experiencing a stressful situation, as perceived by the brain, results in the stimulation of
the hypothalamic–pituitary–adrenal (HPA) axis and the sympathetic–adrenal–medullary (SAM) axis.
The production of adrenocorticotropic hormone by the pituitary gland results in the production of
glucocorticoid hormones. The SAM axis can be activated by stimulation of the adrenal medulla to
produce the catecholamines adrenaline and noradrenaline, as well as by ‘hard-wiring’, through
sympathetic-nervous-system innervation of lymphoid organs. Leukocytes have receptors for stress
hormones that are produced by the pituitary and adrenal glands and can be modulated by the binding
of these hormones to their respective receptors. In addition, noradrenaline produced at nerve endings
can also modulate immune-cell function by binding its receptor at the surface of cells within lymphoid
organs. These interactions are bidirectional in that cytokines produced by immune cells can modulate
the activity of the hypothalamus. APC, antigen-presenting cell; IL-1, interleukin-1; NK, natural killer.

assume that these same individuals would
also be slower to develop immune responses
to pathogens; therefore, they could be at
greater risk of developing more severe illness. Consistent with this argument, adults
who show poorer responses to vaccines also
experience higher rates of clinical illness22.
In agreement with these vaccine studies,
researchers have also shown that distress can
alter an individual’s susceptibility to infection
with respiratory viruses4,23, 24. In a group of 394
healthy volunteers who were inoculated with
one of five strains of respiratory virus, severity of both respiratory infection and clinical
cold symptoms increased in a dose–response
relationship as scores increased on a psychological stress index. The stress index was a

compilation of three common measures: the
number of stressful life events, the degree
that a participant felt that current demands
exceeded his or her ability to cope, and
scores from a negative-emotion word list
(including words such as sad, angry and nervous). Importantly, the risk did not differ
across the five strains of respiratory virus
studied. In further related work from the
same laboratory, stressors that lasted for
1 month or more were the best predictors of
developing colds; volunteers who reported
more enduring interpersonal difficulties
with family or friends were substantially
more likely to develop a cold after inoculation with a rhinovirus23. Similarly, other
researchers reported that individuals who

NATURE REVIEWS | IMMUNOLOGY

developed cold symptoms following inoculation with rhinovirus had higher numbers
of recent stressful life events than those who
did not24.
Studies carried out with human participants in which individuals have been exposed
to a pathogen or a vaccine give researchers a
means of controlling exposure and dosage;
moreover, because immune function can
be assessed before the infectious challenge,
these studies provide excellent data on causality, thereby complementing evidence from
research that addresses the course of naturally
occurring infections25–35. The similarity of the
data from human and rodent studies provides
strong evidence that stress can dysregulate the
humoral and cellular immune responses to
pathogens and increase the risk of developing
infectious disease.
HIV and the herpesviruses are different
from many other viruses, such as rhinoviruses and influenza virus, in that they
remain in a latent state in the body after primary infection. To investigate the possibility
that social stress was a contributor to the
rate of progression in HIV-associated disease, rhesus macaques were inoculated with
simian immunodeficiency virus (SIV)36.
Animals that were assigned to the stable
social condition (the same three animals
met every day) had lower concentrations of
SIV RNA in plasma early after inoculation
and survived longer than those in the unstable social condition (different two-, threeand four-member groups were formed
every day).
Studies of HIV-infected men have also indicated that stress increases the rate of disease
progression. For example, in a longitudinal
study of HIV-positive men who were asymptomatic at entry to the study, faster progression
to AIDS was associated with more stressful
life events and less social or interpersonal support25; indeed, at 5.5 years after entry into the
study, the probability of developing AIDS was
two- to threefold higher in men who were
above the median level for stress or below the
median level for support compared with those
who were below the median level for stress or
above the median level for support. Other
researchers reported that the course of HIV
infection was accelerated in gay men who concealed their homosexual identity compared
with men who did not26.
Considerable anecdotal evidence has supported the relationship between psychological
stress and the development, duration and
recurrence of herpesvirus infections. The cellular immune response has an important role
in controlling the pathophysiology of both
lytic herpesvirus infections and the expression

VOLUME 5 | MARCH 2005 | 2 4 5

© 2005 Nature Publishing Group

PERSPECTIVES

Table 1 | Interactions of hormones and immune cells
Hormone

Expression of receptors
by immune cells

Examples of effects on cell function

References

Glucocorticoids

T and B cells, neutrophils,
monocytes and macrophages

Inhibit inflammation; inhibit the
production of IL-12 by antigenpresenting cells; induce a shift
from production of TH1 to TH2 cytokines

87,88

Substance P

T and B cells, eosinophils,
mast cells, monocytes and
macrophages

Stimulates mitogen-induced
blastogenesis; increases trafficking
of cells from lymph nodes to
peripheral blood; stimulates
monocytes to produce several
cytokines, such as IL-1, IL-6
and TNF

89

Neuropeptide Y

T and B cells, dendritic cells,
monocytes and macrophages

Can downregulate antibody
production to T-cell-dependent
antigens by its effect on dendritic
cells, and T and B cells

90

Corticotropinreleasing hormone

T cells, monocytes
and macrophages

Increases production of IL-1
by monocytes; evidence for
autocrine and/or paracrine
modulation of inflammation

91

Prolactin

T and B cells, granulocytes,
NK cells, monocytes and
macrophages

Can stimulate lymphoid-cell clonal
expansion; might function as an
in vitro co-mitogen for NK cells
and macrophages

Growth hormone

T and B cells, NK cells,
monocytes and macrophages

Helps to maintain competence of
T and B cells, and macrophages;
stimulates antibody production
and NK-cell activity

94

Catecholamines
(adrenaline and
noradrenaline)

T and B cells, NK cells,
monocytes and macrophages

Induce a shift to a TH2 response,
involving antigen-presenting cells
and TH1 cells

95

Serotonin

T and B cells, NK cells,
monocytes and macrophages

Modulates the synthesis of IFN-γ
by NK cells; stimulates the
production of IL-16 (a chemotactic
factor) by T cells

96

92,93

IFN-γ, interferon-γ; IL, interleukin; NK, natural killer; TH, T helper; TNF, tumour-necrosis factor.

and/or replication of latent herpesviruses.
When the cellular immune response is
impaired, one or more herpesviruses can be
reactivated, and herpesvirus infections are
often more severe.
Herpes simplex virus (HSV) is a natural
human pathogen that is characterized by its
ability to cause an acute infection at a peripheral site and to establish a latent infection in the
local sensory ganglia, and stress can exacerbate
HSV lytic infection. Mouse models have been
developed to study the effect of stress on the
pathophysiology of HSV latent and lytic infections. Indeed, several studies carried out during
the past 15 years have provided compelling
experimental evidence that stress not only
increases the development and severity of
HSV infection, in both the peripheral nervous system13,37–39 and the CNS, but also suppresses components of primary13,37,39–41 and
memory13,38,41 cytotoxic T lymphocyte (CTL)
responses to HSV infection.
Surgical and pharmacological approaches
have shown the ability of both the HPA13 and
the SAM41 axes to mediate stress-induced

246

modulation of immunity and HSV-associated
pathology. For example, mice treated with
6-hydroxydopamine (6-OHDA) to induce
peripheral sympathetic denervation were
inhibited in their ability to generate primary
HSV-1-specific CTLs when infected with
the virus41. The suppression of CTL production could result from a large release of
noradrenaline induced by 6-OHDA and
increased levels of corticosterone. In another
study, surgical removal of the adrenal gland
blocked the suppression of HSV-1-specific
CTLs that was induced by restraint stress
and also blocked the production of IL-6
and IFN-γ13.
Relationships between neuroendocrine
activity, immune function and latent HSV
reactivation have also been documented42,43;
infected mice that were exposed to a stressor
showed reactivation of the latent virus,
whereas non-stressed controls showed no
reactivation43. It is important to keep in mind
that these experiments were carried out using
mice in a laboratory setting; however, the data
still provide some insight into how stress

| MARCH 2005 | VOLUME 5

could modulate the immune response to
HSV in humans.
Indeed, psychological stressors have been
linked to more frequent recurrences of lesions
in individuals who are latently infected with
HSV-1 or HSV-2. For example, women who
reported greater persistent stress from events
that lasted longer than 1 week also had more
recurrences of genital herpes28. Similarly,
more chronically distressed individuals had
more frequent recurrences of re-activation of
HSV-1 (REF. 29) and HSV-2 (REF. 30).
The incidence of Herpes zoster (also known
as shingles), which is caused by the reactivation
of latent varicella-zoster virus (VZV), increases
with age, presumably owing to a decline in cellmediated immunity to VZV44. A case–control
study indicated that psychological stress in
healthy community-dwelling older adults
was associated with the occurrence of herpes
zoster31. Other researchers evaluated the possibility that VZV-specific immunity could be
altered by means of a behavioural intervention, such as T’ai chi (also known as ‘meditation through movement’)44. Older adults who

www.nature.com/reviews/immunol

© 2005 Nature Publishing Group

PERSPECTIVES

Epithelial cell

Cutaneous wound

a

Skin

Other chemoattractants
PDGF

Platelet
Neutrophil

T cell

Recruitment of
inflammatory cells

c

Production of CXCL8,
IL-1α, IL-1β, CCL2,
VEGF, TGF-β and TNF

b
Endothelial cell

Macrophage

Blood

Figure 2 | Influence of stress on pro-inflammatory cytokine responses in wound healing. Stress
can influence key pro-inflammatory cytokine responses in the early phase (the first 24 hours) of the healing
of skin wounds, through dysregulation of cytokine secretion at the wound site and recruitment and
activation of circulating peripheral-blood leukocytes that traffic to the wound site48,52,54. Using a skin
wound as an example, blood platelets at the wound site produce platelet-derived growth factors (PDGFs) (a).
Other chemoattractants are also produced by activated parenchymal cells. A concentration gradient
is established, with higher levels of chemoattractants at the wound site attracting immune cells, such as
neutrophils and macrophages. These cells have important roles in the early phases of wound healing.
For example, neutrophils clean the area of bacteria and, together with activated macrophages, they
phagocytose the bacteria and produce cytokines that stimulate the growth of fibroblasts. The leukocytes
transmigrate through the endothelium of the blood-vessel wall to the wound site in the skin (b) and are
activated to proliferate and produce cytokines and chemokines, such as CXC-chemokine ligand 8
(CXCL8; also known as IL-8), IL-1α, IL-1β, transforming growth factor-β (TGF-β), vascular endothelial
growth factor (VEGF), CC-chemokine ligand 2 (CCL2; also known as MCP1) and tumour-necrosis factor
(TNF), at the wound site (c). These cytokines continue to function as chemoattractants for the continued
migration of cells to the site. The proliferative phase of wound healing involves the recruitment and
replication of cells that are required for tissue regeneration and capillary growth. Therefore, the
downregulation of the early inflammatory response by an increase in serum cortisol levels can help
to explain how stress affects wound healing49.

were randomly assigned to T’ai chi showed
a 50% increase in VZV-specific cellular
immunity between the start and the end of
the 15-week intervention compared with no
change in the ‘waiting-list’ control group.
Epstein–Barr virus (EBV) — the aetiological agent of infectious mononucleosis —
is another herpesvirus that establishes latent
infection and can be modulated by psychological stressors. In one early study, West
Point Military Academy (New York, United
States) cadets who were seronegative for EBV
on entry into the academy were followed for
4 years27. Men with particular psychosocial
risk factors (high motivation for a military
career in the face of poorer academic performance) were more likely to develop infectious
mononucleosis and were likely to be hospitalized for longer periods. In addition, these risk
factors were also associated with increased
EBV-specific antibody titres in cadets who
had been infected with EBV but had not
developed obvious clinical symptoms.
A series of studies provided mechanistic
data that revealed the effect of stress on EBV

latency. Medical students had substantially
higher titres of IgG specific for EBV capsid
antigen, and these were associated with more
stressful examination periods compared
with lower-stress periods45. In a further study
of medical students, examination stress produced a significant decrease in the ability of
EBV-specific CTLs to kill EBV-infected autologous B cells45. The results of several studies
have shown that various psychological stressors — including examination stress, caring
for a spouse with dementia and spaceflights
by astronauts — can reactivate latent EBV
and cytomegalovirus (CMV)32–35,45. Together,
these human and animal studies show that
stress can modulate the steady-state expression of latent HSV, EBV and CMV, downregulating the specific T-cell response to the
virus to an extent that is sufficient to result in
viral reactivation. Although the mechanisms
that underlie stress-associated reactivation of
latent herpesviruses are not fully understood, in vitro studies of cells that are latently
infected with EBV have shown that glucocorticoid hormones can reactivate the virus.

NATURE REVIEWS | IMMUNOLOGY

For example, a glucocorticoid hormone,
dexamethasone, can reactivate latent EBV
and enhance the lytic replication of the virus
in EBV-superinfected cells in vitro, but the
catecholamine hormones do not induce
such a response. Other stress hormones —
CRH and ACTH — cannot induce reactivation of latent EBV, but they can enhance lytic
replication in EBV-superinfected cells46.
Different types of stressor can have different effects on reactivation of latent HSV-1
and EBV43,47. For example, although restraintstressed mice did not show evidence of reactivation of latent HSV-1, infectious HSV-1
was isolated from approximately 50% of the
mice that were subjected to social reorganizational stress, despite both stressors resulting
in similar increases in serum corticosterone
levels43. Data from studies of students at
West Point Military Academy also showed
that different types of stress could have an
impact on the reactivation of latent HSV-1
and EBV47. The mechanisms underlying
these differences are not understood, but
clearly, a factor as obvious as disparities in
glucocorticoid hormone levels is not sufficient to explain variations in viral reactivation. Together, these studies highlight the
complex interactions that underlie the relationships between stress, neuroendocrine
activity, immune function and herpesvirus
pathogenesis, and they indicate the many
ways in which these relationships are central
to a lifelong defence against herpesvirus
infections.
Stress and wound healing

Wound repair progresses through several
overlapping stages48. In the initial inflammatory stage, vasoconstriction and blood coagulation are followed by platelet activation and
the release of platelet-derived growth factors
(PDGFs), as well as the release of chemoattractant factors by injured parenchymal cells.
Cytokines and chemokines — such as IL-1α,
IL-1β, transforming growth factor-β (TGF-β),
vascular endothelial growth factor (VEGF),
TNF and CXC-chemokine ligand 8 (CXCL8;
also known as IL-8) — are important in the
early stages of wound healing. These factors
function as chemoattractants, promoting the
migration of phagocytes and other cells to
the wound site, thereby starting the proliferative phase, which involves the recruitment
and replication of cells that are required for
tissue regeneration and capillary regrowth.
The final step, wound remodelling, might
continue for weeks or months. So, the healing
process is a cascade, and success in the later
stages of wound repair depends to a large
extent on initial events48.

VOLUME 5 | MARCH 2005 | 2 4 7

© 2005 Nature Publishing Group

PERSPECTIVES
Immune function has a key role in the
early stages of this cascade (FIG. 2). CXCL8
and pro-inflammatory cytokines, such as IL-1
and TNF, are essential to this effort; they
help to protect against infection and prepare
injured tissue for repair by enhancing the
recruitment and activation of phagocytes49.
Furthermore, cytokines that are released by
recruited cells regulate the ability of fibroblasts and epithelial cells to remodel the
damaged tissue49. IL-1 that is produced early
after tissue injury can regulate the production,
release and activation of metalloproteinases
that are important in the destruction and
remodelling of the wound. IL-1 also regulates
fibroblast chemotaxis and the production of
collagen49. Moreover, IL-1 stimulates the production of other cytokines that are important
for wound healing, including IL-2, IL-6 and
CXCL8 (REF. 49). Accordingly, IL-1 deficits early
in the wound-repair cascade can have adverse
consequences downstream.
Stress disrupts the production of proinflammatory cytokines that are important
for wound healing, a mechanism that produces substantial delays in wound repair.
For example, in a clinical study, women who
were experiencing the long-term stress of
caring for a relative with Alzheimer’s disease
took 24% longer than sociodemographically matched controls to heal a small, standardized dermal wound. Consistent with
these differences in wound repair, peripheralblood leukocytes (PBLs) obtained from carers also produced less IL-1β in response to
lipopolysaccharide (LPS) stimulation50. In a
subsequent study in a different population,
wounds produced in the hard palate 3 days
before important examinations healed an
average of 40% more slowly than identical
wounds made during summer holidays: no
student healed as rapidly during examinations as during the holiday period, and no
student produced as much IL-1β when his or
her PBLs were stimulated with LPS51.
Mouse models have also been developed
to study the effect of stress on wound healing.
These studies have confirmed and extended
the data obtained by studying humans. Mice
that were subjected to restraint stress and had
a standardized 3.5-mm full-thickness cutaneous punch-biopsy wound healed this
wound an average of 27% more slowly than
control mice52. Analysis of the cellularity of
wound sites using cross-sections of dermal
and epidermal layers showed less leukocyte
infiltration of the wound sites in restraintstressed mice at 1 and 3 days after wounding,
compared with controls52. Serum corticosterone levels in the restraint-stressed group
were more than fourfold higher than those

248

of control animals52. Blocking glucocorticoid receptors in restraint-stressed animals,
using RU40555, resulted in healing rates
that were similar to those of control animals52. Accordingly, these data provide evidence that disruption of neuroendocrine
homeostasis modulates the early stages of
wound healing.
Higher levels of glucocorticoids have several adverse effects on various components of
the wound-healing process. For example,
they might slow wound healing by altering
local levels of pro-inflammatory cytokines.
Hübner et al.48 showed that the strong and
early induction of IL-1α, IL-1β and TNF
expression at the site after wounding was significantly reduced after pretreatment of mice
with glucocorticoids. Similarly, human studies have also shown that stress-induced
increases in glucocorticoids can transiently
suppress IL-1β, TNF and PDGF production53.
Accordingly, dysregulation of glucocorticoid
secretion provides one obvious neuroendocrine pathway through which stress alters
wound healing.
In humans, a suction-blister model
enabled investigators to measure immune
responses that are central to the early stages
of wound healing in vivo and occur at the
wound site, providing key data on the
inflammatory response that have direct clinical relevance54,55. The suction-blister model
provides an excellent mechanism to study
the migration of neutrophils and macrophages and the production of cytokines at
wound sites for the first 2 days after wounding. Commonly, after raising several blisters
and removing their roofs (the epidermis),
plastic templates with wells containing a
salt solution and autologous serum are
placed over the lesions, and cells migrate
to the wound sites and collect in the wells.
The serial collection of samples from the
wells as time progresses allows for cell
phenotyping and cytokine measurement as
the local immune response evolves. Using
this approach to study stress and wound
healing, women who reported more stress
produced significantly lower levels of two
pro-inflammatory cytokines (IL-1α and
CXCL8) that are important for the early
stages of wound healing54.
Therefore, convergent data from mouse
and human studies have shown that stress has
substantial adverse effects on wound repair. In
agreement with these laboratory findings,
several studies have shown that greater fear or
distress before surgery is associated with
poorer outcomes, including longer hospitalization, more post-operative complications
and higher rates of rehospitalization56,57.

| MARCH 2005 | VOLUME 5

Stress and inflammation

The pro-inflammatory cytokine IL-6, which
is produced by T cells, B cells, monocytes and
several non-lymphoid cell types, has an
important role in the acute-phase response3.
IL-6 is an important inducer of C-reactive
protein (CRP) by the liver, and the combination of IL-6 and CRP is important in the
process that leads to the development of cardiovascular disease3,58. As previously discussed,
stress induces immune dysregulation partly
through alterations in the production of proinflammatory cytokines. Both physical and
psychological stressors can provoke transient
increases in pro-inflammatory cytokines, particularly in IL-6 (REFS 53,59). In animal models,
both stress and administration of adrenaline
increase levels of IL-6 in the plasma, which is
consistent with evidence that IL-6 production
is stimulated by β-adrenergic receptors, as well
as through other pathways3.
Importantly, negative emotions, such as
depression and anxiety, augment the production of IL-6 (REFS 60–62). Indeed, both stressors
and depression can sensitize the inflammatory response, thereby producing heightened
responsiveness to subsequent stressful events,
as well as to antigen challenge59,61–63. For
example, individuals who reported more
depressive symptoms showed increases in
serum IL-6 levels 2 weeks after vaccination
against influenza-virus infection, whereas
there was little change in IL-6 levels in those
individuals who reported few or no symptoms61. This is consistent with other evidence
of cross-sensitization between cytokines and
stressors in human and animal studies59,62,63.
These stress-related changes have broad implications for health: increased levels of proinflammatory cytokines, such as IL-6, have
been linked to various age-related diseases
and conditions (including cardiovascular
disease, osteoporosis, arthritis, type 2 diabetes, frailty and functional decline) and to
certain cancers (such as chronic lymphocytic
leukaemia)64.
Stress, inflammation and ageing

One recent longitudinal study highlighted the
deleterious longer-term immunological consequences of chronic stress: the average annual
rate of increase in serum IL-6 was about fourfold higher in men and women who were
chronically stressed by caring for a spouse with
dementia than in similar individuals who did
not have caring responsibilities65. Possible
consequences of these different trajectories are
indicated by epidemiological studies of individuals of 65 years of age or older64; within
these population studies, individuals whose
IL-6 values fell within the highest quartile had

www.nature.com/reviews/immunol

© 2005 Nature Publishing Group

PERSPECTIVES
a twofold greater risk of death within the following 4–5 years compared with those whose
IL-6 values were in the lowest quartile64.
Application of the epidemiological risk values
to the data from carers indicated that carers
would, on average, have values that crossed
into the highest quartile around the age of
75, whereas the IL-6 values of control individuals would not reach that level until after
the age of 90.
Another recent study also supports the
hypothesis that chronic stress might be associated with premature ageing of immune
cells. Telomerase activity and telomere length
— two cellular markers that are associated
with ageing — were measured in peripheralblood mononuclear cells obtained from
mothers caring for a chronically ill child, as
well as from mothers of healthy children66.
Carers reported greater stress than controls,
but reports of a higher level of perceived
stress were associated with lower telomerase
activity and shorter telomere length, regardless of whether the mother’s child was ill or
healthy. Reports of high stress levels were also
associated with higher oxidative-stress activity, as measured by levels of F2-isoprostanes,
another independent measure associated
with ageing66.
Taken together, the data regarding the IL-6
levels of carers of spouses with dementia65
and the data regarding telomerase activity
and length66 provide evidence of mechanisms
through which chronic stressors might accelerate the risk of developing many age-related
diseases by ‘premature ageing’ of the immune
response. Indeed, a prospective populationbased cohort study found that the relative risk
for all-cause mortality over a 4-year period
in strained carers was 63% higher than in
control individuals who were not carers67.
Conclusions and future directions

