The long-term trajectory of depression among Latinos in primary care
and its relationship to depression care disparities
Alejandro Interian, Ph.D.a,⁎, Alfonso Ang, Ph.D.b, Michael A. Gara, Ph.D.a,c,
Michael A. Rodriguez, M.D., M.P.H.b, William A. Vega, Ph.D.d
aDepartment of Psychiatry, UMDNJ-Robert Wood Johnson Medical School, Lyons, NJ 08854, USA
bDepartment of Family Medicine, University of California, Los Angeles, CA 90095, USA
cUMDNJ-University Behavioral Healthcare, Lyons, NJ 08854, USA
dUniversity of Southern California, School of Social Work, Los Angeles, CA 90089, USA
Received 3 September 2010; accepted 7 December 2010
Objective: Lower use of medication treatment, poorer doctor–patient communication (DPC) and depression stigma are key contributors to
mental healthcare disparities among Latinos with depression. The current study investigated the relationship between these key variables and
the long-term trajectory of depression in primary care among Latinos.
Method: Participants (N=220) were Latinos presenting to primary care who screened positive for depression. A repeated measures design
was used to assess participants at baseline and 6, 25 and 30 months. Repeated measures included depression (Patient Health Questionnaire-
9), self-reported quality of DPC and stigma pertaining to antidepressants. Using growth curve modeling, participants' depressive symptom
trajectories were examined for a 30-month period. Self-reported utilization of antidepressants, DPC and antidepressant stigma were
examined as predictors of the depressive symptom trajectory. In addition, rates of depression improvement/remission and recurrence/relapse
Results: Improvement/remission was experienced by 69.4% of participants during a 30-month period. Among those who improved/remitted
at 6 or 25 months, 63.4% maintained that improvement/remission by 30 months. The long-term trajectory of depressive symptoms
demonstrated a significant positive association with antidepressant stigma and significant negative associations with use of antidepressant
treatment and quality DPC.
Conclusions: While relapse/recurrence is common, most Latinos in this study experienced improvement in depression over 30 months. For
many, there is a considerable time to reach improvement/remission. Also, these findings confirm the significance of antidepressant
underutilization, DPC and stigma in the long-term outcome of depression among Latinos in primary care.
Published by Elsevier Inc.
Latinos in the US, particularly Mexican–Americans, are
less likely to receive guideline-concordant depression care
, and it is critical to examine how this problem, and its
determinants, are related to the long-term course of
depression among Latinos. The problem of mental health
care disparities among US Latinos, as well as other racial/
ethnic minority groups, has been given wide attention in the
research literature. Visibility to these disparity issues was
enhanced by the Surgeon General's report on culture, race
and ethnicity on mental health , as well as the Institute of
Medicine's outline of a broad range of healthcare disparities,
many of which centered on the lower utilization of mental
health care .
The adequate treatment of depression in primary care
among Latinos is a key area to address for reducing mental
health disparities. Latinos predominantly seek mental health
treatment from primary care [4,5], and a major source of
mental health disparities is the markedly low rate of
depression treatment , as 36% percent of Latinos with
Available online at www.sciencedirect.com
General Hospital Psychiatry xx (2011) xxx–xxx
⁎Corresponding author. Department of Veterans Affairs, New Jersey
Health Care System, Lyons Campus, Mental Health and Behavioral
Sciences, Lyons, NJ 07939-5000, USA. Tel.: +1 908 647 0180x4617.
E-mail address: firstname.lastname@example.org (A. Interian).
0163-8343/$ – see front matter. Published by Elsevier Inc.
depression receive treatment versus 60% of non-Latino
whites. Latinos are also more likely to prematurely
discontinue their antidepressant, relative to non-Latino
whites [7–10]. Also, evidence shows that Latinos initiate
antidepressant treatment at much lower rates, despite being
equally likely to receive a recommendation for an antide-
pressant from their primary care provider .
The stigma associated with antidepressant treatment and
poorer doctor–patient communication (DPC) were consid-
ered a causal factor in lower treatment engagement among
Latinos . Indeed, depression stigma is a particular
concern among a range of Latino groups [13–16], and this
has been shown to be associated with lower treatment
utilization . Thus, efforts that target stigma may
complement existing approaches that improve access to
guideline-concordant care among racial/ethnic minority
patients with depression [18,19].