Great strides have been made in the field of
PNI towards understanding some of the
interactions between the CNS, endocrine
system and immune system, as well towards
understanding how distress modulates
these three complex systems. Although the
mechanisms that underlie these interactions are complex, and although it will
probably take many years to fully understand how these three systems interact, there
are already clear translational implications
from laboratory data.
Herpesvirus infections carry substantial
human costs because the latent viruses are
linked to considerable pain and suffering.
Moreover, the evidence that psychological
stressors can reactivate latent herpesviruses
might have the most notable implications for

people who are already immunosuppressed
(such as patients who have received an organ
transplant or patients infected with HIV),
owing to the risk of these individuals developing EBV-associated B-cell lymphomas.
Indeed, reactivation of latent EBV, HSV-1
and CMV is associated with significant morbidity and mortality of immunosuppressed
patients.
Furthermore, on the basis of speculation
that chronic inflammation might be a contributing factor in up to 15% of all cancer
cases68, stress-induced increases in the inflammatory response could be a broader pathway
that links stress with cancer. Although it is
beyond the scope of this article, the possibility
that the physiological changes associated with
stress could be key factors in cancer risk and
progression has recently been reviewed69.
The results of the vaccine studies are particularly important for individuals who might
be at a higher risk of developing complications that are associated with respiratoryvirus infections, such as older individuals for
whom increased susceptibility to pathogens is
a serious health problem: together, influenza
and pneumonia are the fifth leading cause of
mortality in individuals aged 50 or older16.
Biologically, the largest deleterious or enhancing consequences of stress are likely to occur
when biological vulnerability is greatest: that
is, early and late in life70. Older adults seem to
show greater immunological impairments
associated with distress or depression than
younger adults14,57. However, the studies indicate that vaccine efficacy can be compromised
by psychological stress, even in younger adults
— an important public-health finding in its
own right. These studies should be considered
in the planning of clinical studies using cancer
vaccines. The efficacy of such vaccines will
depend on an optimum immune response.
The possibility that stressors might have a
long-term impact on the developing endocrine
and immune systems of infants and young
children is an important question that has not
been well studied in the PNI field. Indeed,
excellent developmental studies of primates
indicate that early stressors can reverberate
for the life of an individual70.
In accordance with the evidence that
stress delays wound healing, more than 200
studies published in the past 3 decades have
shown beneficial effects from pre-surgical
interventions. These beneficial effects include
decreased anxiety and stress reductions when
hospitalized, fewer post-operative complications, better treatment compliance, less pain
and reduced use of analgesics, and alterations
in various physiological indices56,57. Given the
substantial consequences of stress for wound

NATURE REVIEWS | IMMUNOLOGY

repair, even small reductions in anxiety could
have substantial clinical consequences, both
directly and indirectly57.
More broadly, researchers have used several diverse strategies to modulate immune
function, including relaxation, hypnosis, exercise, classical conditioning, self-disclosure and
cognitive behavioural interventions. These
interventions have generally produced positive endocrine and immune changes5,44,71–74.
Although it is not yet clear to what extent
these positive immunological changes translate
into any concrete improvements in relevant
aspects of health, such as alterations in the incidence, severity or duration of infectious and/or
malignant disease, the preliminary evidence
seems to be promising.
The role that genetics might have in these
complex relationships is unknown, and this is
an important new area that deserves exploration. For example, do individuals who have
one or more variants of the polymorphisms
associated with increased production of cortisol show greater immunological dysregulation
when faced with stressful events?
Several studies have provided convincing
evidence linking stress-induced immune
dysregulation with morbidity and mortality.
Animal models that involve viral infections
have confirmed that stress can exacerbate
morbidity that is associated with a viral
infection37,75–77. Stress can also exacerbate
bacterial infections, such as infections with
Listeria monocytogenes78,79. In both humans
and mice, studies of wound healing show a
direct link between stress-associated immune
dysregulation and health outcome, with
well-documented relationships occurring
between stress hormones, the immune
response and the rate of wound healing50–52,54.
Together, these studies support the hypothesis that morbidity can be directly linked to
stress-induced immune dysregulation.
Using a mouse model, it was also shown
that stress-induced immune dysregulation
can cause mortality80. Restraint-stressed mice
infected with Theiler’s murine encephalomyelitis virus (TMEV) had an increased risk
of dying. TMEV is a Picorna virus, which
can cause CNS lesions. Higher titres of the
virus were observed in the stressed mice
compared with the control mice, and the
underlying mechanism that accounted for
the increased mortality in restraint-stressed
mice was related to corticosterone-induced
immune suppression.
The field of PNI is improving our understanding of the complex physiological
changes that take place in stressful situations
and providing new insights into various
clinical applications. This research field is

VOLUME 5 | MARCH 2005 | 2 4 9

© 2005 Nature Publishing Group

PERSPECTIVES
also contributing to our knowledge of how
the immune system operates in an environment in which there is bidirectional communication with other bodily systems. Despite
the remarkable complexities of the interactions between the CNS, the immune system
and the endocrine system, the researchers are
making good progress at the molecular, cellular and organ-system levels. And, with that
knowledge, the potential for new approaches
to treatment is evident.
Ronald Glaser is at the Department of Molecular
Virology, Immunology and Medical Genetics,
the College of Medicine and Public Health,
the Institute for Behavioral Medical Research and
the Comprehensive Cancer Center, Ohio State
University, Columbus, Ohio 43210, USA.
Janice K. Kiecolt-Glaser is at the Department of
Psychiatry, the College of Medicine and Public
Health, the Institute for Behavioral Medical
Research and the Comprehensive Cancer Center,
Ohio State University, Columbus, Ohio 43210,
USA.

17.

18.

19.

20.

21.

22.
23.

24.

Correspondence to R.G.
e-mail: [e-mail]
doi:10.1038/nri1571
1.
2.

3.

4.

5.

6.

7.
8.

9.

10.

11.

12.

13.

14.

15.

16.

250

25.

26.

Rabin, B. S. Stress, Immune Function, and Health:
the Connection (Wiley–Liss & Sons, New York, 1999).
Ader, R., Felten, D. L. & Cohen, N. (eds)
Psychoneuroimmunology 3rd edn (Academic, San Diego,
2001).
Black, P. H. The inflammatory response is an integral part
of the stress response: implications for atherosclerosis,
insulin resistance, type II diabetes and metabolic
syndrome X. Brain Behav. Immun. 17, 350–364 (2003).
Cohen, S., Tyrrell, D. A. & Smith, A. P. Psychological
stress and susceptibility to the common cold. N. Engl.
J. Med. 325, 606–612 (1991).
Segerstrom, S. C. & Miller, G. E. Psychological stress
and the human immune system: a meta-analytic study
of 30 years of inquiry. Psychol. Bull. 130, 1–37 (2004).
Webster, J. I., Tonelli, L. & Sternberg, E. M.
Neuroendocrine regulation of immunity. Annu. Rev.
Immunol. 20, 125–163 (2002).
Padgett, D. A. & Glaser, R. How stress influences the
immune response. Trends Immunol. 24, 444–448 (2003).
Weigent, D. A. & Blalock, J. E. Production of peptide
hormones and neurotransmitters by the immune system.
Chem. Immunol. 69, 1–30 (1997).
Bellinger, D. L., Lorton, D., Lubahn, C. & Felten, D. L. in
Psychoneuroimmunology 3rd edn Vol. 2 (eds Ader, R.,
Felten, D. L. & Cohen, N.) 55–112 (Academic, San Diego,
2001).
Capuron, L., Ravaud, A., Miller, A. H. & Dantzer, R.
Baseline mood and psychosocial characteristics
of patients developing depressive symptoms during
interleukin-2 and/or interferon-α cancer therapy.
Brain Behav. Immun. 18, 205–213 (2004).
Maier, S. F. Bi-directional immune–brain communication:
implications for understanding stress, pain, and
cognition. Brain Behav. Immun. 17, 69–85 (2003).
Konstantinos, A. P. & Sheridan, J. F. Stress and influenza
viral infection: modulation of proinflammatory cytokine
responses in the lung. Respir. Physiol. 128, 71–77 (2001).
Bonneau, R. H., Padgett, D. A. & Sheridan, J. F.
in Psychoneuroimmunology 3rd edn Vol. 2 (eds
Ader, R., Felten, D. L. & Cohen, N.) 483–498
(Academic, San Diego, 2001).
Kiecolt-Glaser, J. K., Glaser, R., Gravenstein, S.,
Malarkey, W. B. & Sheridan, J. Chronic stress alters
the immune response to influenza virus vaccine in older
adults. Proc. Natl Acad. Sci. USA 93, 3043–3047 (1996).
Vedhara, K. et al. Chronic stress in elderly carers of
dementia patients and antibody response to influenza
vaccination. Lancet 353, 627–631 (1999).
Deng, Y., Jing, Y., Campbell, A. E. & Gravenstein, S.
Age-related impaired type 1 T cell responses to
influenza: reduced activation ex vivo, decreased

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

expansion in CTL culture in vitro, and blunted response
to influenza vaccination in vivo in the elderly. J. Immunol.
172, 3437–3446 (2004).
Miller, G. E. et al. Psychological stress and antibody
response to influenza vaccination: when is the critical
period for stress, and how does it get inside the body?
Psychosom. Med. 66, 215–223 (2004).
Glaser, R. et al. Stress-induced modulation of the
immune response to recombinant hepatitis B vaccine.
Psychosom. Med. 54, 22–29 (1992).
Morag, M., Morag, A., Reichenberg, A., Lerer, B. &
Yirmiya, R. Psychological variables as predictors of
rubella antibody titers and fatigue — a prospective,
double blind study. J. Psychiatr. Res. 33, 389–395
(1999).
Glaser, R., Sheridan, J. F., Malarkey, W. B.,
MacCallum, R. C. & Kiecolt-Glaser, J. K. Chronic stress
modulates the immune response to a pneumococcal
pneumonia vaccine. Psychosom. Med. 62, 804–807
(2000).
Burns, V. E., Drayson, M., Ring, C. & Carroll, D. Perceived
stress and psychological well-being are associated with
antibody status after meningitis C conjugate vaccination.
Psychosom. Med. 64, 963–970 (2002).
Plotkin, S. A. Immunologic correlates of protection induced
by vaccination. Pediatr. Infect. Dis. 20, 73–75 (2001).
Cohen, S. et al. Types of stressors that increase
susceptibility to the common cold in healthy adults.
Health Psychol. 17, 214–223 (1998).
Stone, A. A. et al. Development of common cold
symptoms following experimental rhinovirus infection
is related to prior stressful life events. Behav. Med.
18, 115–120 (1992).
Leserman, J. et al. Progression to AIDS: the effects
of stress, depressive symptoms, and social support.
Psychosom. Med. 61, 397–406 (1999).
Cole, S. W., Kemeny, M. E., Taylor, S. E., Visscher, B. R. &
Fahey, J. L. Accelerated course of human
immunodeficiency virus infection in gay men who conceal
their homosexual identity. Psychosom. Med. 58, 1–13
(1996).
Kasl, S. V., Evans, A. S. & Niederman, J. C. Psychosocial
risk factors in the development of infectious
mononucleosis. Psychosom. Med. 41, 445–466 (1979).
Cohen, F. et al. Persistent stress as a predictor of genital
herpes recurrence. Arch. Intern. Med. 159, 2430–2436
(1999).
Luborsky, L., Mintz, J., Brightman, U. J. & Katcher, A. H.
Herpes simplex and moods: a longitudinal study.
J. Psychosom. Res. 20, 543–548 (1976).
Goldmeier, D. & Johnson, A. Does psychiatric illness
affect the recurrence rate of genital herpes? Br. J. Vener.
Dis. 58, 40–43 (1982).
Schmader, K., Studenski, S., MacMillan, J.,
Grufferman, S. & Cohen, H. J. Are stressful life events
risk factors for herpes zoster? J. Am. Geriatr. Soc.
38, 1188–1194 (1990).
Payne, D. A., Mehta, S. K., Tyring, S. K., Stowe, R. P. &
Pierson, D. L. Incidence of Epstein–Barr virus in astronaut
saliva during spaceflight. Aviat. Space Environ. Med.
70, 1211–1213 (1999).
Mehta, S. K., Stowe, R. P., Feiveson, A. H., Tyring, S. K. &
Pierson, D. L. Reactivation and shedding of
cytomegalovirus in astronauts during spaceflight. J. Infect.
Dis. 182, 1761–1764 (2000).
Prosch, S. et al. A novel link between stress and
human cytomegalovirus (HCMV) infection: sympathetic
hyperactivity stimulates HCMV activation. Virology
272, 357–365 (2000).
Sarid, O., Anson, O., Yaari, A. & Margalith, M. Epstein–Barr
virus specific salivary antibodies as related to stress
caused by examinations. J. Med. Virol. 64, 149–156
(2002).
Capitanio, J. P., Mendoza, S. P., Lerche, N. W. &
Mason, W. A. Social stress results in altered
glucocorticoid regulation and shorter survival in simian
acquired immune deficiency syndrome. Proc. Natl
Acad. Sci. USA 95, 4714–4719 (1998).
Kusnecov, A. V. et al. Decreased herpes simplex viral
immunity and enhanced pathogenesis following stressor
administration in mice. J. Neuroimmunol. 38, 129–137
(1992).
Wonnacott, K. M. & Bonneau, R. H. The effects of stress
on memory cytotoxic T lymphocyte-mediated protection
against herpes simplex virus infection at mucosal sites.
Brain Behav. Immun. 16, 104–117 (2002).
Cao, L., Martin, A., Polakos, N. & Moynihan, J. A. Stress
causes a further decrease in immunity to herpes simplex
virus-1 in immunocompromised hosts. J. Neuroimmunol.
156, 21–30 (2004).

| MARCH 2005 | VOLUME 5

40. Bonneau, R., Sheridan, J., Feng, N. & Glaser, R.
Stress-induced modulation of the primary cellular
immune response to herpes simplex virus infection is
mediated by both adrenal-dependent and independent
mechanisms. J. Neuroimmunol. 42, 167–176 (1993).
41. Leo, N. A. & Bonneau, R. H. Mechanisms underlying
chemical sympathectomy-induced suppression of herpes
simplex virus-specific cytotoxic T lymphocyte activation
and function. J. Neuroimmunol. 110, 45–56 (2000).
42. Halford, W. P., Gebhardt, B. M. & Carr, D. J.
Mechanisms of herpes simplex virus type 1
reactivation. J. Virol. 70, 5051–5060 (1996).
43. Padgett, D. A. et al. Social stress and the reactivation
of latent herpes simplex virus type 1. Proc. Natl Acad.
Sci. USA 95, 7231–7235 (1998).
44. Irwin, M. R., Pike, J. L., Cole, J. C. & Oxman, M. N.
Effects of a behavioral intervention, tai chi chih, on
varicella-zoster virus specific immunity and health
functioning in older adults. Psychosom. Med.
65, 824–830 (2003).
45. Glaser, R. et al. Stress-associated changes in the steady
state expression of latent Epstein–Barr Virus: implications
for chronic fatigue syndrome and cancer. Brain Behav.
Immun. 19, 91–103 (2005).
46. Glaser, R., Kutz, L. A. & Malarkey, W. B. Hormonal
modulation of Epstein–Barr virus replication.
Neuroendocrinology 62, 356–361 (1995).
47. Glaser, R. et al. The differential impact of training stress
and final examination stress on herpesvirus latency at the
United States Military Academy at West Point. Brain
Behav. Immun. 13, 240–251 (1999).
48. Hübner, G. et al. Differential regulation of pro-inflammatory
cytokines during wound healing in normal and
glucocorticoid-treated mice. Cytokine 8, 548–556 (1996).
49. Werner, S. & Grose, R. Regulation of wound healing by
growth factors and cytokines. Physiol. Rev. 83, 835–870
(2003).
50. Kiecolt-Glaser, J. K., Marucha, P. T., Malarkey, W. B.,
Mercado, A. M. & Glaser, R. Slowing of wound healing
by psychological stress. Lancet 346, 1194–1196 (1995).
51. Marucha, P. T., Kiecolt-Glaser, J. K. & Favagehi, M.
Mucosal wound healing is impaired by examination
stress. Psychosom. Med. 60, 362–365 (1998).
52. Padgett, D. A., Marucha, P. T. & Sheridan, J. F. Restraint
stress slows cutaneous wound healing in mice. Brain
Behav. Immun. 12, 64–73 (1998).
53. DeRijk, R. et al. Exercise and circadian rhythm-induced
variations in plasma cortisol differentially regulate
interleukin-1β (IL-1β), IL-6, and tumor necrosis factor-α
(TNF-α) production in humans: high sensitivity of TNF-α
and resistance of IL-6. J. Clin. Endocrinol. Metab.
82, 2182–2192 (1997).
54. Glaser, R. et al. Stress-related changes in
proinflammatory cytokine production in wounds.
Arch. Gen. Psychiatry 56, 450–456 (1999).
55. Follin, P. Skin chamber technique for study of in vivo
exudated human neutrophils. J. Immunol. Methods
232, 55–65 (1999).
56. Contrada, R. J., Leventhal, E. A. & Anderson, J. R.
in International Review of Health Psychology Vol. 3
(eds Maes, S., Leventhal, H. & Johnston, M.) 219–266
(Wiley & Sons, Chichester, 1994).
57. Kiecolt-Glaser, J. K., Page, G. G., Marucha, P. T.,
MacCallum, R. C. & Glaser, R. Psychological influences
on surgical recovery: perspectives from
psychoneuroimmunology. Am. Psychol. 53, 1209–1218
(1998).
58. Black, P. H. Stress and the inflammatory response:
a review of neurogenic inflammation. Brain Behav.
Immun. 16, 622–653 (2002).
59. Zhou, D., Kusnecov, A. W., Shurin, M. R., DePaoli, M. &
Rabin, B. S. Exposure to physical and psychological
stressors elevates plasma interleukin 6: relationship to
the activation of hypothalamic–pituitary–adrenal axis.
Endocrinology 133, 2523–2530 (1993).
60. Penninx, B. W. et al. Inflammatory markers and
depressed mood in older persons: results from the
health, aging, and body composition study. Biol.
Psychiatry 54, 566–572 (2003).
61. Glaser, R., Robles, T., Sheridan, J., Malarkey, W. B. &
Kiecolt-Glaser, J. K. Mild depressive symptoms are
associated with amplified and prolonged inflammatory
responses following influenza vaccination in older adults.
Arch. Gen. Psychiatry 60, 1009–1014 (2003).
62. Maes, M., Ombelet, W., De Jongh, R., Kenis, G. &
Bosmans, E. The inflammatory response following
delivery is amplified in women who previously suffered
from major depression, suggesting that major depression
is accompanied by a sensitization of the inflammatory
response system. J. Affect. Disord. 63, 85–92 (2001).

www.nature.com/reviews/immunol

© 2005 Nature Publishing Group

PERSPECTIVES
63. Johnson, J. D. et al. Prior stressor exposure sensitizes
LPS-induced cytokine production. Brain Behav. Immun.
16, 461–476 (2002).
64. Harris, T. et al. Associations of elevated interleukin-6
and C-reactive protein levels with mortality in the elderly.
Am. J. Med. 106, 506–512 (1999).
65. Kiecolt-Glaser, J. K. et al. Chronic stress and age-related
increases in the proinflammatory cytokine IL-6. Proc. Natl
Acad. Sci. USA 100, 9090–9095 (2003).
66. Epel, E. S. et al. Accelerated telomere shortening
in response to life stress. Proc. Natl Acad. Sci. USA
101, 17312–17315 (2004).
67. Schulz, R. & Beach, S. R. Caregiving as a risk factor
for mortality: the caregiver health effects study. JAMA
282, 2215–2219 (1999).
68. Marx, J. Inflammation and cancer: the link grows
stronger. Science 306, 966–968 (2004).
69. Reiche, E. M. V., Nunes, S. O. V. & Morimoto, H. K.
Stress, depression, the immune system, and cancer.
Lancet Oncol. 5, 617–625 (2004).
70. Coe, C. L. & Lubach, G. R. Critical periods of special
health relevance for psychoneuroimmunology. Brain
Behav. Immun. 17, 3–12 (2003).
71. Antoni, M. H. et al. Cognitive-behavioral stress
management intervention effects on anxiety, 24-hr urinary
norepinephrine output, and T-cytotoxic/suppressor cells
over time among symptomatic HIV-infected gay men.
J. Consult. Clin. Psychol. 68, 31–45 (2000).
72. Fawzy, F. I. et al. Malignant melanoma: effects of an early
structured psychiatric intervention, coping, and affective
state on recurrence and survival 6 years later. Arch. Gen.
Psychiatry 50, 681–689 (1993).
73. Kiecolt-Glaser, J. K. et al. Psychosocial enhancement
of immunocompetence in a geriatric population. Health
Psychol. 4, 25–41 (1985).
74. Andersen, B. L. et al. Psychological, behavioral, and
immune changes after a psychological intervention: a
clinical trial. J. Clin. Oncol. 22, 3570–3580 (2004).
75. Bonneau, R. H., Sheridan, J. F., Feng, N. & Glaser, R.
Stress-induced suppression of herpes simplex virus
(HSV)-specific cytotoxic T lymphocyte and natural killer
cell activity and enhancement of acute pathogenesis
following local HSV infection. Brain Behav. Immun.
5, 170–192 (1991).
76. Johnson, R. R., Storts, R., Welsh T. H. Jr, Welsh, C. J. &
Meagher, M. W. Social stress alters the severity of acute
Theiler’s virus infection. J. Neuroimmunol. 148, 74–85
(2004).
77. Sieve, A. N. et al. Chronic restraint stress during early
Theiler’s virus infection exacerbates the subsequent
demyelinating disease in SJL mice. J. Neuroimmunol.
155, 103–118 (2004).
78. Cao, L., Hudson, C. A. & Lawrence, D. A. Acute
cold/restraint stress inhibits host resistance to Listeria
monocytogenes via β1-adrenergic receptors. Brain
Behav. Immun. 17, 121–133 (2003).
79. Cao, L., Filipov, N. M. & Lawrence, D. A. Sympathetic
nervous system plays a major role in acute
cold/restraint stress inhibition of host resistance
to Listeria monocytogenes. J. Neuroimmunol.
125, 94–102 (2002).
80. Campbell, T. et al. Neuropathogenesis of Theiler’s
virus infection: I. Acute disease. Brain Behav. Immun.
15, 235–254 (2001).
81. Mehta, S. K., Pierson, D. L., Cooley, H., Dubow, R. &
Lugg, D. Epstein–Barr virus reactivation associated with
diminished cell-mediated immunity in Antarctic
expeditioners. J. Med. Virol. 61, 235–240 (2000).
82. Ironson, G. et al. Posttraumatic stress symptoms,
intrusive thoughts, loss and immune function after
Hurricane Andrew. Psychosom. Med. 59, 128–141
(1997).
83. Delahanty, D. L., Dougall, A. L. & Baum, A. in
Psychoneuroimmunology 3rd edn Vol. 2 (eds
Ader, R., Felten, D. L. & Cohen, N.) 335–348
(Academic, San Diego, 2001).
84. Lundberg, U. & Frankenhaeuser, M. Stress and workload
of men and women in high-ranking positions. J. Occup.
Health Psychol. 4, 142–151 (1999).
85. Dhabhar, F. S. Stress, leukocyte trafficking, and the
augmentation of skin immune function. Ann. NY Acad.
Sci. 992, 1–13 (2003).
86. Altemus, M., Rao, B., Dhabhar, F. S., Ding, W. &
Granstein, R. D. Stress-induced changes in skin
barrier function in healthy women. J. Invest. Dermatol.
117, 309–317 (2001).
87. Miller, A. H. et al. Glucocorticoid receptors are
differentially expressed in the cells and tissues of
the immune system. Cell. Immunol. 186, 45–54
(1998).