An additional contributor to depression care disparities
pertains to DPC. There is less doctor–patient discussion of
antidepressant medication among Spanish-speaking Latinos
, a finding that is consistent with Latinos' overall
dissatisfaction with DPC . Doctor–patient communica-
tion may therefore be another target for reducing treatment
disparities, given that good communication is associated
with improved outcomes, adherence and patient satisfaction
in many areas of medicine [21–23].
These issues prompt the question, how do these key
disparity factors influence the course of depressive illness
among Latinos in primary care? It is likely that antidepres-
sant utilization, depression stigma and DPC interact with
one another to influence the trajectory of depression among
Latinos. The stigma associated with depression and its
treatment may leave patients more reluctant to utilize
antidepressant medication. Also, the medical encounter is a
critical point in the delivery of care, where skillful
communication can engage patients in a way that increases
their amenability to antidepressants and ameliorates the
stigma associated with these medications.
The objective of this study was to examine the course of
depression among Latino primary care patients during a
30-month period. Of particular interest was the role of
critical determinants of depression care disparities, such as
antidepressant utilization, depression stigma and DPC, in
influencing the illness course. The study focuses specifically
on antidepressant treatment, as engagement with this type of
treatment has been most concerning among Latinos given
lower utilization, lower adherence and lower preference for
antidepressants [6,7,24]. In addition, national trends in
depression care have shown that treatment is increasingly
relying on antidepressant treatment and less on psychothe-
rapy , significantly increasing the need to understand
engagement with antidepressant treatment among Latinos.
Since our study focuses on the long-term course of
depressive symptoms among Latinos in primary care, we
employed latent variable curve growth modeling to estimate
the trajectory of depressive symptoms, allowing an assess-
ment of the variables that significantly influence that
trajectory. We hypothesized that the course of depressive
symptoms would be improved by antidepressant utilization
and high ratings of DPC, while worsened by stigma.
The data were collected at two large primary care clinics
for underserved populations between January 2006 and June
2008 and have been reported elsewhere . These clinics
were run by Los Angeles County, and most participants had
public forms of insurance, such as Medi-Cal and plans for
reduced cost for outpatient services in public clinics. The
recruitment process involved approaching patients in the
clinic and explaining the study's purpose. Those who
consented to participate were screened for eligibility and
enrolled in the study if they met the following criteria: (a)
screened positive on the Patient Health Questionnaire-2
(PHQ-2) for depression, score ≥3; [27,28] (b) were 18 years
of age or older and (c) spoke English or Spanish. A total of
220 Latinos with probable depression was enrolled based on
this process. Given that the current study emphasized a
naturalistic design, the PHQ-2 was utilized as the depression
criterion, as we sought to describe the course of illness for
individuals who are evaluated in primary care in two-step
process that is geared for busy primary care settings (i.e.,
screened with the PHQ-2, more thoroughly assessed with the
Once enrolled, participants completed a baseline inter-
view (Time 1) in their preferred language (English or
Spanish). These interviews were conducted again after 6
(Time 2), 25 (Time 3) and 30 months (Time 4) from baseline.
Demographic variables were collected during the baseline
assessment, while information pertaining to depressive
symptoms, antidepressant utilization and DPC were collect-
ed during each of the assessments. Stigma related to
antidepressants was assessed during the final two assessment
points (Times 3 and 4). All participants signed an
institutional review board (IRB)-approved consent form
and study procedures were approved by the IRB at
University of California, Los Angeles.
A demographic form inquired about participants' status
on basic demographic variables, such as age, gender, years of
education, employment status, marital status and availability
of health insurance. Antidepressant treatment utilization was
assessed by asking participants at each time point whether
they were “currently taking antidepressants.”
2A. Interian et al. / General Hospital Psychiatry xx (2011) xxx–xxx
2.3.1. Depressive symptoms
The assessment of depression utilized a two-stage
procedure, as recommended by Kroenke et al.  for
busy primary care settings. This assessment strategy relied
on first screening patients with the briefer Patient Health
Questionnaire-2 (PHQ-2) , followed by the longer
PHQ-9 . The PHQ-2 allowed for rapid screening of
depression, while the PHQ-9 permitted for lengthier
assessment of depression. The PHQ-9 also served as the
repeated measure for assessing changes in depressive
symptom levels. The PHQ-2 is a two-item measure, with
each item scored 0–3, resulting in a core of 0–6. A PHQ-2
score of ≥3 has demonstrated a sensitivity of 83% and
specificity 92%, in comparison with a clinician structured
diagnostic interview. The PHQ-9 is a nine-item measure,
with each item scored 0–3, that yields a score ranging from
0–27. Higher scores indicate greater depressive symptoms.
The validity of the PHQ-9 has been supported, including
criterion-related validity with a clinician diagnostic interview
for depression and convergence with other depressive
measures. A score ≥10 served as the threshold for
determining probable depression .
2.3.2. Stigma towards antidepressants
The Latino Scale for Antidepressant Stigma (LSAS)
assesses the stigma concerns that Latinos have reported
regarding antidepressants . It contains seven items with
stigma-related statements pertaining to antidepressants that
are scored on a three-point scale, yielding a score ranging
from 0 to 14. Participants indicate the degree to which they
feel others may agree with each statement and higher scores
indicate greater concerns about how others would view
antidepressant use. The LSAS has acceptable internal
consistency, has a sound factor structure and is associated
with 23% lower utilization of antidepressants. Administra-
tion of the LSAS occurred during the final time points in the
study (Times 3 and 4).
2.3.3. Doctor–patient communication
Items measuring DPC were derived from a previous study
. The current measure included six items rated on a five-
point Likert scale that inquired about the degree to which the
providers explained things, demonstrated respect, listened,
asked for preferences, involved patient in decisions and
worked with the patient to develop a plan. Scoring involved
dividing the total sum by the maximum possible total score,
yielding a score from 0 to 1. Using the current sample, the
DPC scale produced a Cronbach coefficient of .80 at each
In order to utilize complete data, analyses for the current
study focused on the 200 participants who completed the
30-month assessment. A number of analyses were conducted
to describe the number of participants throughout the
different time points according to a criterion of probable
depression (i.e., PHQ-9N9). First, we calculated the
frequency and percentage of participants who were probably
depressed, separately by time point. Next, a calculation was
used to describe the frequency and percentage of participants
who demonstrated their first improvement/remission, sepa-
rately by time point. We operationalized this with a criterion
below the cutoff indicating probable depression (i.e.,
PHQ-9b10). In addition, of those who demonstrated an
improvement/remission, we calculated the frequency who
maintained a below threshold score throughout the remain-
ing time points, as well as the frequency of those who
experienced a recurrence/relapse, which was operationalized
according to meeting criteria for probable depression (i.e.,
Subsequent analyses used a latent variable curve growth
modeling approach, using Mplus software . Growth
modeling allows for an estimate of the growth trajectory by
examining the rate of change on the depression outcome
variable (PHQ-9) over time. The analysis also permits
examining the effect of predictor variables while also
adjusting for covariate variables. Growth modeling requires
fewer subjects to detect a small effect size at a power of .80,
in comparison to traditional repeated measures analyses such
as analysis of variance or analysis of covariance . With
four waves of data over a 30-month period, a linear trajectory
shape was hypothesized. Each individual growth trajectory is
characterized by an intercept, which represents the initial
status or starting point of the outcome, and a slope, which
represents the constant change rate over time. The first step
in this analysis was to describe the trajectory of depression
symptoms. We accomplished this by using measures at four
time points (baseline and Months 6, 25, and 30) to estimate
unconditional growth models for depressive symptomato-
logy. In the next step, we specified the conditional model to
determine the effects of the predictor variables (i.e.,
antidepressant treatment utilization, DPC, stigma), while
adjusting for covariates (age, education, gender, having
health insurance), on the trajectory of depressive symptoms.
Model fit was assessed by the following global fit indices:
chi-square values, Comparative Fit Index (CFI) (with a
cutoff value of .95) and Root Mean Square Error of
Approximation (RMSEA) (with a cutoff value of .06) .
3.1. Sample characteristics
The demographic characteristics of the participants are
presented in Table 1. Most Latino participants were female,
Spanish speakers, with less than a high school education.
Nearly all participants had access to health insurance. Only
a small proportion of participants were fully employed.