88. Elenkov, I. J. & Chrousos, G. P. Stress hormones,
proinflammatory and anti-inflammatory cytokines,
and autoimmunity. Ann. NY Acad. Sci. 966, 290–303
(2002).
89. Lai, J. P., Douglas, S. D. & Ho, W. Z. Human lymphocytes
express substance P and its receptor. J. Neuroimmunol.
86, 80–86 (1998).
90. Petitto, J. M., Huang, Z. & McCarthy, D. B. Molecular
cloning of NPY-Y1 receptor cDNA from rat splenic
lymphocytes: evidence of low levels of mRNA expression
and 125I-NPY binding sites. J. Neuroimmunol. 54, 81–86
(1994).
91. Radulovic, M., Dautzenberg, F. M., Sydow, S.,
Radulovic, J. & Spiess, J. Corticotropin-releasing factor
receptor 1 in mouse spleen: expression after immune
stimulation and identification of receptor-bearing cells.
J. Immunol. 162, 3013–3021 (1999).
92. Clevenger, C. V., Freier, D. O. & Kline, J. B. Prolactin
receptor signal transduction in cells of the immune
system. J. Endocrinol. 157, 187–197 (1998).
93. Sun R., Li, A. L., Wei, H. M. & Tian, Z. G. Expression
of prolactin receptor and response to prolactin
stimulation of human NK cell lines. Cell Res.
14, 67–73 (2004).
94. Welniak, L. A., Sun, R. & Murphy, W. J. The role of
growth hormone in T-cell development and
reconstitution. J. Leukoc. Biol. 71, 381–387 (2002).
95. Sanders, V. M., Kasprowicz, D. J., Kohm, A. P. &
Swanson, M. A. in Psychoneuroimmunology 3rd edn
Vol. 1 (eds Ader, R., Felten, D. L. & Cohen, N.) 132–158
(Academic, San Diego, 2001).

96. Cloez-Tayarani, I., Petit-Bertron, A.-F., Venters, H. D. &
Cavaillon, J.-M. Differential effect of serotonin on cytokine
production in lipopolysaccharide-stimulated human
peripheral blood mononuclear cells: involvement of
5–hydroxytryptamine 2A receptors. Int. Immunol. 15,
233–240 (2003).

Acknowledgements
We thank J. Sheridan, D. Padgett, R. Bonneau, R. Nelson,
N. Quan and V. Sanders for helpful suggestions. Work on this
paper was supported, in part, by grants from the General
Clinical Research Center (Columbus, United States) and the
Comprehensive Cancer Center, (Columbus, United States).

Competing interests statement
The authors declare no competing financial interests.

Online links
DATABASES
The following terms in this article are linked online to:
Entrez Gene:
http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=gene
ACTH | CCL2 | CRH | CRP | CXCL8 | glucocorticoid receptor |
growth hormone | IFN-γ | IL-1α | IL-1β | IL-2 | IL-6 | PDGFs |
prolactin | TGF-β | TNF | VEGF
FURTHER INFORMATION
PNI Research Program: http://pni.psychiatry.ohiostate.edu/jkg/
Access to this interactive links box is free online.

OPINION

Consensual immunity: successdriven development of T-helper-1
and T-helper-2 responses
Pawel Kalinski and Muriel Moser
Abstract | Non-germline-encoded T- and
B-cell receptors allow humans to effectively
deal with rapidly mutating pathogens. Here,
we argue that, in addition to determining the
antigenic specificity of immune responses,
the same receptor systems can also regulate
the T-helper-1/T-helper-2 profile of immunity.
Such a mechanism — based on feedback
from distinct effector cells to dendritic cells,
rather than on instruction from pathogens —
uses the effectiveness of particular effector
cells at targeting and destroying a pathogen
as a reliable, experience-based criterion to
induce and maintain the appropriately
polarized response.

Distinct subsets of CD4+ T cells preferentially
support cell-mediated (type 1) versus
humoral (type 2) immunity1. Type 1 T helper
(TH1) cells promote the cytotoxic effector
functions of natural killer (NK) cells, CD8+
T cells and macrophages. They also promote
antibody-dependent cell-mediated cytotoxicity (ADCC) by supporting B-cell production
of IgG2a in mice and IgG1 in humans. By

NATURE REVIEWS | IMMUNOLOGY

contrast, TH2 cells promote humoral immunity, mediated by B-cell-produced IgG4 and
IgE in humans (and IgG1 and IgE in mice).
Although the proper balance of TH1 and TH2
immunity is as important for the success of
an immune response as its specificity and
overall magnitude1, it still remains unclear
how the TH1/TH2-response profile is matched
to distinct pathogens and to particular
affected tissues.
The previously identified ‘instructive’
mechanisms of the induction of TH1- versus
TH2-dominated responses by dendritic cells
(DCs) use germline-encoded receptors to
identify both distinct sets of conserved
pathogen-specific motifs and endogenous
mediators of tissue damage that are induced
by different pathogen types invading distinct
tissues2–6. Here, we discuss recent evidence for
the existence of an additional highly reliable
mechanism that assures the correctness of
such a match. We propose that the intrinsic
ability of different effector cells to discriminate between different pathogen classes and
to differentially affect DC functions is a

VOLUME 5 | MARCH 2005 | 2 5 1

© 2005 Nature Publishing Group

1
2
3
4
5
6
7

Exhibit G

8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

ADDENDUM TO MPAS IN SUPPORT OF AMICUS CURIAE
2:18-cv-00490-JAM-KJN

Psychological Trauma: Theory, Research, Practice, and Policy
2017, Vol. 9, No. 3, 352–361

© 2016 American Psychological Association
1942-9681/17/$12.00 http://dx.doi.org/10.1037/tra0000177

Trauma and Psychological Distress in Latino Citizen Children Following
Parental Detention and Deportation
Lisseth Rojas-Flores, Mari L. Clements,
and J. Hwang Koo

Judy London
Public Counsel’s Immigrants’ Rights Project, Los
Angeles, California

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Fuller Theological Seminary

The mental health impact of parental detention and deportation on citizen children is a topic of
increasing concern. Forced parent– child separation and parental loss are potentially traumatic events
(PTEs) with adverse effects on children’s mental health. Objective: This study examines posttraumatic stress disorder (PTSD) symptoms and psychological distress among 91 Latino U.S.-born
children (ages 6 to 12), living in mixed-status families with a least 1 undocumented parent at risk
for detention or deportation. Method: Multiagent (child, parent, teacher, clinician) and standardized
assessments were conducted at baseline to assess for child trauma and psychological distress.
Results: Analyses indicate that PTSD symptoms as reported by parent were significantly higher for
children of detained and deported parents compared to citizen children whose parents were either
legal permanent residents or undocumented without prior contact with immigration enforcement.
Similarly, findings revealed differences in child internalizing problems associated with parental
detention and deportation as reported by parent as well as differences in overall child functioning as
reported by clinician. In addition, teachers reported higher externalizing for children with more
exposure to PTEs. Conclusions: These findings lend support to a reconsideration and revision of
immigration enforcement practices to take into consideration the best interest of Latino citizen
children. Trauma-informed assessments and interventions are recommended for this special
population.
Keywords: PTSD, Latino children, citizen children, immigration, deportation, detention

health. Specifically, we examined U.S.-born Latino children’s
mental health, including PTSD and psychological distress, following parental detention or deportation.
Children of immigrants represent 25% of the 69.9 million children in the United States (Zong & Batalova, 2015). Over 88% of
immigrant-origin children (4.5 million) are U.S.-born with a
foreign-born parent (Passel, Cohn, Krogstad, & Gonzalez-Barrera,
2014). Many of these foreign-born parents are unauthorized immigrants at chronic risk of arrest, detention, and/or deportation.
Enforcement efforts have taken the form of worksite and home
raids that sweep undocumented immigrants from families and
communities. From 2002 to 2014, the Office of Immigration
Statistics (2013) reported record-high deportations. In just over 2
years (July 2010 to September 2012), nearly 250,000 parents of
citizen children were deported (Wessler, 2012). The majority of
the deportees had migrated from Latin American countries, including Mexico, Honduras, El Salvador, Guatemala, Cuba, and Brazil
(Office of Immigration Statistics, 2013).
Forced parent– child separation and parental loss are PTEs with
adverse effects on child mental health and academic functioning
(Finkelhor et al., 2009). Children may experience the loss or
potential loss of a parent as particularly traumatic if it occurs in the
context of contact with legal authorities, such as in the case of
incarceration or deportation. Parental incarceration, a recognized
PTE in childhood (Felitti, 2009), is distinguished from other adverse childhood experiences by the unique combination of trauma,
ambiguity, lack of social support, shame, and stigma (Hairston,

Adverse childhood experiences (Alegría, Green, McLaughlin &
Loder, 2015) and immigration status (Castañeda et al., 2015) are
important social determinants of mental disorders. In children,
potentially traumatic events (PTEs) may lead to the development
of posttraumatic stress disorder (PTSD; Finkelhor, Ormrod, &
Turner, 2009). PTSD has debilitating effects on child development
and functioning and is a costly public health issue (U.S. Department of Health & Human Services, 2003). This study examines the
intersection of parental immigration status and children’s mental

This article was published Online First August 8, 2016.
Lisseth Rojas-Flores, Mari L. Clements, and J. Hwang Koo, Graduate
School of Psychology, Fuller Theological Seminary; Judy London, Public
Counsel’s Immigrants’ Rights Project, Los Angeles, California.
This article was developed with a generous grant from the Foundation
for Child Development’s Young Scholars Program to the first author. The
views, policies, and opinions expressed are those of the authors and do not
necessarily reflect those of the Foundation. We would like to acknowledge
the many churches, community centers, and immigration advocacy centers
that referred families to us; our committed research assistants; and the
brave Latino citizen children and their families who have contributed to our
growing understanding of child traumatic stress and the impact of immigration enforcement policies on children.
Correspondence concerning this article should be addressed to
Lisseth Rojas-Flores, Graduate School of Psychology, Fuller Theological Seminary, 180 North Oakland Avenue, Pasadena, CA 91101.
E-mail: [e-mail]
352

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

PTSD AND LATINO CITIZEN CHILDREN

2007). Mounting evidence has indicated that arrest and imprisonment of a parent disrupts parent– child relationships, alters familial
support networks, and impairs children’s mental health (Roberts et
al., 2014). We speculate that the detention and deportation of
unauthorized parents may have similar unintended negative effects
on their U.S.-born progeny.
Emerging research has indicated that parental detention and
deportation increase risk for mental health problems such as severe
psychological distress, anxiety, and depression (Allen, Cisneros, &
Tellez, 2015; Zayas, Aguilar-Gaxiola, Yoon, & Rey, 2015); for
underutilization of care (Chen & Vargas-Bustamante, 2011); and
for involvement with Child Welfare (Rabin, 2011). A few empirical and qualitative studies have examined the effects of parental
legal status on child and adolescent development (e.g., Allen et al.,
2015; Brabeck & Xu, 2010; Dreby, 2012), but this research has
been largely descriptive or retrospective, relying primarily on
parent report of child outcomes (Allen et al., 2015; Brabeck & Xu,
2010; Chaudry et al., 2010).
To the best of our knowledge, only two empirical studies have
examined citizen children and their increased risk for psychological distress subsequent to parental detention or deportation (Allen
et al., 2015; Zayas et al., 2015). Allen and colleagues (2015)
recruited immigrant caregivers who either were in deportation
legal proceedings, had been deported, or were unauthorized without contact with immigration enforcement. In this sample of primarily U.S.-born children, Allen et al. found that children with a
deported parent exhibited more internalizing problems after controlling for trauma history than did children without a deported
parent.
In a recent binational study using child self-report, Zayas and
colleagues (2015) examined the psychological distress of three
groups of citizen children (ages 8 –15 years) who had at least one
parent of Mexican origin. The groups consisted of (a) children
living in Mexico with their deported parents, (b) children living in
the United States with parents affected by detention or deportation,
and (c) children living in the United States whose undocumented
parents were not affected by detention or deportation. Two significant group differences emerged. First, children with parental
history of detention or deportation reported possible attention
deficits. Second, citizen children living in Mexico with deported
parents displayed more depressive symptoms than did other children. Furthermore, all three groups scored within the range of
probable anxiety problems. Notably, no measures of trauma were
reported. To the best of our knowledge, no studies have systematically assessed child PTSD symptoms and overall psychological
distress in this vulnerable population using extrafamilial informants. Using multiple informants (i.e., child, parent, teacher, clinician) and standardized measures, the present study was designed
to examine the psychological impact of parental detention and
deportation on U.S.-born Latino children.
Children of unauthorized parents have been shown to be disproportionally poor and in disadvantaged neighborhoods at risk for
exposure to violence, victimization, and further marginalization
(e.g., Ross & Mirowsky, 2009). In fact, unauthorized status is
highly associated with poverty and low parental education (Yoshikawa, Kholoptseva, & Suárez-Orozco, 2013). Emerging evidence, however, has proposed that precarious parental immigration
status puts citizen children at risk for a gamut of socioemotional
disadvantages beyond the ill effects of poverty and related risk

353

factors (Yoshikawa et al., 2013). Immigration enforcement is a
multifaceted social issue, and its effects on Latino children’s
development need further research.

The Present Study
This study sought to build on prior research on the unintended
mental health consequences of immigration enforcement on Latino
citizen children. To address the intersectional nature of cumulative
risks, we included two comparison groups of citizen children
whose immigrant parents had no contact with U.S. Immigration
and Customs Enforcement (ICE): (a) children of unauthorized
parents with no history of detention or deportation and (b) children
of U.S. legal permanent residents (LPRs). We planned to control
for child lifetime exposure to PTEs and for maternal education as
the best indicator of family’s socioeconomic status (SES). Income
was not included as an SES indicator, because family income was
expected to be substantially reduced following parental detention
or deportation. We examined baseline multiple informant assessment data to test the central hypothesis that Latino U.S.-citizen
children whose parents have been detained and/or deported would
have significantly more psychological distress and PTSD symptoms than would children of parents who had no contact with ICE.

Method
Study Sample
From 2013 through early 2015, undocumented and legal permanent resident parents born in Mexico or Central America (e.g.,
Nicaragua, Honduras, El Salvador, Guatemala), regardless of race
or socioeconomic status, were recruited. Specifically, this study
targeted mixed-status Latino families with U.S.-born citizen children between ages 6 and 12 living in the Southwest. Citizen
children with a current major medical, neurological, or mental
health disorder (e.g., psychosis, autism, Down’s syndrome) were
excluded.

Procedures
Families with precarious legal status were recruited through a
broad network of trusted immigration advocacy agencies,
community-based programs, and churches that work with such
families. Three primary methods were used in recruiting mixedstatus families: (a) individual agency referral, (b) oral presentations at various community-based programs and Latino churches
serving the immigrant community, and (c) a short video advertising the study. Staff at these agencies identified potential study
participants. Using provided scripts, staff invited caregivers who
had at least one child who was born in the United States to
participate. Once a release of information was obtained, contact
information was passed on to the research staff, who then contacted potential participants by phone to explain the study, validate
the child’s age, and schedule the initial visit. Caregivers and
children were interviewed simultaneously in separate rooms at
trusted community agencies or churches. Interviewers were bilingual or bicultural (English or Spanish) master’s-level clinicians.
Interviews lasted approximately 2 hr, including snack breaks.

354

ROJAS-FLORES, CLEMENTS, HWANG KOO, AND LONDON

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Consent and assent forms were reviewed and signed, including
parental consent to obtain school records and to mail a survey to
the child’s teacher. Adult and child participants were informed that
they could choose not to answer any question or to stop the
interview at any time. Confidentiality was discussed, including the
exception for reporting child abuse and neglect. Given the vulnerable legal status of this study’s participants, a “certificate of
confidentiality” was deemed important and obtained. Participants
were compensated with $30 for parents or caregivers, $10 gift card
for teachers, and $15 gift cards for children. All parent and child
measures were available in Spanish and English and were read to
participants.

Measures
Child report. Children were assessed using the UCLA Posttraumatic Stress Disorder Reaction Index (UCLA PTSD-RI; Steinberg,
Brymer, Decker, & Pynoos, 2004). This 22-item, clinicianadministered measure is among the more extensively studied and
widely used assessments of childhood PTSD. The UCLA PTSD-RI
has strong convergent validity with the Diagnostic and Statistical
Manual of Mental Disorders (4th ed.; DSM–IV; American Psychiatric
Association, 1994) diagnosis criteria of experiencing a traumatic
event (Criterion A) and reporting symptoms related to reexperiencing/
intrusive thoughts (Criterion B), avoidance (Criterion C), and hyperarousal (Criterion D). This measure has excellent psychometric properties, with internal consistencies of .82 (Criterion B), .83 (Criterion
C), and .71 (Criterion D). It has been used across a variety of trauma
types, age ranges, settings, and languages, including Spanish (Rodriguez, Steinberg, & Pynoos, 1999; Steinberg et al., 2013). The UCLA
PTSD-RI provides PTSD symptom severity and screens for 13 PTEs
among children of 7–18 years of age, including accidents, physical
and sexual abuse, and domestic violence. In our sample, PTEs
were positively skewed (1.50), with observed scores ranging
from 0 to 8. To normalize this distribution, we recoded the five
scores above 4 to 4.
Children also completed the Center for Epidemiologic Studies
Depression Scale for Children (CES-DSC), a 20-item self-report
depression inventory with scores ranging from 0 to 60 and a
clinical cutoff of Ն15 (Weissman, Orvaschel, & Padian, 1980).
The Spanish version of the CES-DSC has been widely used in
epidemiological research (González et al., 2016). Cronbach alpha
in this study was .81.
Parent report. Parents completed the Behavior Assessment
System for Children–2nd Edition, Parent Rating Scales–Child
(BASC-2 PRS-C; Reynolds & Kamphaus, 2004). The BASC-2
PRS-C is a widely used and well-validated caregiver-report measure
of 160 items on a Likert-type scale ranging from 1 (never) to 4 (almost
always). It yields scores on a wide range of empirically based syndrome scales and two composite scales (Internalizing Problems and
Externalizing Problems). Scores are reported in T scores, and percentiles based on age-specific norms (clinical cutoff Ն70), standardized
using samples of clinical and nonclinical populations sampled to
reflect the general population (Reynolds & Kamphaus, 2004). The
Spanish version of the BASC-2 PRS-C has reliability and validity
support with Spanish-speaking parents (McCloskey, Hess, &
D’Amato, 2003). In the current study, composite score reliabilities for
the BASC-2 PRS-C Externalizing Problems and Internalizing Problems were strong, with Cronbach alphas of .88 and .76, respectively.

Parents also completed the Trauma Symptom Checklist for
Young Children—Spanish Version (TSCYC–SP; Briere, 2005), a
standardized 90-item caregiver report developed to assess traumarelated symptoms in children ages 3–12 (T scores with clinical
cutoff Ն70). The reliability and validity of the TSCYC–SP has
been established in a sample of outpatient children from Spanishspeaking families, with reported Cronbach alphas from .67 to .93
(Wherry et al., 2014). Reliability for the TSCYC–SP scales in the
current study were strong (alphas of .79 to .85).
Teacher report. Teachers completed the BASC-2 Teacher
Rating Scales–Child (BASC-2 TRS-C; Reynolds & Kamphaus,
2004). The BASC-2 TRS-C is a 139-item scale that evaluates
children’s behavioral and emotional functioning. Like the BASC-2
PRS-C, scale scores are reported as T scores, and percentiles are
based on normative data (clinical cutoff score Ն70). In this study,
composite score reliabilities for BASC-2 TRS-C Externalizing
Problems and Internalizing Problems were strong, with Cronbach
alphas of .89 and .82, respectively.
Clinician evaluation. Clinicians used the Child and Adolescent Functional Assessment Scale (CAFAS; Hodges, 2006) to rate
the child’s lowest level of day-to-day functioning across critical
life domains (School, Home, Community, Moods/Emotions, and
Total Dysfunction). Cutoff scores indicating severe, moderate, and
mild impairment are 30, 20, and 10, respectively. The CAFAS has
been widely used in community mental health across the United
States as part of statewide assessments of mental health outcomes
(Bates, 2001). After being trained to 80% agreement using CAFAS
training materials and assessment (Hodges, 2006), two master’slevel clinicians jointly rated each child participant on the basis of
information collected in the structured interviews with parent and
child, as well as the BASC-2 TRS-C scale (teacher report) and
school records.

Results
Descriptive data and correlations for main study variables are
presented in Table 1. Gender was not significantly related to
outcome variables, so it was dropped from all analyses. Surprisingly, neither maternal education nor family income was correlated
with most outcome variables. Higher maternal education was
associated with lower parental TSCYC–SP depression reports
(p ϭ .01), and lower income was significantly correlated with
more PTEs (p ϭ .03). Thus, these SES variables were included as
covariates in analyses of only those specific outcomes. Demographic characteristics of the participant children, grouped by
parental immigration status comparisons, are presented in Table 2.
As expected, groups significantly differed on family income (p ϭ
.002), with legal permanent resident (LPR) families reporting
significantly higher incomes than did either unauthorized group,
and families with a detained or deported parent having both lower
maternal education than did LPR families and more father unemployment than did either LPR or unauthorized without ICE contact
families (p ϭ .002 and p Ͻ .001, respectively).

Risk Exposure for PTEs
After controlling for family income, the groups significantly
differed on lifetime exposure to PTEs on the UCLA PTSD-RI
index by parental immigration status, as shown in Table 3, with

—
.04
.12
.11
Ϫ.19a
Ϫ.09
.00
.03
Ϫ.06
.03
.01
Ϫ.02
.01
Ϫ.01
.04
.13
Ϫ.07
.61
.49
97

Sex
Educationb
Incomec
PTE
PTSD
CES-DSC
PTS total
Anxiety
Depression
Parent int
Parent ext
Teacher int
Teacher ext
Total dys
Home
School
Mood

a

—
.10
.02
.05
.01
Ϫ.17
Ϫ.19a
Ϫ.25‫ء‬
Ϫ.12
.10
.08
.00
Ϫ.19a
Ϫ.08
Ϫ.07
Ϫ.20a
.42
.50
95

2

—
.22‫ء‬
.10
.06
.01
.05
Ϫ.04
.09
.11
Ϫ.04
Ϫ.13
.11
.10
Ϫ.01
.09
1.27
1.35
95

3

—
.26‫ءء‬
.26‫ءء‬
.08
.07
.06
.04
.34‫ءء‬
Ϫ.07
.21
.20
.10
.18
.22‫ء‬
1.69
1.27
97

4

—
.52‫ءءء‬
.02
Ϫ.08
Ϫ.02
Ϫ.03
.21‫ء‬
.01
.10
.21‫ء‬
.05
.02
.42‫ءءء‬
20.40
16.01
96

5

—
.08
Ϫ.06
.08
.04
.13
Ϫ.03
Ϫ.07
.17
.04
.05
.31‫ءء‬
22.97
10.12
97

6

—
.71‫ءءء‬
.77‫ءءء‬
.40‫ءءء‬
.27‫ءء‬
Ϫ.13
Ϫ.03
.47‫ءءء‬
.26‫ء‬
.49‫ءءء‬
.35‫ءءء‬
51.72
10.68
92

7

—
.59‫ءءء‬
.32‫ءء‬
.07
Ϫ.14
Ϫ.01
.43‫ءءء‬
.16
.37‫ءءء‬
.41‫ءءء‬
54.51
11.89
92

8

—
.54‫ءءء‬
.27‫ءء‬
Ϫ.12
.01
.53‫ءءء‬
.35‫ءءء‬
.46‫ءءء‬
.33‫ءء‬
50.32
10.82
92

9

—
.48‫ءءء‬
Ϫ.03
.03
.16
.24‫ء‬
.08
.06
51.16
11.43
96

10

—
.22‫ء‬
.38‫ءءء‬
.42‫ءءء‬
.57‫ءءء‬
.31‫ءء‬
.19a
47.41
10.49
96

11

—
.55‫ءءء‬
Ϫ.02
Ϫ.01
.15
Ϫ.14
48.17
10.56
83

12

—
.27‫ء‬
.27‫ء‬
.37‫ءءء‬
.10
49.49
10.51
83

13

—
.77‫ءءء‬
.57‫ءءء‬
.75‫ءءء‬
13.59
19.08
92

14

—
.26‫ء‬
.37‫ءءء‬
1.74
4.83
92

15

—
.29‫ءء‬
2.07
5.04
92

16

—
6.52
7.77
92

17

Note. PTE ϭ potentially traumatic event; PTSD ϭ UCLA PTSD Reaction Index total score; CES-DSC ϭ Center for Epidemiological Studies Depression Scale for Children; PTS total ϭ Trauma
Symptom Checklist for Young Children (TSCYC) posttraumatic stress total score; Anxiety ϭ TSCYC anxiety score; Depression ϭ TSCYC depression score; Parent int ϭ Behavior Assessment System
for Children–2nd Edition, Parent Rating Scales–Child (BASC-2 PRS-C) internalizing score; Parent ext ϭ BASC-2 PRS-C externalizing score; Teacher int ϭ BASC-2 Teacher Rating Scales–Child
(BASC-2 TRS-C) internalizing score; Teacher ext ϭ BASC-2 TRS-C externalizing score; Total dys ϭ clinician report of Total Dysfunction score from the Child and Adolescent Functional Assessment
Scale (CAFAS); Home ϭ clinician report of Home Behavior score from the CAFAS; School ϭ clinician report of School Behavior score from the CAFAS; Mood ϭ clinician report of Mood score
from the CAFAS).
a
Child gender (0 ϭ female, 1 ϭ male). b Maternal education (0 ϭ less than high school, 1 ϭ high school, GED, or some college). c Parental income (1 ϭ less than $15,000, 2 ϭ $15,000 –34,999,
3 ϭ $35,000 or more).
‫ء‬
p Ͻ .05. ‫ ءء‬p Ͻ .01. ‫ ءءء‬p Ͻ .001.

1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
M
SD
n

1

Variable

Table 1
Bivariate Correlations, Means, and Standard Deviations for Major Study Variables

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

PTSD AND LATINO CITIZEN CHILDREN

355

ROJAS-FLORES, CLEMENTS, HWANG KOO, AND LONDON

356

Table 2
Demographic Characteristics of Citizen Children and Their Families
Parental immigration status

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Variable
Child sex
Male
Female
Mother education
ϽHigh school
High school or higher
Family income
Ͻ$15,000
$15,000–$34,999
Ն$35,000
Father current employment
Full-time
Part-time
Unemployed
Mother current employment
Full-time
Part-time
Unemployed
Parents’ marital status
Married
Never married
Father’s country of origina
Mexico
El Salvador
Guatemala
Honduras
Nicaragua
United States
Mother’s country of origina
Mexico
El Salvador
Guatemala
Honduras
Nicaragua
United States
Parent years in U.S.
M
SD
Child age
M
SD

Detained or
deported
(n ϭ 39)

Unauthorized no history
of detention or
deportation (n ϭ 42)

Legal permanent
resident
(n ϭ 16)

Inferential
statistic
2.13

21
18

29
13

9
7

29
9

22
19

4
12

19
16
4

24
13
3

2
7
7

4
3
29

28
8
3

11
3
1

9
12
16

5
10
27

4
5
7

23
14

33
8

14
2

23
9
4
3
0
0

28
3
6
1
1
2

10
2
2
0
0
1

14
3
4
10
0
6

33
2
4
2
0
1

8
3
3
0
0
2

12.69‫ءءء‬
17.11‫ءء‬

50.43‫ءءء‬

4.48

5.16†
2.16

9.89‫ءء‬

2.75†
14.79
9.91

17.67
5.41

20.00
8.42

9.05
1.82

9.12
2.03

8.63
1.78

.40

Note. Inferential statistics are ␹2 for count data and F(2, 94) for means.
a
Due to small cell counts, ␹2 was computed on Mexico versus Central America countries.
†
p Ͻ .10. ‫ ء‬p Ͻ .05. ‫ ءء‬p Ͻ .01. ‫ ءءء‬p Ͻ .001.

children of detained or deported parents experiencing significantly
more lifetime PTEs than did children of LPRs (p ϭ .02). Even
when child reports of parental deportation or detention as a PTE
were excluded, the groups significantly differed on PTEs, F(2,
2
91) ϭ 3.62, p ϭ .03, ␩p ϭ .07, again with children of detained or
deported parents reporting significantly more PTEs than did children of LPRs (p ϭ .03). Children of unauthorized parents with no
contact with ICE were not significantly different from children
with detained or deported parents but tended to have more PTEs
than did children of LPRs (p ϭ .06). Overall, exposure was high
across groups, with 35% of the sample reporting exposure to one
PTE, 21% to two, 14% to three, and 12% to four or more PTEs in
their lifetime, with an average exposure of 1.69 PTEs (SD ϭ 1.27).

PTSD and Psychological Distress
Child outcomes by parental immigration status were examined
in a series of univariate and multivariate analyses of variance
(ANOVAs), controlling for maternal education when indicated,
with Bonferroni pairwise post hoc comparisons (see Table 3).
Analyses were conducted both with and without controlling for
lifetime exposure to PTEs, and the results were essentially identical. Thus, for ease of interpretation, only analyses not controlling
for lifetime exposure are presented here.
Child report of PTSD symptoms. Per the UCLA PTSD-RI
child report, 29% of all child participants met criteria for full
(19%) or partial (10%) PTSD diagnoses. There were no significant

PTSD AND LATINO CITIZEN CHILDREN

357

Table 3
Citizen Children PTSD and Psychological Distress by Parental Immigration Status
Parental immigration status
Detained or deported
Measure

n

UCLA PTSD-RI
Lifetime PTEsa
Total severity score
Criterion B
Criterion C
Criterion D
CES-DCS

38

TSCYCb
PTS overall total
Anxiety
Depression
Anger/Aggression
PTS Intrusion
PTS Avoidance
PTS Arousal
Dissociation
BASC-2 PRS-C
Int Prob total
Anxiety
Depression
Somatization
Ext Prob total
Hyperactivity
Aggression
Conduct
problems

37

BASC-2 TRS-C
Int Prob total
Anxiety
Depression
Somatization
Ext Prob total
Hyperactivity
Aggression
Conduct
problems

33

CAFAS
Overall dysfunction
Home
School
Mood/Emotions

M

SD

35

Unauthorized no history of
detention or deportation

Legal permanent resident

n

n

M

SD

␩2
p

M

SD

1.19b
17.50
5.19
6.63
5.69
19.69

1.11
18.11
6.67
7.02
5.55
6.77

3.84‫ء‬
.43
.66
.18
.65
1.05

46.14b
48.07b
43.79b
45.93
47.36
47.36b
45.71b
44.93

5.35
6.04
4.04
3.85
6.59
5.40
5.92
3.36

9.70‫ءءء‬
7.73‫ءءء‬
10.21‫ءءء‬
4.67‫ء‬
5.90‫ءء‬
6.08‫ءء‬
9.96‫ءءء‬
1.23

.18
.15
.19
.10
.12
.12
.19
.03

45.53b
53.07
44.47b
41.33b
43.87
45.87
43.73
44.00

7.25
11.00
7.46
4.85
8.07
10.69
6.24
6.59

4.17‫ء‬
1.56
5.24‫ءء‬
3.14‫ء‬
1.66
.92
1.54
2.00

.08
.03
.10
.06
.03
.02
.03
.04

46.46
46.62
46.31
48.38
49.49
47.54
45.00
44.46

7.24
9.00
6.10
9.07
10.51
6.17
4.16
3.48

.55
.05
1.26
.36
2.66†
1.76
2.35
3.25‫ء‬

6.32
2.58
.00
6.17

13.41‫ءءء‬
1.60
4.86‫ءء‬
17.10‫ءءء‬

F

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Child reports
42
1.97a
21.92
6.92
7.68
7.32
23.95

1.33
14.39
4.45
6.54
4.99
11.33

16
1.67
20.12
5.90
7.79
6.43
23.32

1.23
16.92
5.84
7.13
4.99
9.94

.08
.01
.01
Ͻ.01
.01
.02

Parent reports
39
57.62a
60.57a
56.59a
51.59a
56.49a
55.62a
57.92a
48.22

12.40
14.44
11.76
8.37
14.27
11.09
11.69
8.00

54.67a
58.00
54.13a
49.05a
49.41
50.59
47.62
50.28

12.45
11.73
10.84
11.92
10.37
11.97
8.39
10.50

39

14
48.56b
51.54b
47.18b
47.15b
47.41b
49.44b
49.41b
47.69

7.64
7.97
8.64
5.23
8.13
7.83
8.18
6.74

49.90
54.12
49.17
46.69
46.81
48.64
44.98
47.74

10.78
11.45
10.84
9.69
11.15
11.73
9.16
11.59

42

15

Teacher reports
37
49.61
47.52
50.18
51.30
52.48
52.88
51.42
52.36

13.54
9.57
10.65
15.98
13.53
13.49
12.47
13.39

13
47.49
47.35
47.59
49.35
48.24
48.68
48.24
48.19

8.36
7.95
7.06
8.42
8.15
9.31
7.47
7.87

.01
Ͻ.01
.03
.01
.06
.04
.06
.08

Clinician reports
42
25.14a
2.86
4.00a
11.71a

23.44
6.23
6.95
8.22

15
7.38b
1.19
1.19b
3.33b

12.31
3.95
3.23
5.26

4.00b
.67
.00b
3.33b

.23
.03
.10
.28

Note. Means with differing subscripts are significantly different in Bonferroni corrected pairwise comparisons. PTSD ϭ posttraumatic stress disorder;
UCLA PTSD-RI ϭ UCLA PTSD Reaction Index; PTEs ϭ potentially traumatic events; CESD-DSC ϭ Center for Epidemiologic Studies Depression Scale
for Children; TSCYC ϭ Trauma Symptom Checklist for Young Children; PTS ϭ posttraumatic stress; BASC-2—PRS ϭ Behavior Assessment System
for Children—2nd Edition, Parent Rating Scales-Child; Int Prob ϭ Internalizing Problems; Ext Prob ϭ Externalizing Problems; BASC-2 TRS-C ϭ BASC-2
Teacher Rating Scales–Child; CAFAS ϭ Child and Adolescent Functional Assessment Scale.
a
Analysis of covariance (ANCOVA) controlling for family income was conducted. Raw means and standard deviations reported. b ANCOVA of the total
score and multivariate analysis of variance of the scale scores were conducted for TSCYC, controlling for maternal education. Raw means and standard
deviations are presented here.
†
p Ͻ .10. ‫ ء‬p Ͻ .05. ‫ ءء‬p Ͻ .01. ‫ ءءء‬p Ͻ .001.

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

358

ROJAS-FLORES, CLEMENTS, HWANG KOO, AND LONDON

differences by parental immigration status on total PTSD symp2
toms, F(2, 93) ϭ 0.43, p ϭ .65, ␩p ϭ .01, or individual criteria
2
(Wilks’s ␭ ϭ .96), F(6, 184) ϭ 0.68, p ϭ .67, ␩p ϭ .02.
Parent report of PTSD child symptoms. Five families had
invalid TSCYC–SP Response Level and Atypical Response Scale
scores and were dropped from TSCYC–SP analyses. Controlling
for maternal education, total symptoms differed by parental immi2
gration status, F(2, 86) ϭ 9.70, p Ͻ .001, ␩p ϭ .18. Similarly, a
multivariate analysis of covariance of TSCYC–SP Anxiety, Depression, Anger/Aggression, Intrusion, Avoidance, Arousal, and
Dissociation controlling for maternal education indicated a significant multivariate effect of parental immigration status (Wilks’s
␭ ϭ .64), F(12, 168) ϭ 3.53, p Ͻ .001, with significant univariate
effects for all scales except Dissociation. In each case, per parent
report, children of detained and deported parents demonstrated
higher levels of trauma symptoms than did children of either LPR
parents (ps ranging from Ͻ .001 to .03) or unauthorized parents
with no ICE contact (ps ranging from Ͻ .001 to .01; see Table 3).
Child reports of psychological distress. There was no effect
of parent immigration status on child self-reports of depression on
2
the CES-DSC, F(2, 94) ϭ 1.05, p ϭ .35, ␩p ϭ .02.
Parent report of child psychological distress. As shown in
Table 3, there was a significant univariate effect of parent immigration status on BASC-2 PRS-C total internalizing, F(2, 93) ϭ
2
4.17, p ϭ .02, ␩p ϭ .08, with children of detained or deported
parents reported to have more internalizing problems than did
children of LPRs (p ϭ .02). Multivariate analysis of variance
(MANOVA) analyses revealed a marginal main effect of parent
immigration status on the three BASC-2 PRS-C internalizing
subscales (Wilks’s ␭ ϭ .88), F(6, 182) ϭ 2.05, p ϭ .06, with
children of detained or deported parents scoring higher on Depression (p ϭ .009) and Somatization (p ϭ .04) per parent report than
did children of LPRs.
In contrast, there were no significant effects of parent immigration status on parent-reported total externalizing, F(2, 93) ϭ 1.67,
2
p ϭ .20, ␩p ϭ .03, or on Hyperactivity, Aggression, or Conduct
Problems (Wilks’s ␭ ϭ .95), F(8, 182) ϭ 0.83, p ϭ .55.
Teacher report of child psychological distress. Univariate
and MANOVA results demonstrated no significant effects for
parental immigration status on teacher BASC-2-TRS total inter2
nalizing score, F(2, 80) ϭ 0.55, p ϭ .58, ␩p ϭ .01, or on the
Anxiety, Depression, or Somatization scales (Wilks’s ␭ ϭ .95),
F(6, 156) ϭ 0.61, p ϭ .72 (see Table 3).
Teacher reports of total externalizing symptoms were only margin2
ally different by group, F(2, 80) ϭ 2.66, p ϭ .08, ␩p ϭ .06. A
MANOVA analysis of Hyperactivity, Aggression, and Conduct Problems was also not significant (Wilks’s ␭ ϭ .90), F(6, 156) ϭ 1.35,
p ϭ .24, but a significant univariate effect was found for Conduct
Problems, with children of detained or deported parents tending to
have more of these problems than did children of LPRs (p ϭ .06).
Clinician report of overall child functioning. There was a
significant univariate main effect for parental immigration status on
clinician’s CAFAS overall child dysfunction, F(2, 89) ϭ 13.41, p Ͻ
2
.001, ␩p ϭ .23, with children of detained or deported parents exhibiting poorer functioning than did children of both LPRs and unauthorized parents with no ICE contact (ps Ͻ .001). Similarly,
MANOVA analyses of the Home, School, and Moods scales revealed
a significant effect of parent immigration status (Wilks’s ␭ ϭ .69),
F(6, 176) ϭ 5.87, p Ͻ .001, with children of detained or deported

parents having poorer scores on School Behavior than did children of
LPRs (p ϭ .03) or unauthorized parents with no ICE contact (p ϭ
.04), and poorer scores on Moods than did children of either LPRs or
unauthorized parents with no ICE contact (ps Ͻ .001).

Multiple Informant Comparisons
The correlations between ratings of internalizing constructs
were typically modest. For instance, teacher, parent, and child
ratings of depression were uncorrelated, as were parent and child
reports of PTSD symptoms. In contrast, clinician CAFAS ratings
of Moods and Emotions were significantly correlated with both
parent TSCYC–SP Depression, r(92) ϭ .36, p Ͻ .001, and child
CES-DSC, r(92) ϭ .31, p ϭ .003. Adult reports of externalizing
problems were more consistently and strongly related, with correlations among parent and teacher BASC-2 Total Externalizing and
between BASC-2 and clinician CAFAS Home Behavior and
School Behavior ranging from .37 to .57 (all ps Ͻ .001). Direct
comparisons of means were possible for only parent and teacher
BASC-2 scores. Parents reported higher Total Internalizing than
did teachers, t(83) ϭ 3.70, p ϭ .03, but reports of Total Externalizing did not differ, t(83) ϭ Ϫ1.40, p ϭ .17.

Discussion
The need to detect children with PTSD-related symptoms and
psychological distress is pertinent given the evidence for the
detrimental effects of early childhood adversity in the overall
mental health of children. This is significant for Latino citizen
children whose parents are undocumented and at high risk for
detention or deportation, which often lead to forced parent– child
separation. In light of complex immigration policies, our findings
provide some support for the need for clinical and public policy
interventions on behalf of this vulnerable child population.

Impact of Parental Detention or Deportation on
Citizen Children’s PTSD
Taken together, the reports of multiple informants (parent,
teacher, clinician, and child) indicate that citizen children of detained and deported parents experience more psychological distress
and trauma compared to peers whose parents had no involvement
with immigration enforcement. Higher levels of parent-reported
PTSD symptoms in children of detained and deported parents
imply that forced parental separation resulting from immigration
enforcement is particularly detrimental to children’s mental health.
The unpredictability and uncertainty associated with such separations may have exacerbated PTSD symptoms (see Grillon et al.,
2009). As such, our findings suggest that the current and heightened enforcement of immigration laws poses a serious public
health challenge to U.S.-born children of undocumented parents.
Not only is PTSD recognized as a high priority public health issue
(U.S. Department of Health & Human Services, 2003), but child
PTEs, such as losing a parent, pose serious risks for lifelong
mental and medical illnesses (Felitti, 2009; Putnam, Harris, &
Putnam, 2013).
Specifically, the children of detained and deported parents were
rated on the TSCYC–SP as endorsing more symptoms in all three
DSM–IV PTSD criterion domains as well in total posttraumatic

PTSD AND LATINO CITIZEN CHILDREN

symptoms. Although to the best of our knowledge, no other study
has used the TSCYC–SP with this population, our results seem
congruent with those of prior studies reporting on the validity of
the TSCYC–SP (e.g., Wherry et al., 2014). TSCYC–SP scores
were significantly correlated with BASC-2 internalizing problems
but not with child self-ratings, consistent with research documenting divergence in such child and parent reports (Briere, 2005; De
Los Reyes & Kazdin, 2005).

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Impact of Parental Detention or Deportation on
Citizen Child Psychological Distress
Children of detained or deported parents were rated by parents
and clinicians as higher in internalizing problems and in negative
moods and emotions compared to children of LPRs and parents
who had no contact with ICE. The overlap of depressive and
anxious symptoms with PTSD is significant, and thus these findings are consistent with the findings of prior empirical research in
showing significantly increased rates of depression and anxiety
problems among children with PTSD symptomatology (Samuelson, Krueger, Burnett, & Wilson, 2010). Depression and anxiety
pose immediate developmental challenges to child functioning
(Kendall et al., 2010) and pose higher risk for future mental health
problems (Lopez, Turner, & Saavedra, 2005). Furthermore, our
findings corroborate and extend those of Zayas et al. (2015) and
Allen et al. (2015) and are also congruent with findings of previous
studies documenting the negative mental health outcomes associated with parental separation (Chaudry et al., 2010; SuárezOrozco, Bang, & Kim, 2011).
Children with more PTEs were also rated by parents and teachers as having more externalizing problems (BASC-2) and by
clinicians as having more total dysfunction (CAFAS). This finding
aligns with previous research showing that children with traumarelated symptoms are at risk of misdiagnosis (e.g., with attention
deficit/hyperactivity disorder or conduct difficulties), particularly
in the absence of assessments for complex trauma (e.g., Kletzka &
Siegfried, 2008). Our findings underscore the need for educating
parents and teachers on symptoms associated with PTEs.

Intersection Between Poverty, Exposure to PTEs, and
the Loss of a Parent
Exposure to multiple PTEs was common across our sample,
with 35% of the children reporting experiencing one PTE and 47%
endorsing two or more PTEs. This high prevalence of PTE exposure is concerning given the negative short- and long-term consequences of childhood PTE exposure (Appleyard, Egeland, van
Dulmen, & Sroufe, 2005). Consistent with the literature, more
PTEs were related to increased child PTSD scores. Although there
were no differences by parental immigration status, the PTSD
prevalence in our sample was high (19% of the children meeting
all DSM–IV criteria and 10% meeting partial criteria) per child
report.
Emerging research in childhood adversity describes synergy as
the interaction of two or more PTEs, or adverse events, so that
their combined effect is greater than the sum of their individual
effects (Putnam et al., 2013). Putnam and colleagues (2013) documented the synergy of adverse events with loss of a parent among
adult males with three or more PTEs. They found that poverty, the

359

most potent adverse childhood event in males, is synergistic with
the loss of a parent. Putnam and colleagues’ research is particularly
relevant to citizen children of detained or deported parents who
have lost, or have the impending possibility of losing, a parent due
to U.S. immigration enforcement. These findings thus call for a
reconsideration and reduction of unnecessary detainment of undocumented parents and consequent parent– child separation.

Implications for Health Services, Policies, and
Future Directions
Researchers have argued that child PTEs are the most preventable causes of debilitating mental illnesses, such as PTSD, depression, and anxiety (Finkelhor, Ormrod, & Turner, 2007; National
Research Council & Institute of Medicine, 2009). Particularly for
children who have been multiply victimized, preventing future
PTEs may be the most effective intervention (Finkelhor et al.,
2007). This is notable for our sample of citizen children. A call for
action to prevent forced parental separation and constant threat of
potential loss of a parent due to immigration enforcement is
gravely needed.
Given the high endorsement of PTEs in our sample, more
trauma-informed, developmentally appropriate systems placed at
multiple levels (e.g., home, school) would assist Latino citizen
children and their families. Trauma-informed intervention and
prevention programs for this vulnerable population should target
synergistic adverse events, such poverty and loss of a parent.
Furthermore, affordable and culturally relevant services are warranted not only for children of detained or deported parents but
also for citizen children of parents living in the shadows. A
reevaluation of immigration policies that have significant effects
on access to health services is also extremely relevant to the
well-being of Latino citizen children (for a review see Rodríguez,
Young, & Wallace, 2015).
On the basis of findings with children of incarcerated parents
(Roberts et al., 2014), we suspect that witnessing parental detainment may be particularly detrimental. Future research should investigate the effects of witnessing the arrest or detention of undocumented parents on child PTSD symptoms, and this information
should be used in reviewing policies involving undocumented immigrants with children. Arrest protocols should consider the children’s
best interest.

Study Strengths and Limitations
The cross-sectional nature of this study and its relatively small
sample size limit the ability to infer causation and to generalize
findings to other ethnic and racial immigrant groups. Future studies should also examine South American Latino groups. Statistics
show undocumented South American immigrants tend to fare
better economically in the United States, in part due to higher
levels of education and different migratory routes than for immigrants coming from the Central American cone (Stoney, Batalova
& Russell, 2013). Such SES dynamics would be important to
understand in their interaction with immigration enforcement and
child well-being.
Finally, some children exposed to PTEs, including parental
detention or deportation, do not exhibit high levels of mental

360

ROJAS-FLORES, CLEMENTS, HWANG KOO, AND LONDON

health symptoms. These findings highlight an underlying resilience in the face of adversity that should be understood and
supported in all children of immigrants, regardless of parental legal
status. Future research should explore mediating factors, such as
family or community support, religious coping, hope, and cognitive processing of PTEs.