Table 2 summarizes the repeated measures across the four
time points to allow the reader to examine specific scores or
rates of antidepressant medication use at the particular time
point. Table 2 shows that at any given time point, a quarter
3A. Interian et al. / General Hospital Psychiatry xx (2011) xxx–xxx
to a third of all participants reported utilizing antidepressant
medication. Also, the mean PHQ-9 score at the first two
time points met the criterion for probable depression and
the converse was the case for the final two time points.
3.2. PHQ-9 caseness across time
A total of 86% of the sample had a PHQ-9 score
indicating probable depression at baseline (PHQ-9N9).
Table 3 summarizes PHQ-9 outcomes during subsequent
time points. The first column reports the rate of participants
during Time 2 (6 months), Time 3 (25 months) or Time 4
(30 months) that scored above the cutoff for probable
depression (PHQ-9 b10). By 25 months, fewer than half of
participants scored in the probably depressed range. In the
second column, the percentage of participants who
improved/remitted for the first time since baseline was
26.6% by Time 2, 31.8% by Time 3 and 11% by Time 4,
resulting in a total of 69.4% who improved/recovered at
some point by 30 months. As shown in the third column, of
those who improved/recovered at Time 2, 63% maintained
this status at Time 4. Also, 63.6% of those improving/
recovering at Time 3, maintained their below threshold
depression at Time 4. Thus, of the 101 who improved/
recovered at Times 2 or 3, 63.4% maintained a below
threshold status at Time 4. Conversely, 36.6% had a
recurrence/relapse by Time 4.
3.3. Trajectory of depressive symptoms
The latent trajectory model result showed a good fit (chi-
square p=.41; CFI=.981; RMSEA=.025). Table 4 sum-
marizes the results of the latent trajectory modeling. The
intercept of PHQ-9 scores was significant, indicating that
participants had significantly different PHQ-9 scores at
baseline. More importantly, the model showed a significant
slope or trajectory of PHQ-9 scores, indicating that
depression severity significantly decreased across time.
Several predictors were independently significant with the
30-month trajectory of PHQ-9 symptoms, adjusting for
gender, education, age and health insurance status (none of
these covariates were significant in the model). Three
antidepressant utilization predictors (Time 1, Time 2,
Time 4) showed significant negative associations with the
trajectory of PHQ-9 score, where self-reported “taking
antidepressants” was associated with a trajectory of decreas-
ing PHQ-9 scores. DPC during Time 4 also demonstrated a
significant negative association with the trajectory of PHQ-9
scores. Higher ratings of DPC were associated with a
decreasing trajectory of depression. This measure was not
significantly related to the depression trajectory during
Times 1 and 2. Stigma during Time 4 was also significantly
associated with the PHQ-9 trajectory in the positive
direction, where higher stigma scores were significantly
associated with increasing PHQ-9 scores. Stigma during
time 3 was not significantly related to depression trajectory.
A review of the coefficients in Table 4 shows that stigma
scores had the strongest relationship to the trajectory of
PHQ-9 scores. This relationship is illustrated in Fig. 1,
which plots the course of PHQ-9 symptoms, according to
high and low levels of stigma (divided using a median split).
This plot shows that the PHQ-9 scores for participants
Sociodemographic characteristics of the sample (N=200)
65 and older
Less than high school
Some high school
High school graduate
College graduate or additional higher education
Full-time or part-time
Full-time or part-time students
Single or never married
Married or living with partner
Separated or divorced
Depression severity, antidepressant utilization, stigma, and DPC across the
four time points (N=200)
Baseline (time 1)
6 months (time 2)
25 months (time 3)
30 months (time 4)
25 months (time 3)
30 months (time 4)
6 months (time 2)
25 months (time 3)
30 months (time 4)
Depression care utilization
Currently taking antidepressants (time 1)
Currently taking antidepressants (time 2)
Currently taking antidepressants (time 3)
Currently taking antidepressants (time 4)
aPatient Health Questionnaire–9.
bLatino Scale for Antidepressant Stigma.
4A. Interian et al. / General Hospital Psychiatry xx (2011) xxx–xxx
scoring high and low on time 4 stigma (30 months) start to
diverge after baseline.