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

References
Alegría, M., Green, G. J., McLaughlin, K., & Loder, S. (2015). Disparities
in child and adolescent mental health and mental health services in the
U.S. Retrieved from https://philanthropynewyork.org/sites/default/files/
resources/Disparities_in_child_and_adolescent_health.pdf
Allen, B., Cisneros, E. M., & Tellez, A. (2015). The children left behind:
The impact of parental deportation on mental health. Journal of Child
and Family Studies, 24, 386 –392. http://dx.doi.org/10.1007/s10826013-9848-5
American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author.
Appleyard, K., Egeland, B., van Dulmen, M. H., & Sroufe, L. A. (2005).
When more is not better: The role of cumulative risk in child behavior
outcomes. Journal of Child Psychology and Psychiatry, 46, 235–245.
http://dx.doi.org/10.1111/j.1469-7610.2004.00351.x
Bates, M. P. (2001). The Child and Adolescent Functional Assessment
Scale (CAFAS): Review and current status. Clinical Child and Family
Psychology Review, 4, 63– 84. http://dx.doi.org/10.1023/A:1009
528727345
Brabeck, K. M., & Xu, Q. (2010). The impact of detention and deportation
on Latino immigrant children and families: A quantitative exploration.
Hispanic Journal of Behavioral Sciences, 32, 341–361. http://dx.doi.org/
10.1177/0739986310374053
Briere, J. (2005). Trauma Symptom Checklist for Young Children
(TSCYC): Professional manual. Odessa, FL: Psychological Assessment
Resources.
Castañeda, H., Holmes, S. M., Madrigal, D. S., Young, M. E., Beyeler, N.,
& Quesada, J. (2015). Immigration as a social determinant of health.
Annual Review of Public Health, 36, 375–392. http://dx.doi.org/10.1146/
annurev-publhealth-032013-182419
Chaudry, A., Capps, R., Pedroza, J. M., Castaneda, R. M., Santos, R., & Scott,
M. M. (2010). Facing our future: Children in the aftermath of immigration
enforcement. Retrieved from http://www.urban.org/uploadedpdf/
412020_FacingOurFuture_final.pdf
Chen, J., & Vargas-Bustamante, A. (2011). Estimating the effects of
immigration status on mental health care utilizations in the United
States. Journal of Immigrant and Minority Health, 13, 671– 680. http://
dx.doi.org/10.1007/s10903-011-9445-x
De Los Reyes, A., & Kazdin, A. E. (2005). Informant discrepancies in the
assessment of childhood psychopathology: A critical review, theoretical
framework, and recommendations for further study. Psychological Bulletin, 131, 483–509. http://dx.doi.org/10.1037/0033-2909.131.4.483
Dreby, J. (2012). The burden of deportation on children in Mexican
immigrant families. Journal of Marriage and Family, 74, 829 – 845.
http://dx.doi.org/10.1111/j.1741-3737.2012.00989.x
Felitti, V. J. (2009). Adverse childhood experiences and adult health.
Academic Pediatrics, 9, 131–132. http://dx.doi.org/10.1016/j.acap.2009
.03.001
Finkelhor, D., Ormrod, R. K., & Turner, H. A. (2007). Re-victimization
patterns in a national longitudinal sample of children and youth. Child
Abuse & Neglect, 31, 479 –502. http://dx.doi.org/10.1016/j.chiabu.2006
.03.012
Finkelhor, D., Ormrod, R. K., & Turner, H. A. (2009). Lifetime assessment
of poly-victimization in a national sample of children and youth. Child
Abuse and Neglect, 33, 403– 411. http://dx.doi.org/10.1016/j.chiabu
.2008.09.012

González, P., Nuñez, A., Merz, E., Brintz, C., Weitzman, O., Navas, E. L.,
. . . Gallo, L. C. (2016). Measurement from properties of the Center for
Epidemiologic Studies Depression Scale (CES-D 10): Findings HCHS/
SOL. Psychological Assessment. Advance online publication. http://dx
.doi.org/10.1037/pas0000330
Grillon, C., Pine, D. S., Lissek, S., Rabin, S., Bonne, O., & Vythilingam,
M. (2009). Increased anxiety during anticipation of unpredictable aversive stimuli in posttraumatic stress disorder but not in generalized
anxiety disorder. Biological Psychiatry, 66, 47–53. http://dx.doi.org/10
.1016/j.biopsych.2008.12.028
Hairston, C. F. (2007). Focus on the children with incarcerated parents:
An overview of the research literature. Retrieved from http://www.f2f
.ca.gov/res/pdf/FocusOnChildrenWith.pdf
Hodges, K. (2006). CAFAS self-training manual and blank scoring forms.
Ypsilanti, MI: Eastern Michigan University.
Kendall, P. C., Compton, S. N., Walkup, J. T., Birmaher, B., Albano,
A. M., Sherrill, J., . . . Piacentini, J. (2010). Clinical characteristics of
anxiety disordered youth. Journal of Anxiety Disorders, 24, 360 –365.
http://dx.doi.org/10.1016/j.janxdis.2010.01.009
Kletzka, N. T., & Siegfried, C. (2008). Helping children in child welfare
systems heal from trauma: A systems integration approach. Juvenile &
Family Court Journal, 59, 7–20. http://dx.doi.org/10.1111/j.1755-6988
.2008.00018.x
Lopez, B., Turner, R. J., & Saavedra, L. M. (2005). Anxiety and risk for
substance dependence among late adolescents/young adults. Journal of
Anxiety Disorders, 19, 275–294. http://dx.doi.org/10.1016/j.janxdis
.2004.03.001
McCloskey, D. M., Hess, R. S., & D’Amato, R. C. (2003). Evaluating the
utility of the Spanish version of the Behavior Assessment System for
Children-Parent Report System. Journal of Psychoeducational Assessment, 21, 325–337. http://dx.doi.org/10.1177/073428290302100402
National Research Council & Institute of Medicine. (2009). Preventing
mental, emotional, and behavioral disorders among young people:
Progress and possibilities. Washington, DC: The National Academies
Press.
Office of Immigration Statistics. (2013). Immigration enforcement actions:
2013. Retrieved from http://www.dhs.gov/sites/default/files/publications/
ois_enforcement_ar_2013.pdf
Passel, J. S., Cohn, D., Krogstad, J. M., & Gonzalez-Barrera, A. (2014). As
growth stalls, unauthorized immigrant population becomes more settled.
Retrieved from http://www.pewhispanic.org/2014/09/03/as-growthstalls-unauthorized-immigrant-population-becomes-more-settled/
Putnam, K. T., Harris, W. W., & Putnam, F. W. (2013). Synergistic
childhood adversities and complex adult psychopathology. Journal of
Traumatic Stress, 26, 435– 442. http://dx.doi.org/10.1002/jts.21833
Rabin, N. (2011). Disappearing parents: Immigration enforcement and the
child welfare system (Arizona Legal Studies Discussion Paper No.
11–26). Connecticut Law Review, 44, No. 1. Retrieved from http://ssrn
.com/abstractϭ1794263
Reynolds, C. R., & Kamphaus, R. W. (2004). Behavior Assessment System
for Children (BASC-2; 2nd ed.). Bloomington, MN: Pearson.
Roberts, Y. H., Snyder, F. J., Kaufman, J. S., Finley, M. K., Griffin, A.,
Anderson, J., . . . Crusto, C. A. (2014). Children exposed to the arrest of
a family member: Associations with mental health. Journal of Child and
Family Studies, 23, 214 –224. http://dx.doi.org/10.1007/s10826-0139717-2
Rodríguez, M. A., Young, M. E., & Wallace, S. P. (2015). Creating
conditions to support healthy people: State policies that affect the health
of undocumented immigrants and their families. Los Angeles, CA:
University of California Global Health Institute.
Rodriguez, N., Steinberg, A., & Pynoos, R. S. (1999). UCLA PTSD Index
for DSM–IV instrument information: Child version, parent version,
adolescent version. Los Angeles, CA: UCLA Trauma Psychiatry.

This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

PTSD AND LATINO CITIZEN CHILDREN
Ross, C. E., & Mirowsky, J. (2009). Neighborhood disorder, subjective
alienation, and distress. Journal of Health and Social Behavior, 50,
49 – 64. http://dx.doi.org/10.1177/002214650905000104
Samuelson, K. W., Krueger, C. E., Burnett, C., & Wilson, C. K. (2010).
Neuropsychological functioning in children with posttraumatic stress
disorder. Child Neuropsychology, 16, 119 –133. http://dx.doi.org/10
.1080/09297040903190782
Steinberg, A. M., Brymer, M. J., Decker, K. B., & Pynoos, R. S. (2004).
The University of California at Los Angeles Post-Traumatic Stress
Disorder Reaction Index. Current Psychiatry Reports, 6, 96 –100. http://
dx.doi.org/10.1007/s11920-004-0048-2
Steinberg, A. M., Brymer, M. J., Kim, S., Briggs, E. C., Ippen, C. G.,
Ostrowski, S. A., . . . Pynoos, R. S. (2013). Psychometric properties of
the UCLA PTSD Reaction Index: Part I. Journal of Traumatic Stress,
26, 1–9. http://dx.doi.org/10.1002/jts.21780
Stoney, S., Batalova, J., & Russell, J. (2013). South American immigrants
in the United States. Migration Information Source. Retrieved from
http://www.migrationpolicy.org/article/south-american-immigrantsunited-states
Suárez-Orozco, C., Bang, H. J., & Kim, H. Y. (2011). I felt like my heart
was staying behind: Psychological implications of immigrant family
separations & reunifications for immigrant youth. Journal of Adolescent
Research, 21, 222–257. http://dx.doi.org/10.1177/0743558410376830
U.S. Department of Health & Human Services. (2003). President’s New
Freedom Commission: Achieving the promise—Transforming mental
health in America. Retrieved from http://govinfo.library.unt.edu/
mentalhealthcommission/reports/FinalReport/downloads/downloads
.html
Weissman, M. M., Orvaschel, H., & Padian, N. (1980). Children’s symptom and social functioning self-report scales: Comparison of mothers’

View publication stats

361

and children’s reports. Journal of Nervous and Mental Disease, 168,
736 –740. http://dx.doi.org/10.1097/00005053-198012000-00005
Wessler, S. F. (2012, December 17). Nearly 250,000 deportations of
parents of U.S. citizens in just over two years. Retrieved from http://
colorlines.com/archives/2012/12/us_deports_more_than_200k_parents
.html
Wherry, J. N., Hunsaker, S. F., Pettigrew, H. V., Conaway, I., TrejosCastillo, E., Caldera, Y., . . . Cordero, L. (2014). Reliability and validity
of the Trauma Symptom Checklist for Young Children—Spanish version. Journal of Child & Adolescent Trauma, 7, 3–12. http://dx.doi.org/
10.1007/s40653-014-0003-2
Yoshikawa, H., Kholoptseva, J., & Suárez-Orozco, C. (2013). The roles of
public policies and community-based organization in the developmental
consequences of parent undocumented status. Society for Research in
Child Development Social Policy Report, 27, 1–23. Retrieved from
http://www.srcd.org/sites/default/files/documents/E-News/spr_273.pdf
Zayas, L. H., Aguilar-Gaxiola, S., Yoon, H., & Rey, G. N. (2015). The
distress of citizen-children with detained and deported parents. Journal
of Child and Family Studies, 24, 3213–3223. http://dx.doi.org/10.1007/
s10826-015-0124-8
Zong, J., & Batalova, J. (2015). Frequently requested statistics on immigrants and immigration in the United States. Retrieved from http://www
.migrationpolicy.org/article/frequently-requested-statisticsimmigrantsand-immigration-united-states#Children%20with%20Immigrant%20
Parents

Received November 6, 2015
Revision received June 28, 2016
Accepted July 1, 2016 Ⅲ

1
2
3
4
5
6
7

Exhibit H

8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

ADDENDUM TO MPAS IN SUPPORT OF AMICUS CURIAE
2:18-cv-00490-JAM-KJN

Author's personal copy

Social Science & Medicine 70 (2010) 7–16

Contents lists available at ScienceDirect

Social Science & Medicine
journal homepage: www.elsevier.com/locate/socscimed

War exposure, daily stressors, and mental health in conflict and post-conflict
settings: Bridging the divide between trauma-focused and
psychosocial frameworks
Kenneth E. Miller a, *, Andrew Rasmussen b
a

Harvard School of Public Health, Research Program on Children and Global Adversity, François-Xavier Bagnoud Center for Health and Human Rights,
651 Huntington Avenue, Boston, MA 02115, USA
New York University School of Medicine, NY, USA

b

a r t i c l e i n f o

a b s t r a c t

Article history:
Available online 23 October 2009

This paper seeks to bridge the divisive split between advocates of trauma-focused and psychosocial
approaches to understanding and addressing mental health needs in conflict and post-conflict settings
by emphasizing the role that daily stressors play in mediating direct war exposure and mental health
outcomes. The authors argue that trauma-focused advocates tend to overemphasize the impact of direct
war exposure on mental health, and fail to consider the contribution of stressful social and material
conditions (daily stressors). Drawing on the findings of recent studies that have examined the relationship of both war exposure and daily stressors to mental health status, a model is proposed in which
daily stressors partially mediate the relationship of war exposure to mental health. Based on that model,
and on the growing body of research that supports it, an integrative, sequenced approach to intervention
is proposed in which daily stressors are first addressed, and specialized interventions are then provided
for individuals whose distress does not abate with the repair of the social ecology.
Ó 2009 Elsevier Ltd. All rights reserved.

Keywords:
War
Daily stressors
Mental health
Trauma
Psychosocial
Interventions

As the papers in this special issue of Social Science and Medicine
make clear, interest in the psychological effects of organized
violence has grown tremendously over the past 25 years. As in any
growing field of inquiry, a number of controversial issues have
emerged in research and practice with war-affected populations.
Particularly salient among these issues is the conflict between
advocates of what we refer to in this paper as trauma-focused versus
psychosocial approaches to understanding and addressing the
mental health needs of communities affected by armed conflict.
Underlying these two approaches are fundamentally different
assumptions regarding the factors that most influence mental
health in conflict and post-conflict settings. For trauma-focused
advocates, the critical factor is direct exposure to the violence and
destruction of wardthe types of potentially traumatic exposure
typically assessed by war-events checklists (e.g., physical assault,
the destruction of one’s home, the disappearance or death of loved
ones in the Harvard Trauma Questionnaire; Mollica et al., 1992). In
contrast, for psychosocial advocates attention is focused primarily

* Corresponding author. Boston University, Boston Centre for Refugee Health and
Human Rights, 770 Albandy St., Dowling 7, Boston, MA, USA. Tel.: þ1 510 326 7111.
E-mail addresses: [e-mail], [e-mail] (K.E. Miller).
0277-9536/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.socscimed.2009.09.029

on the stressful social and material conditions caused or worsened
by armed conflictdconditions such as poverty, malnutrition,
displacement into overcrowded and impoverished refugee camps,
strife and divisions within communities, the destruction of social
networks and the resulting loss of social and material support, and
the ostracism and struggle for survival of groups such as former
child soldiers, widows, sexual assault survivors, orphans, and
people with war-related disabilities (Boothby, Strang, & Wessells,
2006; Miller & Rasco, 2004; Wessells & Monteiro, 2004). Where
trauma-focused advocates primarily see evidence of enduring warrelated trauma requiring specialized clinical treatment (Neuner &
Elbert, 2007; Yule, 2002), psychosocial advocates see distress
rooted largely in the stressful conditions of everyday life in settings
of organized violence. From a psychosocial viewpoint, altering
those stressful conditions is likely to improve people’s mental
health, while also fostering their inherent capacity to recoverd
with adequate social support and the passing of timedfrom the
lingering effects of exposure to war-related violence and loss
(Betancourt & Williams, 2008; Boothby et al., 2006). Conversely,
trauma-focused advocates believe that ameliorating symptoms of
war-related trauma will not only improve mental health, but will
also enable people to cope more effectively with ongoing environmental stressors.

Author's personal copy

8

K.E. Miller, A. Rasmussen / Social Science & Medicine 70 (2010) 7–16

The debate between trauma-focused and psychosocial
approaches has been fueled in part by differences that are not easily
reconciled. Such differences include disagreement over the extent
to which people are vulnerable or resilient in the face of extreme
and persistent stress (Bonanno, 2004; Kostelny & Wessells, 2004;
Neuner & Elbert, 2007); the ethics and efficiency of individualized
clinical interventions in settings where distress is widespread and
mental health resources are scarce (Inter-Agency Standing
Committee, 2007; Miller & Rasco, 2004; Neuner, Karunakara, &
Elbert, 2004); and the appropriateness of applying Western
psychiatric diagnoses such as post-traumatic stress disorder (PTSD)
and trauma-focused clinical treatments such as narrative exposure
therapy (Neuner, Karunakara, et al., 2004; Neuner, Schauer, et al.,
2004) and EMDR (Shapiro, 2001) to war-affected populations that
are overwhelmingly non-Western (Bracken, Giller, & Summerfield,
1995; Kostelny & Wessells, 2004; Miller, Kulkarni, & Kushner, 2006;
Summerfield, 1999; Wessells & Monteiro, 2004).
Beyond such differences, we suggest that the debate between
advocates of trauma-focused and psychosocial approaches has also
been driven by an empirical framework thatduntil recentlydhas
failed to capture the various pathways by which organized violence
impacts mental health. Research on the psychological impact of
armed conflict has traditionally focused rather narrowly on
examining the relationship between direct war exposure and
mental health. Implicit in this focus is a simple direct effects model
to explain psychological distress in settings of organized violence.
In that model, depicted in Fig. 1, there is a straight line with an
arrow leading from war exposure to mental health, reflecting the
direct effect that exposure is believed to have on mental health
status. The model in Fig. 1 does not include any intervening variables (such as daily stressors) that might either partly or fully
explain the impact of war exposure on mental health.
Armed conflict undoubtedly has profound effects on those who
experience it directly. However, organized violence also generates
or exacerbates a host of highly stressful conditions or daily
stressors, such as poverty, social marginalization, isolation, inadequate housing, and changes in family structure and functioning.
Only recently have researchers begun exploring what happens
when daily stressors are added to the model in Fig. 1. Several
studies, in settings as diverse as Afghanistan (Miller et al., 2008;
Panter-Brick, Eggerman, Mojadidi, & McDade, 2008), Chad (Rasmussen et al., in press), Sri Lanka (Fernando, Miller, & Berger, in
press), Lebanon (Farhood et al., 1993), Algeria (de Jong et al., 2004),
and the West Bank (al-Krenawi, Lev-Wiesel, & Sehwail, 2007), have
now examined the role of daily stressors in helping to explain the
high rates of psychological distress so often found among survivors
of armed conflict. Thus far, the data have consistently shown that
daily stressors also have powerful effects on mental health
outcomes. In our review of those findings in this paper, we suggest
that that the overly simplistic conceptual model in Fig. 1 has
unfortunately led trauma-focused advocates to overestimate the
magnitude of the direct effects of direct war exposure in explaining
psychological distress within communities. We further suggest that
this in turn has contributed to an emphasis on trauma-focused
interventions aimed at alleviating war-related PTSD in situations
where greater attention to daily stressors may have yielded greater
benefits.

Exposure to
War-Related
Violence and
Loss

At the same time, the available data also suggest that a narrowly
psychosocial focus is likely to underestimate the adverse impact
that exposure to armed conflict can have on mental health and
psychosocial functioning. Interventions that exclusively target daily
stressors risk overlooking the need for more specialized treatment
for persistently traumatized or depressed individuals (see Hubbard
& Pearson, 2004, for an excellent discussion of this issue). To our
knowledge, there have been no published studies showing that
altering stressful social and material conditions is in and of itself
sufficient to foster the resolution of severe and persistent trauma or
unresolved grief.
Implicit in much of the discourse regarding the psychosocial
framework is a conceptual model that places great emphasis on
precisely those variablesddaily stressorsdthat are missing from
the trauma-focused model. To the extent that the psychological
impact of armed conflict is seen to operate largely or wholly
through the stressful social and material conditions it creates, the
psychosocial conceptual model may be considered fully mediateddthat is, other factors (i.e., daily stressors) largely or fully
account for the impact that armed conflict has on mental health.
This model is depicted in Fig. 2. The dashed arrow between war
exposure and mental health is meant to reflect the fully mediated
effect accorded to direct war exposure in explaining psychological
distress within a psychosocial framework.
In fact, neither the direct effects model that guides traumafocused interventions (Fig. 1), nor the fully mediated model that
informs many psychosocial programs (Fig. 2), is consistent with
what we are learning about the relative contribution of war
exposure and daily stressors to mental health. As we discuss below,
war exposure does exert a significant direct effect on mental health,
beyond the stressful social and material conditions it creates.
However, the addition of daily stressors to the model does two
important things: (1) it significantly increases the overall explanatory power of the model, and (2) it consistently weakensdthough
by no means eliminatesdthe direct relationship between war
exposure and mental health. To reflect these findings, we adopt the
model recently delineated by Fernando et al. (in press) based on
their research with war and disaster-affected youth in eastern Sri
Lanka. This model, depicted in Fig. 3, includes both war exposure
and daily stressors as predictors of mental health status, and
illustrates the role that daily stressors may play in partially mediating the relationship of war exposure to psychological distress. In
the model, armed conflict results in exposure to violence and loss,
which in turn directly affect mental health and psychosocial functioning. Exposure to armed conflict also gives rise to a constellation
of daily stressors, which in turn affect psychological wellbeing
(partial mediation). Importantly, the model also includes daily
stressors unrelated to armed conflict. This reflects a point to which
we return belowdnamely, that not all distress (including psychological trauma) in situations of armed conflict is related to the
violence itself or to the stressful conditions it so often generates.
In the discussion that follows, we draw on an increasingly robust
set of empirical findings that support this more complex model,
and examine the different pathways through which organized
violence appears to exert its effects on psychological wellbeing. We
recognize that the utility of any model that purports to explain
distress among survivors of armed conflict ultimately lies in its

Mental Health

Fig. 1. Direct effects model of the relationship between war exposure and mental health.

Author's personal copy

K.E. Miller, A. Rasmussen / Social Science & Medicine 70 (2010) 7–16

9

Stressful Material
and Social
Conditions
(Daily Stressors)

Exposure to
War-Related
Violence and Loss

Mental Health

Fig. 2. Mediation by daily stressors of the relationship between war exposure and mental health.

capacity to inform the development of interventions and the allocation of scarce resources. It is our hope that the model discussed
here, and the integrative approach to intervention that we propose
based on it, will contribute to that end, while also helping to bridge
the problematic divide between trauma-focused and psychosocial
frameworks.
We begin by examining those factors that have led to a gradual
shift among researchers away from the direct effects model, and
towards a greater consideration of the ways in which war exposure
and daily stressors both contribute to mental health difficulties.
Woven into that discussion are findings from recent studies that
illustrate and support the partial mediation model in Fig. 3. We
then explore why daily stressors are so powerfully linked to mental
health, drawing on findings from research on the adverse impact of
chronic stress, and on theory concerning the relative importance of
proximal versus distal stressors to mental health (Kanner, Coyne,
Schaefer, & Lazarus, 1981; Rowlison & Felner, 1988). We also
consider shortcomings of the term daily stressors; for example, as
we note below, some ‘‘daily stressors’’ do not actually occur on
a daily basis. Moreover, the category lumps together chronic low
level stressors (e.g., overcrowded housing) with events that are
potentially quite traumatic (child sexual abuse). Finally, we suggest
an integrated and sequenced model of intervention that addresses
the effects both of war exposure and of the stressful social and
material conditions to which armed conflict invariably gives rise.
Looking beyond the direct effects model
Recent interest in examining the ways in which war exposure
and daily stressors might both contribute to mental health status
has its roots in three sets of research findings: (1) the consistently
large amount of unexplained variance in mental health outcomes
when war exposure is used as the sole predictor of psychological
distress (i.e., concern over the limited explanatory power of the
direct effects model); (2) research with refugees in developed

nations showing that that post-migration or exile-related stressors
such as social isolation, unemployment and discrimination
consistently predict levels of psychiatric symptomatology as well
as, or better than, pre-migration exposure to organized violence
(Steel, Silove, Bird, McGorry, & Mohan, 1999; for a meta-analysis
examining pre- and post-migration stressors see Porter & Haslam,
2005); and (3) studies of non-war-affected populations in which
so-called ‘‘daily hassles’’ are often more highly associated with
mental health symptom severity than major life events (Kanner
et al., 1981; Rowlison & Felner, 1988). We consider each of these
factors below. First, however, we briefly consider the origin of the
direct effects model and the particular context that gave rise to it.
Origin of the direct effects model
Research on the mental health effects of organized violence on
civilians began in earnest in the 1980s, and followed two rather
distinct tracks. In apartheid era South Africa and in Latin American
countries suffering under prolonged state terror and civil war,
psychologists adopted a broad view of the pathways by which
´
organized violence influenced mental health (Buitrago Cuellar,
´
´
2004; Dawes & Donald, 1994; Gibson, 1989; Martın Baro, 1989;
Melville & Lykes, 1992; Straker, 1988). The effects of direct exposure
to physical violence were seen against the backdrop of the structural violence that formed that stressful context of everyday life
(poverty, discrimination, and marginalization; Farmer, 2004;
Dawes & Donald, 1994). In a similar vein, the impact of violence was
analyzed at all levels of the social ecology, from individual mental
health to the functioning of families and communities (Buitrago
´
´
´
Cuellar, 2004; CODEPU, 1989; Martın Baro, 1989).
In contrast, researchers in North America, Europe, and Australia
tended to view the mental health needs of refugees recently arrived
from Latin America and Southeast Asia through the lens of Western
psychiatry and the recently developed diagnosis of PTSD. Although
it was developed based on research and clinical work with

Daily Stressors
Caused or
Worsened by
Armed Conflict

Exposure to
Armed Conflict

Mental Health

Daily Stressors
Unrelated to
Armed Conflict

Fig. 3. Daily stressors as partially mediating the relationship of armed conflict to mental health and psychosocial status. Adapted from Fernando et al., in press.