Overall, depression does likely remit after two and a half
years, but it is likely to recur. Our results showed that
approximately 55% no longer met criteria for probable
depression by 25 months. This indicates that many patients
improve, approximating results reported in a previous non-
Latino primary care sample, where approximately 45% no
longer met major depression criteria at 24 months . In the
current study, relapse/recurrence seems to be concerning,
given that 37% of those whose depression improved/
recovered experienced a relapse/recurrence of symptoms.
This finding is consistent with the attention given to relapse/
recurrence in the depression treatment research literature
[35,36]. In addition, it is concerning that only a minority of
participants had improved/recovered at 6 months, leaving a
considerable number to experience the burden and suffering
of depression for substantial period of time. Thus, our data
show that efforts should target earlier improvement/remis-
sion among Latino primary care patients, as well as reduce
rates of relapse/recurrence. Our overall findings pertaining to
relapse/recurrence, in combination with high rates of
underutilization of treatment, are consistent with recent
findings showing that minority groups are more likely to
experience recurrent types of major depression compared to
non-Latino whites .
Our study suggests that improving key mental health
disparity determinants hold promise for achieving earlier
treatment benefit and maintenance of gains, as our predictors
were shown to influence the course of depressive symptoms.
Latent trajectory model of PHQ-9 (depression) trajectory over 30 months
Slope (PHQ-9 Trajectory)
Demographic variables (covariates)
High school graduate
Health insurance status
Depression outcome rates across a 30-month period
Time periodProbable depression
Baseline (Time 1)
6 months (Time 2)
25 months (Time 3)
30 months (Time 4)
aFigures represent the frequency of participants scoring below probable depression for the first time since baseline, across time periods. The subsample
includes the 173 participants with a probable depression at baseline (PHQ-9N9).
bFigures represent frequency of participants who were who, after demonstrating improvement/remission, maintained a subthreshold depression score at
30 months,separately by time period. Percentages were calculated based on the number showing improvement/remission during each correspondingtime period.
cFigures represent frequency of participants who, after demonstrating improvement/remission, experienced relapse/recurrence during one of the subsequent
time periods. Percentages were calculated based on the number showing improvement/remission during each corresponding time period.
Note. Lines represent stigma as measured by the LSAS and were
dichotomized using a median split.
Fig. 1. PHQ-9 depression scores across time for patients with high and
5A. Interian et al. / General Hospital Psychiatry xx (2011) xxx–xxx
Of note was the important role of stigma as a barrier to
depressive symptom remission among Latino primary care
patients with depression. Specifically, the results of this
study showed that the presence of stigma was associated
with an increased likelihood of depressive symptom
persistence. This effect was observed to be independent
of antidepressant use. Such a finding adds to the accruing
literature documenting the treatment complicating effect of
stigma on mental health outcomes [38–40]. Stigma is a
frequently discussed barrier to mental healthcare among
Latinos and the current study bolsters the significance of
this issue by documenting the deleterious effect of stigma
on long-term depression outcomes among Latinos in
The results of the study also showed low rates of
antidepressant utilization. Our rates of self-reported antide-
pressant utilization, ranging from 28 to 33%, are consistent
with rates of antidepressant utilization previously reported
among Latinos with depression in primary care .
Moreover, our findings directly link antidepressant utiliza-
tion with long-term outcomes, showing that non-utilization
worsened the trajectory of depressive symptoms. It also
seems likely that these low rates of antidepressant utilization
contributed to the low number of participants achieving
depression symptom improvement/remission by 6 months.
While a consideration is whether participants accessed other
treatments in specialty care settings, our data indicate that
very small numbers attended specialty care (i.e., between 3
and 5 participants at each time point). Thus, the study's
results confirm that improving the utilization of antidepres-
sants in primary care among Latinos remains a priority.
We further hypothesized that DPC that was perceived to
be favorable would enhance depression outcomes, a finding
that was partially supported by the growth modeling. This
finding adds to the accumulating literature that links
important health outcomes to DPC [21–23]. The current
results support the examination of communication ingredi-
ents that may be particularly effective for Spanish-speaking
Latinos. These findings also support the use of interventions
targeting the communication of Latino patients  or
physicians , particularly those that are patient centered.