Author's personal copy

10

K.E. Miller, A. Rasmussen / Social Science & Medicine 70 (2010) 7–16

American veterans of the Viet Nam War, the diagnostic criteria for
PTSD clearly specified that it was intended as the diagnosis of
choice when post-traumatic symptoms arose following any sort of
traumatic event that entailed at least the perception of lifethreatening danger beyond one’s control. For many refugees, who
had survived terrifying experiences of extreme violence and
showed visible signs of distress, PTSD seemed ideally suited for
classifying their experience of distress. Given the salience of their
war stories, it was understandably assumed that their high levels of
distress were the result of their exposure to the frightening
violence and destruction from which they had escaped (Arroyo &
Eth, 1986; Kinzie, Sack, Angell, Clarke, & Ben, 1989; Kinzie, Sack,
Angell, Manson, & Rath, 1986). Within a short time, research with
refugees had become focused heavily on assessing the ‘‘dose-effect’’
relationship between direct war exposure and PTSD symptom
levels (Mollica et al., 1998; Smith, Perrin, Yule, Hacam, & Stuvland,
2002); that is, the emphasis was on examining the extent to which
degree of war exposure predicted or accounted for severity of PTSD
symptoms or the likelihood of receiving a diagnosis of PTSD. As we
suggested earlier, underlying this emphasis was an assumption that
war exposure represented the critical determinant of distress
among survivors of political violence (and that PTSD represented
the critical mental health impact of war exposure). Although the
strength of the association between exposure and PTSD varied
considerably across studies, consistent evidence of a dose-effect
relationship emerged over time (Fox & Tang, 2000; Jaranson et al.,
2004; Mollica et al., 1999; Tang & Fox, 2001). War exposure was
clearly linked to the development of PTSD symptoms, and greater
exposure was predictive of greater PTSD symptomatology. Based at
least partly on this research, clinical guidelines and recommendations were developed and widely disseminated regarding the
treatment of traumatized refugees (Basoglu, 1998; van der Veer,
1999; Varvin & Hauff, 1998). This combination of PTSD-focused
research and clinical work with refugees was also critical in
launching a worldwide interest in the mental health of waraffected populations, and continues to play a critical role in
procuring critical resources in refugee populations in disaster and
resettlement settings (Breslau, 2004), much as research and treatment with returning Viet Nam veterans spurred a growth in
resources for veterans and active duty soldiers today.
Rapid growth in the field of traumatology has fueled global
interest in the study of PTSD. Researchers trained in Western
psychiatry and clinical psychology have increasingly adopted the
trauma-focused framework developed in the West, shifting the
focus of research in non-Western societies affected by armed
conflict to the study of PTSD (and related psychopathology) and its
relation to war exposure (Fox & Tang, 2000; Lopes Cardozo et al.,
2004; Neuner, Karunakara, et al., 2004; Neuner, Schauer, et al.,
2004; Thabet & Vostanis, 2000). A similar pattern of results has
emerged to that found in earlier studies of refugees resettled in
Western societies: greater direct exposure to war events is associated with higher levels of PTSD symptoms (see Barenbaum, Ruchkin, & Schwab-Stone, 2004; de Jong, 2002, for excellent reviews).
Although much of our focus in this paper is on the mental health
correlates of exposure to the indirect effects of armed conflict (i.e.,
daily stressors), we also recognize the profoundly distressing
nature of direct exposure to armed conflict.
Unexplained variance in the direct effects model and the inclusion of
daily stressors
The recent addition of daily stressors to the direct effects model
reflects a growing concern that variability in the degree of direct
war exposure leaves a substantial proportion of variance in mental
health outcomes, including PTSD, unexplained when war exposure

is used as the sole predictor of distress. Having just established that
exposure does predict PTSD symptom levels, we also note that such
prediction is far from perfect; in fact, war exposure typically
accounts for less than 25% of the variance in PTSD symptoms, and
often much less than that. For example, in their study of mental
health among adults in the Afghan capital of Kabul, Miller et al.
(2008) found a correlation of .39 between level of war exposure (as
measured by total score on the Afghan War Experiences Scale) and
level of PTSD symptomatology (assessed using the Impact of Events
Scale-Revised; Weiss & Marmar, 1997). Squaring that correlation
coefficient, we find that war exposure in the war-torn city of Kabul
accounted for only about 15% of the variance in PTSD symptom
levels.
Findings have been similar in other studies. In their study of
factors influencing the mental health of youth in eastern Sri Lanka
(a region of the country badly affected by civil war and natural
disaster), Fernando et al. (in press) found that war and disaster
exposure accounted for a mere 8% of the variance in PTSD symptom
levels. In a study of Palestinian youth in the West Bank, al-Krenawi
et al. (2007) assessed exposure to political violence as well as
various forms of violence within the family. They found a correlation of .14 between exposure to political violence and scores on the
Brief Symptom Inventory, suggesting that direct exposure
accounted for only about 2% of the variance in distress among the
youth in their sample (as we note below, family violence was
a considerably stronger predictor of distress in their sample). And
in a study of predictors of distress among Darfurian refugees in
refugee camps in Chad, Rasmussen et al. (in press) found that only
about 1% of variance in PTSD symptom levels was attributable to
the violence experienced in Sudan.
In all of these studies, only a small proportion of the variance in
PTSD symptom severity levels was related to the degree of exposure to armed conflict. This same pattern is found extensively in
studies using war exposure as the sole predictor of distress, and it is
even more pronounced when outcomes other than PTSD are used
as dependent variables (e.g., depression, functional impairment).
This robust finding would appear to call into question the widespread assumption that the degree of war exposure is the critical
determinant of mental health severity in conflict and post-conflict
societies. The large amount of unexplained variance has led
researchers to ask what other variables beyond war exposure might
be contributing to levels of distress or psychiatric symptomatology.
The decision to focus on daily stressors was informed at least partly
by recent studies of refugees resettled in developed nations, for
whom post-migration stressors have been found to predict mental
health status at least as strongly as prior history of war exposure.
The salience of post-migration stressors among refugees in
developed nations
As described above, early studies of refugees’ mental health
needs focused on measuring the ‘‘dose–effect’’ relationship
between war exposure and psychopathology, primarily PTSD but
also depression, anxiety, and functional impairment. Research
lagged behind the experience of clinicians and resettlement
workers, however, who noted that the experience of resettlement
confronted refugees with a host of stressful challenges, ranging
from a lack of culturally relevant competencies to inadequate
housing, poverty, social isolation, discrimination, anddfor undocumented individuals, a chronic fear of discovery and deportation
(Birman et al., 2005; Silove, 1999). As researchers began adding
these post-migration or exile-related stressors to their models, they
discovered that post-migration stressors accounted for equal or
greater variance in symptomatology relative to pre-migration war
exposure. Post-migration stressors have been consistently stronger

Author's personal copy

K.E. Miller, A. Rasmussen / Social Science & Medicine 70 (2010) 7–16

predictors than war exposure of depression, while war exposure
has tended to be more strongly related than post-migration
stressors to PTSD (Ellis, MacDonald, Lincoln, & Cabral, 2008; GorstUnsworth & Goldenberg, 1998; Miller et al., 2002; Montgomery,
2008; Steel et al., 1999). Nonetheless, post-migration stressors have
also consistently been related to PTSD symptom levels among
refugees, though the mechanism by which they may affect PTSD
symptomatology remains unclear at present (e.g., do they deplete
coping resources, thereby leaving people more vulnerable to the
impact of prior war exposure? Are some exile-related stressors
themselves traumatogenic?). In any case, the significance of
these findings for practitioners cannot be overstated. Clearly,
a narrow focus on healing war-related trauma among refugees
risks overlooking significant sources of current environmental
stress that might readily be targeted for intervention (Miller, 1999;
Silove, 1999).
There is an interesting similarity between findings regarding the
salience of post-migration stressors among refugees and the clinical significance of post-disaster stressors in settings of natural
calamities. For example, in the wake of the bushfire disaster in
southeastern Australia in 1983, Clayer, Bookless-Pratz, and McFarlane (1985) found that post-disaster stressors such as financial
hardship and difficulties rebuilding were as powerfully related to
survivors’ mental health as their actual exposure to the fire itself. In
a similar vein, McFarlane (1995) notes that that post-disaster
stressors that ensued from the an earthquake that hit the Yunnan
province of China accounted for twice the variance in PTSD
symptoms compared with the actual experiences of injury, loss,
and threat resulting from direct exposure to the earthquake. And in
the previously mentioned study by Fernando et al. (in press), the
impact of the tsunami that hit Sri Lanka in 2004 was at least partly
mediated by the stressful living conditions it createddthat is, daily
stressors resulting from the tsunami (e.g., displacement to refugee
camps, inability to get basic needs met) were at least as powerful as
actual exposure to the tsunami in predicting symptoms of distress.
There is a natural parallel between the concepts of postmigration and post-disaster stressors, on the one hand, and the
idea of daily stressors in conflict and post-conflict settings, on the
other. In each case, the reference is to constellations of stressors
that are generated or exacerbated by highly distressing and
potentially traumatic situations. In the case of resettled refugees,
armed conflict exposes to people to violence and loss, but it also
forces them into exile, where they are confronted with a host of
potentially stressful challenges related to adapting to life as refugees or asylum seekers. In conflict and post-conflict settings, armed
conflict gives rise to (or worsens) the social and material conditions
of everyday life. Survivors of organized violence are thus confronted with a set of enduring and stressful phenomena with which
they must contend while also coping with the impact of direct
exposure to situations of violent conflict.
As researchers began exploring the salience of daily stressors in
conflict and post-conflict settings, two related questions became
central:
1) To what extent do daily stressors help account for high levels of
unexplained variance in levels of distress? That is, to what
extent does the addition of daily stressors strengthen the
explanatory power of the direct effects model?
2) To what extent do daily stressors function to mediate or explain
the relationship between war exposure and distress?
Several studies have examined the relative contribution of war
exposure and daily stressors to levels of distress without specifically looking at whether daily stressors actually mediate the relation of war exposure to mental health. The findings have generally

11

been quite consistent: daily stressors have shown strong and
significantly related main effects on mental health outcomes,
including PTSD (al-Krenawi et al., 2007; Catani, Schauer, & Neuner,
2008; Farhood et al., 1993; Miller et al., 2008). Moreover, in two
recent studies (Fernando et al., in press; Rasmussen et al., in press),
the addition of daily stressors to the model substantially weakened,
though did not eliminate, the relationship of war exposure to
mental health status. Perhaps most importantly, the addition of
daily stressors significantly increased the explanatory power of
models predicting levels of distress, disorder, or functional
impairment.
In the Miller et al. (2008) study of mental health in Afghanistan,
locally salient daily stressors were first identified through interviews with community members and with input of an expert panel
of local Afghans. The Afghan Daily Stressors Scale (ADSS) was
created based on these qualitative data. Sample items on the ADSS
include overcrowded housing, poverty, unemployment, the security situation, violence in the home, poor health, air pollution, and
traffic congestion making public transportation extremely difficult
(in a subsequent study of university students in Kabul by PanterBrick et al. (2008), two subscales were identified in the ADSS,
socioeconomic stressors and family stressors; not all items loaded
on these two scales, however). The ADSS was then used together
with a measure of war exposure to predict levels of PTSD, depression, anxiety, functional impairment, and a locally derived measure
of general distress, the Afghan Symptom Checklist (Miller, Omidian,
et al., 2006). The addition of daily stressors significantly increased
the explanatory power of each model (i.e., the predictive power for
each mental health outcome); it also lowered (but did not eliminate) the predictive power of war exposure on all mental health
outcomes. Moreover, daily stressors moderated the effect of direct
war exposure, so that the effect of war experiences was weaker
among those who experienced more severe daily stressors. Had
daily stressors not been included in the analysis, the predictive
power of war exposure would have been deceptively inflated
because it masked the contribution of daily stressors. With regard
to specific outcomes, daily stressors were better at predicting
depression, functional impairment, and general distress. Among
women, war exposure and daily stressors were comparably strong
predictors of PTSD, while among men, only daily stressors predicted PTSD.
In their study of Palestinian youth in the West Bank, al-Krenawi
et al. (2007) found that family violence, including spousal violence,
parental violence against children and violence between siblings,
better predicted children’s mental health status (B ¼ .38) than their
level of exposure to political violence in the community (B ¼ .08).
Family violence, specifically child physical abuse, was also found to
strongly predict PTSD symptom levels among Tamil children in the
northeast of Sri Lanka (Catani et al., 2008). The likelihood of child
abuse was related to paternal substance abuse and war exposure,
suggesting that the relationship of war-related violence to children’s mental health may have been mediated at least partly through
the impact of violence on fathers’ substance use and increased
likelihood of engaging in violent parenting.
In one of the earliest studies looking at daily stressors and war
exposure as predictors of mental health, Farhood et al. (1993) found
that among Lebanese families, ‘‘daily hassles’’ associated with the
breakdown in community services, economic hardship, and difficulty maintaining contact with family and friends as a result of the
war were all better predictors of distress than the constant threat of
war-related violence.
To our knowledge, only three studies have looked specifically at
the mediating role of daily stressors. Rasmussen et al. (in press)
formally tested the role of refugee camp-related daily hassles in
mediating the relationship of prior war exposure to mental health

Author's personal copy

12

K.E. Miller, A. Rasmussen / Social Science & Medicine 70 (2010) 7–16

among Darfurian refugees in neighboring Chad. Despite the high
level of extreme violence to which the refugees had been exposed,
daily stressors related to a lack of basic needs and a lack of safety in
the camps were better predictors of PTSD than war exposure; in
fact, daily stressors fully mediated the relationship of war exposure
to PTSD. Both war exposure and daily stressors predicted levels of
depression, while current level of perceived safety mediated the
relationship of war exposure to functional impairment. Daily
stressors significantly enhanced the overall explanatory power for
all mental health outcomes, including the indigenous constructs
majnun and hozun.
As noted earlier, Fernando et al. (in press) examined the role of
daily stressors in mediating the relationship of war and disaster
exposure to several mental health outcomes including PTSD,
depression, anxiety, and psychosocial functioning among Sri Lankan youth in the eastern district of Amparada region badly affected
by civil war and a tsunami that killed over 35,000 people on the
island in December, 2004 (SAFMA, 2008). Daily stressors were
identified through focus groups with youth of all three ethnic
groups (Sinhalese, Tamil, and Muslim), and data were used to create
the Children’s Daily Stressor Scale (CDSS). A factor analysis of the
CDSS revealed three subscales or sets of daily stressors: deprivation, child abuse, and inter-parental violence. The addition of these
three subscales to the regression model substantially reduced the
relationship of war and disaster exposure to all mental health
variables; however, the relationship remained significant in all
cases, suggesting that war exposure and daily stressors both were
contributing significantly to levels of distress and functional difficulties. Mediational analysis further revealed that the relationship
of war and disaster exposure to mental health was partially
mediated by deprivation and child abuse, a finding consistent with
the apparent meditational role of paternal child abuse in the study
of Tamil youth cited above (Catani et al., 2008).
Although the focus of this paper is on the impact of armed
conflict and daily stressors on civilians, a recent finding by Betancourt (2008) with former child soldiers in Sierra Leone is germane
to our discussion. Based on two waves of data from 156 youth of
both sexes, the authors found that having wounded or killed others
was significantly related to levels of anxiety and hostility. However,
when current stigma (perceived discrimination) related to their
status as former child soldiers was added to the analysis, the relationship between wounding or killing and mental health status was
no longer significant (i.e., the main effect was no longer significant).
That is, the current experience of feeling stigmatized within their
communities (a form of chronic daily stress) fully mediated the
relationship between the experience of wounding or killing as
a combatant and the subsequent development of psychiatric
symptomatology. Interestingly, stigma did not mediate the relationship between the experience of rape and self-reported levels of
anxiety or hostility. That is, the experience of having been sexually
assaulted exerted a main effect on mental health regardless of the
level of stigma subsequently experienced in the community.1
In all of these studies, the addition of daily stressors significantly
increased the explanatory power of models predicting psychiatric
symptomatology, including symptoms of PTSD in those studies that
assessed it. War exposure generally remained an important
contributing factor, and as the findings of Betancourt (2008)
suggest, it may be that some types of war exposure (e.g., rape) are

1
Betancourt et al. note that their measure of stigma may have failed to
adequately capture the particular experience of stigma/discrimination experienced
specifically by survivors of sexual assault, which may partly explain the absence of
a mediating effect of stigma on the relationship between sexual assault and anxiety
or hostility.

particularly likely to influence mental health directly. Clearly,
however, the data consistently underscore the importance of taking
into account the stressful social and material conditions of
everyday life when seeking to understand and address patterns of
distress in conflict and post-conflict settings.
The importance of daily stressors to mental health status in
non-war-affected populations
The findings we have reviewed above are consistent with
research on major life events and ‘‘daily hassles’’ in non-waraffected populations. In a somewhat counter-intuitive yet highly
robust set of findings, numerous studies have found that the
cumulative effect of daily hasslesdthe lower level stressors of
everyday lifedis more strongly predictive of psychological distress
than exposure to major life eventsdthe sort of acutely stressful
experiences measured by such life events checklists as the widely
used Holmes and Rahe Scale (1967) (Johnson & Sherman, 1997;
Kanner et al., 1981; Rowlison & Felner, 1988; Ruffin, 1993). Although
both popular and professional attention is consistently drawn to
dramatic and potentially traumatic events, whether in peaceful
societies or settings of armed conflict, there is a substantial and
growing body of evidence which suggests that it may be the less
dramatic but more enduring stressful conditions of everyday life
that eventually take the greatest toll on people’s psychological
wellbeing.
Why are daily stressors so stressful?
Having established the important contribution that daily
stressors make to mental health and psychosocial functioning in
conflict and post-conflict settings, it may be fruitful ask why daily
stressors are so impactful. We suggest four reasons. First, daily
stressors represent proximal or immediate stressors, whereas war
exposure is often more of a distal experience, particular in postconflict settings or situations of low intensity warfare where
violence is episodic rather than constant. Poverty, social isolation,
and overcrowded housing confront people on a daily basis; specific
acts of political violence, though highly distressing, may have
occurred a year or more in the past, and thus simply be less
psychologically salient. Research with non-clinical community
samples has shown that survivors of traumatic events are generally
far more resilient than clinical studies and case reports tend to
suggest (Bonanno, 2004), and that with adequate support and the
passing of time, the majority of trauma survivors are likely to regain
their psychological equilibrium (Foa & Rothbaum, 2001). Consequently, distally experienced war exposure may be highly traumatic in the immediate wake of the exposure, but no longer be
experienced as traumatic during assessments conducted after
a significant period of time has passed (Thabet & Vostanis, 2000). In
contrast, daily stressors represent ongoing and often chronic
threats to psychological wellbeing; therefore, their effects are likely
to continue being felt even with the passing of time. Because of
their chronicity, daily stressors may gradually erode people’s
coping resources and tax their mental health. Kubiak (2005) has
suggested that chronic daily stress may gradually diminish people’s
capacity to cope effectively with potentially traumatic life events,
thereby increasing the likelihood of such events causing enduring
symptoms of PTSD. Sapolsky (2004) has documented the numerous
ways in which continuous exposure to stressful circumstancesdincluding lower level, non-traumatic stressorsdgradually
erodes physical and psychological health, and leaves people
increasingly vulnerable to both physical and psychological illness.
More specifically, research on the psychophysiology of stress
suggests that the human stress response is evolutionarily quite well

Author's personal copy

K.E. Miller, A. Rasmussen / Social Science & Medicine 70 (2010) 7–16

adapted to helping us cope effectively with exposure to acute, lifethreatening events, which may help explain why, as noted earlier,
only a minority of people exposed to potentially traumatic experiences actually develop PTSD or other psychiatric disorders
(Bonanno, 2004; Foa & Rothbaum, 2001). In contrast, chronic stress
exposure maintains the stress response system (specifically the
sympathetic nervous system and the hypothalamic-pituitaryadrenal axis) in a state of continuous activation, which in turn has
been linked (via the effects of prolonged exposure to epinephrine,
norepinephrine, and glucocorticoids) to increased risk of both
physical and emotional disorder (Christopher, 2004; Gunnar &
Quevedo, 2007; Sapolsky, 2004).
Second, daily stressors are stressful in part because they are
noxious stimuli that are largely beyond people’s control (just as
direct war experiences are beyond control). Lack of access to clean
water, vulnerability to sexual assault while gathering firewood on
the outskirts of a refugee camp, overcrowded and unsafe housing,
loneliness and a lack of social support because one’s family has
been killed or dispersed due to violencedthese are all stressful
circumstances that may lead people to feel a fundamental lack of
control over the basic resources on which their physical and
psychological wellbeing depend. As Sapolsky (2004) has noted,
such lack of control over unpleasant or aversive events contributes
powerfully to the perception of those events as stressful. This
suggests that psychosocial interventions that foster a greater sense
of control over challenging circumstances may hold considerable
promise as an approach to reducing stress (and thereby improving
mental and physical wellbeing) in conflict and post-conflict
settings.
Third, daily stressors are pervasive within conflict-affected
populations, whereas direct war exposure is highly variable in
many conflict and post-conflict settings (Macksoud & Aber, 1996).
Everyone in a refugee camp has been displaced, and everyone must
contend with the numerous challenges and hardships of camp life.
However, not all camp residents have necessarily been directly
exposed to the violence that caused the displacement. This was the
case in the previously mentioned study of Darfurian refugees in
Chadian refugee camps (Rasmussen et al., in press). Fully 25% of the
refugees in the study reported having had no direct exposure to
organized violence in Sudan; everyone in the study, however, was
exposed to the deprivation and vulnerability of life in the camps.
The situation was similar among Guatemalan refugees in southern
Mexico, many of whom had escaped into Mexico upon hearing of
massacres in neighboring villages in the early 1980s (Manz, 1988).
Finally, the term daily stressors includes a wide range of stressful
phenomena, some of which may be quite traumatic (e.g., child
physical and sexual abuse, intimate partner violence). The inclusion
of potentially traumatic experiences in the same category as lower
intensity chronic stressors such as lack of access to education or
overcrowded housing is a problematic issue to which we return
below; here we note merely that such potentially traumatic events
may account for at least some of the consistently strong relationship that has been found between daily stressors and mental health
status, including PTSD.
Are daily stressors really daily? Unpacking the construct
In seeking to broaden the focus of research beyond the effects of
direct war exposure, we suggested that research on major life
events and daily hassles might offer a useful framework. On closer
inspection, however, there are some reasonable objections that
might be raised to this parallel. First, daily hassles are generally
conceptualized as just that: hassles that occur on a daily basis.
However, some of the phenomena we have considered in the
category of daily stressors do not necessarily occur daily; in fact,

13

they may occur only episodically, yet still have a significantly
adverse impact on mental health. A child may be sexually abused
periodically by a relative or a teacher, a woman may be beaten
recurrently though intermittently by her husband, and poisonous
snakes may enter homes in refugee camps only occasionally.
Though not a daily occurrence, such events are clearly likely to
represent significant sources of stress (and distress). What is likely
to be daily in these examples is the realistic fear of recurrence and
the experience of vulnerability that such intermittent events may
elicit.
A second objection is the inclusion under the label daily stressors
of such a broad range of stressors of highly varied intensity. As
noted in the previous section, measures of daily stressors have
sometimes included items that may be quite traumatic in their
intensity; this does not seem consistent with the lower level types
of chronic stress that were intended by the concept of daily hassles
(Kanner et al., 1981; Rowlison & Felner, 1988). We find merit in this
concern, and propose a distinction between lower intensity and
potentially traumatic daily stressors. Lower intensity stressors
include such experiences as overcrowded housing, lack of access to
education and employment, and social isolation resulting from the
loss of social networks. Potentially traumatic daily stressors, in
contrast, would include experiences such as physical and sexual
abuse of children, spousal abuse, and criminal acts not directly
related to armed conflict (sexual assault in or around refugee camps
by other camp residents or local officials). The concept of potentially traumatic daily stressors is important because it underscores
the reality that even in settings of armed conflict, there are sources
of psychological trauma other than exposure to the conflict itself.
This point was underscored in a recent study of children’s mental
health in Afghanistan (Panter-Brick, Eggerman, Gonzalez, & Safdar,
2009). Not surprisingly, exposure to violence was strongly predictive of both PTSD and depression symptom levels; however, much
of the violence that children reported was not directly related to
war exposure, including domestic and community violence, accidents, and medical treatment. In fact, the authors note that ‘‘Some
children identified severe beatings, a severe accident, or a frightening medical treatment as more traumatic than having witnessed
parents or grandparents being killed in rocket attacks’’ (p. 8). This is
a critical point when planning interventions. A focus on healing the
effects of previously experienced war trauma may seem profoundly
out of sync to a child who is currently being beaten or sexually
abused at home or in the community.
Implications for intervention: a sequenced, integrated model
The findings from the literature we have reviewed in this paper
suggest the potential utility of an integrated approach to intervention that addresses, in a sequential manner, both daily stressors
(low intensity and potentially traumatic) and war exposure. They
also suggest the utility of an empirically-informed set of guidelines
for the allocation of mental health resources and the development
of interventions aimed at improving mental health and psychosocial wellbeing in conflict and post-conflict settings. We note that
the guidelines we propose are consistent with those suggested by
Barenbaum et al. (2004), Betancourt and Williams (2008), and
Bolton and Betancourt (2004).
Guideline 1: It is important to undertake a rapid and contextually
grounded assessment of locally salient daily stressors before developing mental health and psychosocial interventions
There are both similarities and differences in the types of daily
stressors that are salient in different geographic, economic,
cultural, and sociopolitical contexts. Moreover, the salience of

Author's personal copy

14

K.E. Miller, A. Rasmussen / Social Science & Medicine 70 (2010) 7–16

particular stressors is likely to vary by age and gender; for example,
children may be particularly vulnerable to school-related
problems, parental neglect or abandonment, and physical or sexual
abuse, while women may struggle with domestic violence or high
rates of reproductive health-related problems. Numerous reports
describe quick and efficient methods (focus groups, free-listing, key
informant interviews) for identifying locally salient daily stressors,
as well as resources available to help people cope with or modify
those stressors (Bolton & Tang, 2002; de Jong & van Ommeren,
2002; Miller, Fernando, & Berger, in press).
Guideline 2: Before providing specialized clinical services that target
psychological trauma, first address those daily stressors that are
particularly salient and can be affected through targeted
interventions.
Advocates of psychosocial approaches have long maintained
that reducing ongoing sources of stress that tax coping resources,
and reestablishing social ties that foster emotional and material
support, are likely to go a long way towards improving mental
health in war-affected communities. The data we have reviewed are
consistent with this position, for several reasons.
First, daily stressors clearly exert a direct effect on mental
health. By targeting particularly impactful stressors for change, we
can expect to see a direct benefit in terms of reduced distress and
improved psychosocial functioning. Second, daily stressors
contribute to continuously high levels of stress, and it seems
reasonable to infer that coping with continuous stressorsdpoverty,
family violence, unsafe housing, social isolationdis likely to place
considerable demands on people’s coping resources. To the extent
that interventions are able to reduce the occurrence and/or intensity of such stressors, coping resources will be less taxed and thus
be more available for healing from any persistent effects of warrelated violence and loss. In short, by altering the social and
material environments in ways that improve mental health, the
need for formal and resource-intensive mental health services may
be reduced (Bolton & Betancourt, 2004).
Third, findings from research on stress and social support
suggest that strengthening social support networks is likely to exert
significant beneficial effects on mental health, and may in fact
buffer against the development of PTSD in the wake of exposure to
potentially traumatic stress (Norris et al., 2002). This may in turn
have the beneficial effect of reducing the need for specialized
mental health care.
Finally, as Bolton and Betancourt (2004) have noted, by
improving mental health to the extent possible through psychosocial interventions aimed at reducing daily stressors (or, phrased
more positively, by improving the quality of the social and material
ecology), it will be easier to identify those individuals whose
persistent distress does not abate with the reduction of daily stress
and who may in fact require specialized assistance.
Guideline 3: When specialized mental health interventions are indicated, interventions should go beyond PTSD to address the diverse
forms of distress that may result from exposure to war-related violence
and loss.
Although researchers and clinicians have a shown a strong
interest in PTSD, the studies we have reviewed show that war
exposure is also related to a variety of other forms of distress,
including depression, anxiety, impaired social functioning, and
various local idioms of distress. While recognizing that many
symptoms of PTSD are found transculturally and may benefit from
clinical intervention, we share Breslau’s (2004) concern (see also
Barenbaum et al., 2004; de Jong, 2002; Miller, Kulkarni, et al., 2006;

Miller, Omidian, et al., 2006; Summerfield, 1999) that a narrow
focus on treating PTSD may reflect the interest of mental health
professionals more than it does the actual priorities of community
members regarding their own mental health.
Guideline 4: It is essential to take into account that not all symptoms of
trauma are necessarily related to conflict exposure. Even in situations
of armed conflict, there are other sources of psychological trauma.
Although this may seem intuitive or self-evident, we have been
struck by how few studies of mental health in war-affected
communities have assessed exposure to forms of traumatic stress
other than direct war exposure. Given what is known about the
increased risk for PTSD and more complex forms of trauma caused
by experiences such as child abuse (Garbarino &Ganzel, 2000; Terr,
1990) and intimate partner violence (Stein & Kennedy, 2001), it
seems imperative to us to consider the inclusion of potentially
traumatic daily stressors such these in any assessment of factors
contributing to psychological distress in conflict and post-conflict
settings. As we suggested earlier, the relevance and impact of
mental health or psychosocial interventions are likely to be
considerably enhanced when they are seen as targeting those
sources of stress that are most immediately and severely affecting
people. A programmatic focus on healing the effects of previously
experienced war exposure is likely to have limited impact on
individuals who are facing ongoing exposure to traumatic stress in
their homes or communities.

Conclusion
In this paper, we have sought to bridge the longstanding and
unhelpful division between advocates of trauma-focused and
psychosocial approaches to understanding and addressing mental
health needs in conflict and post-conflict settings. We have suggested that among the various factors underlying this split in the
field is a fundamental difference in perception regarding what
factors most critically affect mental health in the wake of organized
violence. Until recently, there was a paucity of data to inform this
largely opinion-based and experience-driven difference in
perception. Presently, however, there are sufficient data to permit
an empirically-informed discussion. We believe that the findings
we have presented make a compelling case for the inclusion of daily
stressors in any model purporting to explain patterns of distress in
war-affected populations. In fact, the available data suggest that
addressing daily stressors should be a priority in the development
of mental health policy, the allocation of scarce resources, and the
design of interventions to assist war-affected communities. Daily
stressors are strongly related to the severity of psychological
distress and psychiatric symptomatology; and, because they are
ongoing, may be targeted for change through well-designed
intervention programs.
The inclusion and prioritization of daily stressors by no means
negates the value of more specialized clinical interventions for
highly distressed individuals whose symptoms do not abate with
the normalization of their environment through the reduction of
daily stressors. War exposure does exert a direct and adverse effect
on mental health, though the data are not consistent with the
current trend towards conceptualizing that effect solely in terms of
PTSD. We suggest that a broad range of specialized interventions
should be brought to bear, from culturally informed adaptations of
Western treatment strategies (Hubbard & Pearson, 2004) to the use
of traditional healers whose explanatory models and methods of
treatment are more likely to be familiar to community members
(de Jong, 2004).

Author's personal copy

K.E. Miller, A. Rasmussen / Social Science & Medicine 70 (2010) 7–16

As we suggested in the introduction to this paper, the difference
in underlying conceptual models is by no means the only point of
disagreement between trauma-focused and psychosocial advocates. Other issues must also be addressed, such as the appropriateness of applying Western diagnoses, the efficacy of
professionally staffed clinical treatments in non-Western cultural
contexts, and even the appropriateness of mental health outcomes
altogether. We recognize, for example, that many advocates of
psychosocial interventions eschew a narrow focus on reducing
psychopathology as the desired outcome of their interventions,
opting instead to focus on strengthening families and communities
and promoting positive outcomes in children (Boothby et al., 2006;
Kostelny & Wessells, 2004). It is our hope that in seeking to build
a bridge between more clinically focused and psychosocially
oriented approaches, we have at least helped initiate a discussion
that may lead to further exploration of common ground and
collaboration between the advocates of these two influential
frameworks.
References
al-Krenawi, A., Lev-Wiesel, R., & Sehwail, M. (2007). Psychological symptomatology
among Palestinian children living with political violence. Child and Adolescent
Mental Health, 12, 27–31.
Arroyo, W., & Eth, S. (1986). Children traumatized by Central American Warfare. In
R. Pynoos, & S. Eth (Eds.), Post-traumatic stress disorder in children (pp. 101–120).
Washington, DC: American Psychiatric Press.
Barenbaum, J., Ruchkin, V., & Schwab-Stone, M. (2004). The psychosocial aspects of
children exposed to war: practice and policy initiatives. Journal of Child
Psychology and Psychiatry, 45, 41–62.
Basoglu, M. (1998). Behavioral and cognitive treatments of survivors of torture. In
J. Jaronson, & M. Popkin (Eds.), Caring for victims of torture (pp. 131–148).
Washington, DC: American Psychiatric Publishing.
Betancourt, T. S. Stigma and the psychosocial adjustment and social reintegration of
former child soldiers in Sierra Leone. Presentation at the International Society
for Traumatic Stress Studies (ISTSS) 2008 Annual Meeting, Chicago, IL,
November, 2008.
Betancourt, T. S., & Williams, T. (2008). Building an evidence base on mental health
interventions for children affected by armed conflict. Intervention, 6, 39–56.
Birman, D., Ho, J., Pulley, E., Batia, K., Everson, M. L., Ellis, H., et al. (2005).
Mental health interventions for refugee children in resettlement: white
paper II. National Child Traumatic Stress Network. Retrieved on March 1,
2009, from. http://www.nctsnet.org/nctsn_assets/pdfs/promising_practices/
MH_Interventions_for_Refugee_Children.pdf.
Bolton, P., & Betancourt, T. (2004). Mental health in postwar Afghanistan. Journal of
the American Medical Association, 292, 626–628.
Bolton, P., & Tang, A. (2002). An alternative approach to cross-cultural function
assessment. Social Psychiatry and Psychiatric Epidemiology, 37, 537–543.
Bonanno, G. (2004). Loss, trauma, and human resilience: have we underestimated
the human capacity to thrive after extremely aversive events? American
Psychologist, 59, 202–208.
Boothby, N., Strang, A., & Wessells, M. (2006). A world turned upside down.
Bloomfield, CT: Kumarian Press.
Bracken, P. J., Giller, J. E., & Summerfield, D. (1995). Psychological responses to war
and atrocity: the limitations of current concepts. Social Science & Medicine, 40,
1073–1082.
Breslau, J. (2004). Cultures of trauma: anthropological views of posttraumatic stress
disorder in international health. Culture, Medicine, and Psychiatry, 28, 113–126.
´
Buitrago Cuellar, J. (2004). Internally displaced Colombians: the recovery of victims
of violence within a psychosocial framework. In K. E. Miller, & L. M. Rasco (Eds.),
The mental health of refugees: Ecological approaches to healing and adaptation
(pp. 229–262). Mahwah, NJ: Lawrence Erlbaum Associates, Inc.
Catani, C., Schauer, E., & Neuner, F. (2008). Beyond individual war trauma: domestic
violence against children in Afghanistan and Sri Lanka. Journal of Marital and
Family Therapy, 34, 165–176.
Christopher, M. (2004). A broader view of trauma: a biopsychosocial-evolutionary
view of the role of the traumatic stress response in the emergence of pathology
and growth. Clinical Psychology Review, 24, 75–98.
Clayer, J. R., Bookless-Pratz, C., & McFarlane, A. (1985). The health and social impact of
ash wednesday bushfires: Survey of the 12 months following the bushfires of
February 1983. South Australia: Mental Health and Evaluation Centre, South
Australian Health Commission.
CODEPU. (1989). The effects of torture and political repression in a sample of
Chilean families. Social Science & Medicine Special Issue: Political violence and
health in the Third World, 28(7), 735–740.
Dawes, A., & Donald, D. (1994). Understanding the psychological consequences of
adversity. In Dawes., & D. Donald (Eds.), Psychological perspectives from South
African research (pp. 1–27). Claremont, South Africa: David Philip Publishers, Lts.

15

Ellis, H., MacDonald, H., Lincoln, A., & Cabral, H. (2008). Mental health of Somali
adolescent refugees: the role of trauma, stress, and perceived discrimination.
Journal of Consulting and Clinical Psychology, 76, 184–193.
Farhood, L., Zurayk, H., Chaya, M., Saadeh, F., Meshefedjian, G., & Sedani, T. (1993).
The impact of war on the physical and mental health of the family: the Lebanese experience. Social Science & Medicine, 36, 1555–1567.
Farmer, P. (2004). An anthropology of structural violence. Current Anthropology,
45(3), 305–325.
Fernando, G., Miller, K., & Berger, D. Growing Pains: the impact of disaster-related
and daily stressors on the psychological and psychosocial functioning of youth
in Sri Lanka. Child Development, in press.
Foa, E., & Rothbaum, B. (2001). Treating the trauma of rape. New York: The Guildford
Press.
Fox, S., & Tang, S. (2000). The Sierra Leonean refugee experience: traumatic events
and psychiatric sequelae. Journal of Nervous and Mental Disease, 188, 490–495.
Garbarino, J., & Ganzel, B. (2000). The human ecology of early risk. In J. Shankoff, &
S. Meisels (Eds.), Handbook of early childhood intervention (2nd ed.). (pp. 76–93)
New York: Cambridge University Press.
Gibson, K. (1989). Children in political violence. Social Science & Medicine, 28,
659–667.
Gorst-Unsworth, C., & Goldenberg, E. (1998). Psychological sequelae of torture and
organized violence suffered by refugees from Iraq: trauma-related factors
compared with social factors in exile. British Journal of Psychiatry, 172, 90–94.
Gunnar, M., & Quevedo, K. (2007). The neurobiology of stress and development.
Annual Review of Psychology, 58, 145–173.
Holmes, & Rahe. (1967). Holmes-Rahe life changes scale. Journal of Psychosomatic
Research, 11, 213–218.
Hubbard, J., & Pearson, N. (2004). Sierra Leonean refugees in Guinea: addressing the
mental health effects of massive community violence. In K. E. Miller, &
L. M. Rasco (Eds.), The mental health of refugees: Ecological approaches to healing
and adaptation (pp. 95–132). Mahwah, NJ: Lawrence Erlbaum Associates, Inc.
Inter-Agency Standing Committee. (2007). IASC guidelines on mental health and
psychosocial support in emergency settings. Available from http://www.
humanitarianinfo.org/iasc/content/products/.
Jaranson, J. M., Butcher, J., Halcon, L., Johnson, D. R., Robertson, C., Savik, K., et al.
(2004). Somali and Oromos refugees: correlates of torture and trauma history.
American Journal of Public Health, 94, 591–598.
Johnson, J., & Sherman, M. (1997). Daily hassles mediate the relationship between
major life events and psychiatric symptomatology: longitudinal findings from
an adolescent sample. Journal of Clinical and Consulting Psychology, 16, 389–
404.
de Jong, J. (2002). Public mental health, traumatic stress and human rights violations in low-income countries. In J. de Jong (Ed.), Trauma, war, and violence:
Public mental health in socio-cultural context (pp. 1–91). New York: Kluwer.
de Jong, J. (2004). Public mental health and culture: disasters as a challenge to
western mental health care models, the self, and PTSD. In J. Wilson, &
B. Drozdek (Eds.), Broken spirits: The treatment of traumatized asylum seekers,
refugees, war and torture victims (pp. 158–178). New York: Brunner-Routledge.
de Jong, J., Komproe, I., van Ommeren, M., el Masri, M., Khaled, N., van de Put, W., et
al. (2004). Lifetime events and posttraumatic stress disorder in 4 postconflict
settings. Journal of the American Medical Association, 286, 555–562.
de Jong, J., & van Ommeren, M. (2002). Toward a culture-informed epidemiology:
combining qualitative and quantitative research in transcultural contexts.
Transcultural Psychiatry, 39(4), 422–433.
Kanner, A. D., Coyne, J. C., Schaefer, C., & Lazarus, R. S. (1981). Comparisons of two
models of stress measurement: daily hassles an uplifts versus major life events.
Journal of Behavioral Medicine, 4, 1–39.
Kinzie, J. D., Sack, W., Angell, R., Clarke, G., & Ben, R. (1989). A three-year follow-up
of Cambodian young people traumatized as children. The American Academy of
Child and Adolescent Psychiatry, 28, 501–504.
Kinzie, J. D., Sack, W., Angell, R., Manson, S., & Rath, B. (1986). The psychiatric effects
of massive trauma on Cambodian children: I. The children. Journal of the
American Academy of Child and Adolescent Psychiatry, 25, 370–376.
Kostelny, K., & Wessells, M. (2004). Internally displaced East Timorese: challenges
and lessons of large-scale emergency assistance. In K. E. Miller, & L. M. Rasco
(Eds.), The mental health of refugees: Ecological approaches to healing and
adaptation (pp. 187–225). Mahwah, NJ: Lawrence Erlbaum Associates, Inc.
Kubiak, S. (2005). Trauma and cumulative adversity in women of a disadvantaged
social location. American Journal of Orthopsychiatry, 75, 451–465.
Lopes Cardozo, B., Bilukha, O., Gotway Crawford, C., Shaikh, I., Wolfe, M., Gerber, M.,
et al. (2004). Mental health, social functioning, and disability in postwar
Afghanistan. Journal of the American Medical Association, 292, 575–584.
Macksoud, M. S., & Aber, J. L. (1996). The war experience and psychosocial development of children in Lebanon. Child Development, 67, 70–88.
Manz, B. (1988). Refugees of a hidden war. New York: SUNY Press.
´
´
Martın Baro, I. (1989). Political violence and war as causes of psychosocial trauma in
El Salvador. International Journal of Mental Health, 18, 3–20.
McFarlane, A. C. (1995). Helping the victims of disasters. In J. R. Freedy, &
S. E. Hobfall (Eds.), Traumatic stress: From theory to practice (pp. 287–314). New
York: Plenum Press.
Melville, M., & Lykes, B. (1992). Guatemalan Indian children and the sociocultural
effects of government-sponsored terrorism. Social Science & Medicine, 34,
33–48.
Miller, K. E. (1999). Rethinking a familiar model: psychotherapy and the mental
health of refugees. Journal of Contemporary Psychotherapy, 29(4), 283–306.

Author's personal copy

16

K.E. Miller, A. Rasmussen / Social Science & Medicine 70 (2010) 7–16

Miller, K.E., Fernando, G., & Berger, D. Daily stressors in the lives of Sri Lankan youth:
A mixed-methods approach to assessment in a context of war and natural
disaster. Intervention. (In press).
Miller, K. E., Kulkarni, M., & Kushner, H. (2006). Beyond trauma-focused psychiatric
epidemiology: bridging research and practice with war-affected populations.
American Journal of Orthopsychiatry, 76, 409–422.
Miller, K. E., Omidian, P., Rasmussen, A., Yaqubi, A., Daudzai, H., Nasiri, M., et al.
(2008). Daily stressors, war experiences, and mental health in Afghanistan.
Transcultural Psychiatry, 45, 611–639.
Miller, K. E., Omidian, P., Qurashy, A. S., Nasiry, M. N., Quarshy, N., Nasiry, S., et al.
(2006). The Afghan symptom checklist: a culturally grounded approach to
mental health assessment in a conflict zone. American Journal of Orthopsychiatry, 76, 423–433.
Miller, K. E., & Rasco, L. M. (2004). An ecological framework for addressing the
mental health needs of refugee communities. In K. E. Miller, & L. M. Rasco (Eds.),
The mental health of refugees: Ecological approaches to healing and adaptation
(pp. 1–64). Mahwah, NJ: Lawrence Erlbaum Associates, Inc.
Miller, K. E., Weine, S., Ramic, A., Brkic, N., Djuric Bjedic, Z., Smajkic, A., et al. (2002).
The relative contribution of war experiences and exile-related stressors to
levels of psychological distress among Bosnian refugees. Journal of Traumatic
Stress, 15, 377–387.
Mollica, R., McInnes, K., Pham, T., Fawzi, M., Smith, C., Murphy, E., et al. (1998). The
dose–effect relationships between torture and psychiatric symptoms in Vietnamese ex-political detainees and a comparison group. Journal of Nervous and
Mental Disease, 186, 543–553.
Mollica, R. F., Caspi-Yavin, Y., Bollini, P., Truong, T., Tor, S., & Lavelle, J. (1992). The
Harvard trauma questionnaire: validating a cross-cultural instrument for
measuring torture, trauma, and post-traumatic stress disorder in Indochinese
refugees. Journal of Nervous and Mental Disease, 180, 111–116.
Mollica, R. F., McInnes, K., Sarajlic, N., Lavelle, J., Sarajlic, I., & Massagli, M. P. (1999).
Disability associated with health status in Bosnian refugees living in Croatia.
Journal of the American Medical Association, 281(5), 433–439.
Montgomery, E. (2008). Long-term effects of organized violence on young Middle
Eastern refugees’ mental health. Social Science & Medicine, 67, 1596–1603.
Neuner, F., & Elbert, T. (2007). The mental health disaster in conflict settings: can
scientific research help? BMC Public Health, 7, 275.
Neuner, F., Karunakara, U., & Elbert, T. (2004). A comparison of narrative exposure
therapy, supportive counseling, and psychoeducation for treating posttraumatic
stress disorder in an African refugee settlement. Journal of Consulting and
Clinical Psychology, 72, 579–587.
Neuner, F., Schauer, M., Karunakara, U., Laschik, C., Robert, C., & Elbert, T. (2004).
Psychological trauma and evidenced for enhanced vulnerability for posttraumatic stress disorder through previous trauma among West Nile refugees.
BMC Psychiatry, 4, 1–7.
Norris, F. H., Friedman, M., Watson, P., Byrne, C., Diaz, E., & Kaniasty, K. (2002).
60,000 Disaster victims speak: an empirical review of the empirical literature,
1981–2001. Psychiatry, 65, 207–239.
Panter-Brick, C., Eggerman, M., Gonzalez, V., & Safdar, S. (2009). Violence, suffering,
and mental health in Afghanistan: a school-based survey. The Lancet.
doi:10.1016/S0140-6736(09)61080-1. Published online, August 21, 2009.
Panter-Brick, C., Eggerman, M., Mojadidi, A., & McDade, T. (2008). Social stressors,
mental health, and physiological stress in an urban elite of young Afghans in
Kabul. American Journal of Human Biology, 20, 627–641.