The results of the current study provide support for a
number of clinical services and practice strategies. Strategies
to increase engagement with antidepressant treatment will
likely yield improved outcomes. Currently, there are a
number of interventions available that hold promise
improving engagement among Latinos [41,43]. These
interventions also have the potential to address depression
stigma by either increasing patient treatment empowerment
or by helping patients view antidepressant treatment in ways
that are more consonant with models of coping within their
community (e.g., struggling against problems instead of
taking medications). Engagement strategies can also include
awareness efforts that utilize culturally-focused educational
approaches such as fotonovelas . Finally, while DPC can
be improved by adopting key components of patient-
centered communication (e.g., eliciting patient treatment
preference, collaborative treatment planning), it should also
include communicating an understanding of the stigma
issues associated with antidepressant use among Latinos
. Such communication can also help patients find ways
in which antidepressant treatment can be viewed compatibly
with their own models of coping with depression .
The current study has a number of limitations. First,
our results linking depression trajectory and stigma are
correlative and cannot differentiate between a number of
interpretations. For example, it may be that worsened course
of symptoms leads to greater stigma, as self-esteem and
social functioning may be negatively affected by the
persistence of depressive symptoms. Another interpretation
is that the perception of stigma towards depression
complicated remission. Accordingly, baseline stigma has
been shown to prospectively predict self-esteem at 6 and
24 months follow-up.Consistentwiththisinterpretation,
stigma can be seen as early as 6 months. While another
interpretation to consider is that stigma negatively influences
depression remission via reduced treatment engagement, our
findings show that stigma's effect on depression is indepen-
dent of antidepressant utilization. An addition interpretation
would be that depression severity and stigma interrelate in
multiple ways, having reciprocal effects, including an
interplay with treatment engagement. While we cannot
differentiate between these interpretations, our results
highlight the importance of stigma in the long-term course
of depressive illness among Latinos in primary care.
Second, although the PHQ-9 has good agreement with a
clinician interview for major depression , it is not a
structured psychiatric research interview for assessing major
depression. Therefore, care should be taken when comparing
rates of depression recurrence with those of previous primary
care studies [34,35], as these studies employed a structured
diagnostic interview. The current results should therefore be
interpreted as a preliminary estimate of the course of
depression among Latinos in primary care.
Third, a measure of stigma was included in our research
protocol during the later time points of the study, which
provided an incomplete picture of how stigma affects care at
all data points. While our results establish the key influence
of antidepressant stigma during late follow-up periods, they
do not help us understand the role of stigma when patients
Fourth, some caution is warranted in interpreting the
results given that the majority of this Latino sample had
access to health insurance. This may raise some questions
pertaining to generalizability given that low rates of health
insurance funding are observed among Latinos . Despite
this limitation, however, this study adds to existing evidence
showing that disparities in care continue to be observed, even
among Latinos who have access to health insurance [11,46].
In conclusion, our results illustrate the long-term course
of depressive symptoms among Latinos in primary care and
6 A. Interian et al. / General Hospital Psychiatry xx (2011) xxx–xxx
that it is characterized by slow improvement and relapse/
recurrence in a significant proportion of cases. Issues
frequently discussed as relevant to Latino depression care
disparities play a key role in the long-term course of
depressive illness in primary care.
This project was supported by grants 62454 and 62609
from the Robert Wood Johnson Foundation (Dr. Vega,
principal investigator). Dr. Interian was also funded by grant
K23-MH074860 from the National Institute of Mental
Health. This project was also supported by the Robert
Wood Johnson Foundation's Finding Answers: Disparities
Research for Change National Program (RWJF ID 59748).
Dr. Interian completed this work while at UMDNJ-Robert
Wood Johnson Medical School.
 Gonzalez HM, Vega WA, Williams DR, et al. Depression Care in the
United States: too little for too few. Arch Gen Psychiatry 2010;67:
 Department of Health & Human Services. Mental health: culture, race,
and ethnicity — a supplement to mental health: a report of the surgeon
general. Rockville (MD): U.S. Department of Health and Human
 Smedley BD, Stith AY, Nelson AR. Unequal treatment: confronting
racial and ethnic disparities in healthcare. Washington (DC): National
Academies Press; 2003.
 Vega WA, Kolody B, Aguilar-Gaxiola S, et al. Gaps in service
utilization by Mexican Americans with mental health problems. Am J
 Alegria M, Mulvaney-Day N, Woo M, et al. Correlates of past-year
mental health service use among Latinos: results from the National
Latino and Asian American Study. Am J Publ Health 2007;97:76–83.