Porter, M., & Haslam, N. (2005). Predisplacement and postdisplacement factors
associated with mental health of refugees and internally displaced persons.
Journal of the American Medical Association, 294(5), 602–612.
Rasmussen, A., Nguyen, L., Wilkinson, J., Vundla, S., Raghavan, S., Miller, K. E., et al.
Rates and impact of trauma and current stressors among Darfuri refugees in
Eastern Chad. American Journal of Orthopsychiatry, in press.
Rowlison, R., & Felner, R. (1988). Major life events, hassles, and adaptation in
adolescence: confounding in the conceptualization and measurement of life
stress. Journal of Personality and Social Psychology, 55, 432–444.
Ruffin, C. (1993). Stress and health: little hassles vs. major life events. Australian
Psychologist, 28, 201–208.
Sapolsky, R. (2004). Why zebras don’t get ulcers. New York: Owl Books.
Shapiro, F. (2001). Eye movement desensitization and reprocessing: Basic principles,
protocols and procedures (2nd ed.). New York: Guilford Press.
Silove, D. (1999). The psychosocial effects of torture, mass human rights violations,
and refugee trauma. Journal of Nervous and Mental Disease, 187, 200–207.
Smith, P., Perrin, S., Yule, W., Hacam, B., & Stuvland, R. (2002). War exposure among
children from Bosnia-Hercegovina: psychological adjustment in a community
sample. Journal of Traumatic Stress, 15, 147–156.
South Asia Free Media Association. (2008). Sri Lanka: ethnic conflict, civil war, and
politics. http://www.lankalibrary.com/pol/militancy_history.htm Retrieved
June 3, 2008.
Steel, Z., Silove, D., Bird, K., McGorry, P., & Mohan, P. (1999). Pathways from war
trauma to posttraumatic stress symptoms among Tamil asylum seekers, refugees, and immigrants. Journal of Traumatic Stress, 12, 421–435.
Stein, M., & Kennedy, C. (2001). Major depressive and post-traumatic stress disorder
comorbidity in female victims of intimate partner violence. Journal of Affective
Disorders, 66, 133–138.
Straker, G. (1988). Post-traumatic stress disorder: a reaction to state-supported
child abuse and neglect. Child Abuse and Neglect, 12, 383–395.
Summerfield, D. (1999). A critique of seven assumptions behind psychological trauma
programmes in war-affected countries. Social Science & Medicine, 48, 1449–1462.
Tang, S., & Fox, S. (2001). Traumatic experiences and the mental health of Senegalese refugees. Journal of Nervous and Mental Disease, 189, 507–512.
Terr, L. (1990). Too scared to cry: Psychic trauma in childhood. New York: Harper &
Row.
Thabet, A., & Vostanis, P. (2000). Post-traumatic stress disorder reactions in children
of war: a longitudinal study. Child Abuse and Neglect, 24, 291–298.
van der Veer, G. (1999). Counseling and therapy with refugees and victims of trauma:
Psychological problems of victims of war, torture, and repression. New York:
Wiley.
Varvin, S., & Hauff, E. (1998). Psychoanalytically oriented psychotherapy with
torture survivors. In J. Jaronson, & M. Popkin (Eds.), Caring for victims of torture
(pp. 107–116). Washington, DC: American Psychiatric Publishing.
Weiss, D., & Marmar, C. (1997). The impact of events scale-revised. In J. Wilson, &
T. Keane (Eds.), Assessing psychological trauma and PTSD: A handbook for practitioners (pp. 399–411). New York: Guilford Press.
Wessells, M., & Monteiro, C. (2004). Internally displaced Angolans: a child-focused,
community-based intervention. In K. E. Miller, & L. M. Rasco (Eds.), The mental
health of refugees: Ecological approaches to healing and adaptation (pp. 67–94).
Mahwah, NJ: Lawrence Erlbaum Associates, Inc.
Yule, W. (2002). Alleviating the effects of war and displacement on children.
Traumatology, 8, 160–180.

1
2
3
4
5
6
7

Exhibit I

8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28

ADDENDUM TO MPAS IN SUPPORT OF AMICUS CURIAE
2:18-cv-00490-JAM-KJN

RESEARCH LETTERS

Research letters

Mental health of detained asylum seekers
Allen S Keller, Barry Rosenfeld, Chau Trinh-Shevrin, Chris Meserve, Emily Sachs, Jonathan A Leviss, Elizabeth Singer, Hawthorne Smith,
John Wilkinson, Glen Kim, Kathleen Allden, Douglas Ford
Asylum seekers arriving in the USA are likely to be held in
detention for months or years pending adjudication of their
asylum claims. We interviewed 70 asylum seekers detained in
New York, New Jersey, and Pennsylvania. We used self-report
questionnaires to assess symptoms of anxiety, depression,
and post-traumatic stress disorder. At baseline, 54 (77%)
participants had clinically significant symptoms of anxiety,
60 (86%) of depression, and 35 (50%) of post-traumatic
stress disorder; all symptoms were significantly correlated
with length of detention (p=0·004, 0·017, and 0·019,
respectively). At follow-up, participants who had been
released had marked reductions in all psychological
symptoms, but those still detained were more distressed than
at baseline. Our findings suggest detention of asylum seekers
exacerbates psychological symptoms.

Lancet 2003; 362: 1721–23

Worldwide, there is a growing trend toward detention of
asylum seekers arriving in industrialised countries for
months or even years pending adjudication of their asylum
claims.1 In the USA, 5000 asylum seekers are estimated to
be held in detention,1–3 although reliable statistics on the
number of detained asylum seekers are unavailable.
Detention of asylum seekers has concerned health
professionals and human rights advocates, in part because
of the potential detrimental effects on the mental health of
detainees.1,3 Research on this subject, however, has been
limited by difficulties in gaining access to detention
centres. The Bellevue New York University (NYU)
Program for Survivors of Torture and Physicians for
Human Rights have done a systematic and longitudinal
study of the effects of postmigration detention on the
mental health of asylum seekers.
Detained (n=35)

The US Immigration and Naturalization Services (INS)
permitted access to detention facilities in New York, New
Jersey, and Pennsylvania. These facilities included two INS
detention centres run by private contractors. These jails are
virtually windowless converted warehouses where only
non-criminal INS detainees are incarcerated. Additionally,
access was permitted to three local government-run jails in
which criminals are also held. In all of these facilities,
asylum seekers are heavily guarded, required to wear jail
uniforms, and are shackled whenever they are transported
outside of the detention facilities. The INS did not allow us
access to a random sample of detained asylum seekers at
these facilities. Therefore, we asked six local organisations
providing pro-bono legal representation to detained asylum
seekers to contact clients and ask about their willingness to
be interviewed. Detainees were informed by researchers of
the voluntary nature of the study and that participation
would not affect their asylum applications. We obtained
written informed consent from all participants. The study
was approved by the Institutional Review Board of New
York University School of Medicine and a review
committee for Physicians for Human Rights.
Detainees were interviewed by physicians experienced in
caring for refugees; translators assisted if necessary.
Standardised psychological symptom measures were used:
the Hopkins symptom checklist-25 (HSCL-25)4 and the
post-traumatic stress disorder subscale of the Harvard
trauma questionnaire (HTQ).5 Both measures have been
used in studies of refugee populations and previously
translated and back-translated in several languages,
including French, Spanish, and Arabic. For participants
who spoke other languages, scales were translated by the
interpreter. Demographic information and history of
Released (n=26)

Number (%) above
recommended cut-off*

Symptom scores,
mean (SD)

Number (%) above
recommended cut-off*

Baseline
Anxiety
Depression
PTSD

28 (80%)
30 (86%)
19 (54%)

2·40 (0·71)
2·52 (0·69)
2·52 (0·62)

19 (73%)
22 (85%)
10 (39%)

2·33 (0·72)
2·45 (0·65)
2·45 (0·62)

Follow-up
Anxiety
Depression
PTSD

30 (86%)
31 (89%)
21 (60%)

2·58 (0·80)
2·73 (0·70)
2·63 (0·71)

9 (35%)
10 (39%)
3 (12%)

1·59 (0·56)
1·65 (0·59)
1·80 (0·56)

Change in symptom scores†
Anxiety
Depression
PTSD

··
··
··

0·17 (0·61)
0·21 (0·42)
0·12 (0·51)

··
··
··

Symptom scores,
mean (SD)

–0·75 (0·84)
–0·80 (0·81)
–0·64 (0·64)

PTSD=post-traumatic stress disorder. *Cut-offs: 1·75 for HSCL-25 depression and anxiety subscales, 2·5 for HTQ. †p not significant for changes in any symptom

score in detained group. p=0·0001 for all three symptom score changes in released group. Mean (SD) values are group mean at each assessment.
Change=change in mean score from baseline assessment.

Psychological symptoms of asylum seekers still detained and those released at follow-up

THE LANCET • Vol 362 • November 22, 2003 • www.thelancet.com

For personal use. Only reproduce with permission from The Lancet.

1721

RESEARCH LETTERS

1·00

Change in symptom distress scores

0·80

Released (n=26)
Detained (n=35)

0·60
0·40
0·20
0·00
–0·20
–0·40
–0·60
–0·80
–1·00
Anxiety
(HSCL-25)

Depression
(HSCL-25)

Changes in psychological distress at follow-up
PTSD=post-traumatic stress disorder.

traumatic experiences were elicited from participants’
asylum applications. Detainees who could be located were
followed up 2 months or more after the initial interview to
assess psychological changes.
Spearman correlation coefficients were used to analyse
the relation between length of detention, which
had a skewed distribution, and psychological distress.
Independent sample t tests were used to assess differences
in psychological distress between baseline and follow-up
for patients who had been released and those still detained
at follow-up.
Between Jan 1, 2001, and June 15, 2002, 87 detainees
(73% of the caseload for the six agencies providing
referrals) were referred to the study. Of these 87, 17 were
excluded from the study: one was deported; ten were
released before interview; one had been granted asylum but
was still awaiting release; three did not complete the
interview questionnaires; one withdrew the asylum claim;
and one lost pro-bono legal support. Analyses are based on
the remaining 70 participants (56 male, 14 female, mean
age 28 years [SD 7·3; range 15–52]). Follow-up interviews
were done at a median of 101 days (62–299) with 61
participants; 35 were still in detention and 26 had been
released. Of the 26 released, 22 had been granted asylum
and four released without asylum for medical or
humanitarian reasons. Of the nine participants lost to
follow-up: two had been deported; five could not be
located; one had been transferred to another facility; and
one refused to be interviewed. In April, 2003, 40 (57%) of
the 70 participants had been granted asylum in the USA,
and 14 individuals (20%) had been denied asylum and
deported. Of the 40 individuals granted asylum, the
median length of detention had been 7 months (2–42).
61 (87%) participants were detained in two INS
detention centres and nine (13%) in three local
government-run jails. The median length of detention
before initial interview was 5 months (1–54). Most
participants were from Africa (n=54), seven were from
eastern Europe, four from Asia, two from the Middle East,
and three from South America. 28 interviews were done in
English, 17 in French, seven in Arabic, and 18 in other
languages.
52 (74%) detainees had been tortured before
immigration, 47 (67%) had been imprisoned in their native
country, 41 (59%) reported the murder of a family member

1722

or friend, and 18 (26%) reported
sexual assault. 54 (77%) detainees
had clinically significant symptoms of
anxiety, 60 (86%) of depression,
and 35 (50%) of post-traumatic
stress disorder. 18 (26%) participants
reported thoughts of suicide while in
detention, and two reported having
attempted suicide.
49 (70%) participants perceived their
mental health as having worsened
substantially while in detention, and
this perception was supported by
Spearman correlations between length
of time in detention and baseline
levels of anxiety (rs=0·34, p=0·004),
depression (0·28, p=0·017), and
post-traumatic stress disorder (0·28,
p=0·019). Baseline mental health
PTSD
scores did not differ significantly
(HTQ)
between detainees eventually released
and those who remained in detention,
although differences were significant at
follow-up (table). Participants still
detained at follow-up had increased symptom scores for
anxiety, depression, and post-traumatic stress disorder,
whereas those who had been released had lower scores on
all three scales (p<0·0001; figure).
Nearly all the detainees in our study had clinically
significant symptoms of anxiety, depression, or posttraumatic stress disorder, which worsened with time in
detention and improved on release. Our findings support
anecdotal observations of other researchers and highlight
the concerns raised by health professionals about the
adverse effect of detention on asylum seekers.1,2
A limitation of our study is that there was no
comparison group of non-detained asylum seekers.
Although our sample of released participants was
confounded by the fact that most were also granted asylum,
the significant correlation between symptom severity at
baseline and length of time in detention supports the
hypothesis that detention significantly contributed to
psychological distress. However, our reliance on self-report
questionnaires rather than diagnostic interviews might have
increased the proportion of individuals assessed as having
clinically significant distress. Additionally, sampling was
not random, which might limit the general applicability of
our results to the entire population of detained asylum
seekers. Furthermore, although we have no reason to think
that detained asylum seekers represented by pro-bono legal
groups differ from other detained asylum seekers, we
cannot be certain.
Some participants could have deliberately exaggerated
psychological symptoms, past traumatic experiences, or
both, in order to bolster their asylum claims—despite being
informed that their asylum application would not be
affected by their responses. Nevertheless, the large
proportion of participants ultimately granted asylum lends
credence to their reports. Furthermore, there was no
difference in reported distress at baseline or premigration
traumatic experiences between detainees who were or were
not granted asylum. Finally, despite a marked reduction in
symptoms after release, many participants still had high
levels of psychological distress, suggesting that symptom
endorsement was not solely motivated by secondary gain.
Despite the limitations of this study, our results suggest
that detaining asylum seekers exacerbates symptoms of
depression, anxiety, and post-traumatic stress disorder in
this vulnerable population. Our findings suggest that

THE LANCET • Vol 362 • November 22, 2003 • www.thelancet.com

For personal use. Only reproduce with permission from The Lancet.

RESEARCH LETTERS

policies concerning detention of asylum seekers should be
reviewed, and highlight the need for mental health
intervention to address the psychological needs of these
individuals.

Association of intercellular adhesion
molecule-1 gene with type 1
diabetes

Contributors
A Keller, D Ford, C Trinh-Shevrin, J Leviss, H Smith, K Allden, and
G Kim conceived and designed the study. A Keller, C Meserve,
E Sachs, J Leviss, E Singer, H Smith, and J Wilkinson participated in
the acquisition of data. E Sachs coordinated data collection.
C Trinh-Shevrin and B Rosenfeld did statistical analyses. A Keller,
B Rosenfeld, C Trinh-Shevrin, D Ford, C Meserve, J Leviss, E Sachs,
and B Rosenfeld participated in analysis and interpretation of data. All
authors helped to write the report.

Conflict of interest statement
None declared.

Acknowledgments
This study was done by the Bellevue/NYU Program for Survivors of
Torture and Physicians for Human Rights. We are grateful to
Eleanor Acer, Anwen Hughes, Lin Piwowarczyk, and Jon Hubbard for
their assistance in developing this project. Zachary Steele, Derrick
Silove, and Richard Mollica also provided invaluable insight into issues
concerning the health of detained asylum seekers. We thank the INS
staff at headquarters and at the detention facilities for their courtesy and
cooperation with this study. We thank the following organisations for
their cooperation in referring individuals for participation in this study:
American Friends Service Committee, Catholic Legal Immigration
Network, Circle York, The Lawyers Committee for Human Rights, New
York Association for New Americans, and Hebrew Immigrant Aid
Society. We also thank Suzanne Dieter for background research;
Melanie Jay, Alyssa Finlay, Amina Chaudry, and Joshua Lee for helping
with interviews; Caroline Lai and Angela Lee for data entry; and
Vincent Iacopino, Frank Davidoff, and Barbara Ayotte for reviewing the
manuscript. Support for this research was provided by a grant from the
Jeht Fund of the New York Community Trust and the Morton K and
Jane Blaustein Foundation. A Keller was supported by a Soros
Advocacy Fellowship from the Medicine as a Profession Program of the
Open Society Institute. Funding sources had no role in study design,
data collection, data analysis, data interpretation, writing of the report,
or decision to submit the paper for publication.
1

2

3
4

5

Keller A, Ford D, Sachs E, et al. From persecution to prison: the
health consequences of detention for asylum seekers. Boston and
New York City: Physicians for Human Rights and the Bellevue/NYU
Program for Survivors of Torture, 2003. http://www.phrusa.org/
campaigns/asylum_network/detention_execSummary/(accessed
October, 2003).
Silove D, Steel Z, Mollica R. Detention of asylum seekers: assault on
health, human rights, and social development. Lancet 2001; 357:
1436–37.
Amnesty International. Lost in the labyrinth: detention of asylum
seekers. New York: Amnesty International, 1999.
Mollica RF, Wyshak G, de Marneffe D, Khoun F, Lavelle J.
Indochinese versions of the Hopkins symptom checklist-25: a
screening instrument for the psychiatric care of refugees.
Am J Psychiatry 1987; 144: 497–500.
Mollica RF, Caspi-Yavin Y, Bollini P, Truong T, Tor S, Lavelle J.
The Harvard trauma questionnaire: validating a cross-cultural
instrument for measuring torture, trauma, and post traumatic stress
disorder in refugees. J Nerv Ment Dis 1992; 180: 111–16.

Bellevue/New York University (NYU) Program for Survivors of
Torture, Division of Primary Care Medicine, Department of
Medicine (A S Keller MD, C Meserve MD, H Smith PhD, E Sachs BA,
J Wilkinson MS), Institute for Urban and Global Health
(C Trinh-Shevrin MS), and Department of Emergency Services
(E Singer MD), NYU School of Medicine, New York, NY 10016,
USA; Department of Psychology, Fordham University, Bronx, NY
(B Rosenfeld PhD); New York City Health and Hospitals
Corporation, New York, NY (J A Leviss MD); Department of
Medicine, Harvard Medical School, Boston, MA, USA (G Kim MD);
Department of Psychiatry, Dartmouth Medical School, Hanover,
NH, USA (K Allden MD); and Physicians for Human Rights, Boston,
MA (D Ford JD, A S Keller)
Correspondence to: Dr Allen S Keller, Bellevue/NYU Program for
Survivors of Torture, Bellevue Hospital, 462 First Avenue, CD710,
New York, NY 10016, USA
(e-mail: [e-mail])

Sergey Nejentsev, Cristian Guja, Rose McCormack, Jason Cooper,
Joanna M M Howson, Sarah Nutland, Helen Rance, Neil Walker,
Dag Undlien, Kjersti S Ronningen, Eva Tuomilehto-Wolf,
Jaakko Tuomilehto, Constantin Ionescu-Tirgoviste, Edwin A M Gale,
Polly J Bingley, Kathleen M Gillespie, David A Savage,
Dennis J Carson, Chris C Patterson, A Peter Maxwell, John A Todd
Intercellular adhesion molecule-1 (ICAM-1) functions via its
ligands, the leucocyte integrins, in adhesion of immune cells to
endothelial cells and in T cell activation. The third immunoglobulin-like extracellular domain binds integrin Mac-1 and
contains a common non-conservative aminoacid polymorphism,
G241R. Phenotypically, ICAM-1 has been associated with type 1
diabetes, a T-cell-mediated autoimmune disease. We assessed
two independent datasets, and noted that R241 was associated
with lower risk of type 1 diabetes than is G241 (3695 families,
relative risk 0·91, p=0·03; 446 families, 0·60, p=0·006). Our
data indicate an aetiological role for ICAM-1 in type 1 diabetes,
which needs to be confirmed in future genetic and functional
experiments.

Lancet 2003; 362: 1723–24

The molecular mechanisms underlying type 1 diabetes are
partly known. Three disease loci have been identified so far:
the human leucocyte antigen (HLA) complex, the variable
number tandem repeat locus located in the promoter region of
the insulin gene (INS), and the cytotoxic T lymphocyteassociated antigen-4 gene (CTLA4).1 The known functions of
these genes suggest that T-cell activity is an important pathway
in development of type 1 diabetes. Results of research in the
mouse shows that ICAM-1 function during immune priming
is necessary for the generation of effector T cells capable of
destroying pancreatic insulin-producing ␤ cells.2 Genetic
analysis of the ICAM-1 gene in families affected by type 1
diabetes could indicate whether its function has a causal role in
the disease.
ICAM1 is located on chromosome 19p13 in a region that
has shown some evidence of linkage to type 1 diabetes.1 Two
non-synonymous single nucleotide polymorphisms are known
to be frequent in European populations: G→A in exon 4
encoding G241R, and A→G in exon 6 encoding K469E
(rs1799969 and rs5030382, respectively; http://www.ncbi.nlm.
nih.gov/SNP/). The K469E polymorphism has been investigated in type 1 diabetes in small samples, with variable results.3
We assessed association of these two ICAM1 coding polymorphisms in a sample of 3695 families affected by type 1 diabetes.
All family members were white, from Europe or the USA,
with at least one affected child in every family (table). Mean
age at onset of the affected offspring was 9·3 years (range
0–50). We obtained approval from the relevant research ethics
committees and written informed consent of participants.
DNA samples were genotyped with Invader (Third Wave
Technologies, Madison, WI, USA) and TaqMan (Perkin
Elmer Applied Biosystems, Foster City, CA, USA) assays. We
recorded 99·5% concordance between these methods in 1242
samples tested to assess error in genotyping of the G241R
polymorphism. Statistical analysis was done with STATA
(version 8.1). Calculations of p values and 95% CI of relative
risk (RR) were based on robust variance estimates, used to
correct for clustering of affected individuals within families.
K469E did not show any association with the disease: K469
was transmitted 1931 times of 3897 (49·6%, p=0·59). By
contrast, R241 was transmitted significantly less often: 938
transmissions of 1974 (47·5%, p=0·03; table). To control for
the possibility that genotyping error or transmission ratio

THE LANCET • Vol 362 • November 22, 2003 • www.thelancet.com

For personal use. Only reproduce with permission from The Lancet.

1723