 AlegriaM, Chatterji P, WellsK, et al. Disparityin depressiontreatment
among racial and ethnic minority populations in the United States.
Psychiatr Serv 2008;59:1264–72.
 Olfson M, Marcus SC, Tedeschi M, et al. Continuity of antidepressant
treatment for adults with depression in the United States. Am J
 Sleath B, Rubin RH, Huston SA. Hispanic ethnicity, physician-patient
communication, and antidepressant adherence. Compr Psychiatry
 Warden D, Rush AJ, Wisniewski SR, et al. What predicts attrition in
second step medication treatments for depression?: a STAR⁎D Report.
Int J Neuropsychopharmacol 2009;12:459–73.
 Warden D, Trivedi M, Wisniewski SR, et al. Predictors of attrition
during initial (Citalopram) treatment for depression: a STAR⁎D report.
Am J Psychiatry 2007;164:1189–97.
 Miranda J, Cooper LA. Disparities in care for depression among
primary care patients. J Gen Intern Med 2004;19:120–6.
 Vega WA, Karno M, Alegria M, et al. Research issues for improving
treatment of U.S. Hispanics with persistent mental disorders. Psychiatr
 Cabassa LJ, Lester R, Zayas LH. “It's like being in a labryrinth:”
Hispanic immigrants' perceptions of depression and attitudes towards
treatments. J Immigr Minor Health 2007;9:1–16.
 Interian A, Martinez IE, Guarnaccia PJ, et al. A qualitative analysis of
the perception of stigma among Latinos receiving antidepressants.
Psychiatr Serv 2007;58:1591–4.
 Martinez Pincay IE, Guarnaccia PJ. “It's like going through an
earthquake”: Anthropological perspectives on depression among
Lation immigrants. J Immigr Health 2007;9:1557–912.
 Nadeem E, Lange JM, Edge D, et al. Does stigma keep poor young
immigrant and U.S.-born black and Latina women from seeking
mental health care? Psychiatr Serv 2007;58:1547–54.
 Interian A, Ang A, Gara MA, et al. Stigma and depression treatment
utilization among Latinos: utility of four stigma measures. Psychiatr
 Miranda J, Duan N, Sherbourne C, et al. Improving care for minorities:
can quality improvement interventions improve care and outcomes for
depressed minorities? Results of a randomized, controlled trial. Health
Serv Res 2003;38:613–30.
 Wells KB, Sherbourne C, Schoenbaum M, et al. Impact of
disseminating quality improvement programs for depression in
managed primary care: a randomized controlled trial. JAMA 2000;
 Morales LS, Cunningham WE, Brown JA, et al. Are Latinos less
satisfied with communication by health care providers? J Gen Intern
 Bertakis K, Roter D, Putnam S. The relationship of physician medical
interview style to patient satisfaction. J Fam Pract 1991;32:135–6.
 Bull SA, Hu XH, Hunkeler EM, et al. Discontinuation of use and
switching of antidepressants: influence of patient-physician commu-
nication. JAMA 2002;288:1403–9.
 Clever SL, Ford DE, Rubenstein LV, et al. Primary care patients'
involvement in decision-making is associated with improvement in
depression. Med Care 2006;44:398–405.
 Cooper LA, Gonzales JJ, Gallo JJ, et al. The acceptability of treatment
for depression among African-American, Hispanic, and white primary
care patients. Med Care 2003;41:479–89.
 Olfson M, Marcus SC. National patterns in antidepressant medication
treatment. Arch Gen Psych 2009;66:848–56.
 Bazargan M, Ani C, Hindman D, et al. Correlates of complementary
and alternative medicine utilization in depressed, underserved African
American and Hispanic patients in primary care settings. J Altern
Complement Med 2008;14:537–44.
 KroenkeK,Spitzer RL, Williams JB.The PatientHealth Questionnaire-
2: validity of a two-item depression screener. Med Care 2003;41:
 Löwe B, Kroenke K, Gräfe K. Detecting and monitoring depression
with a two-item questionnaire (PHQ-2). J Psychosom Res 2005;58:
 Kroenke K, Spitzer RL, Williams JB. The PHQ-9: validity of a brief
depression severity measure. J Gen Intern Med 2001;16:606–13.
 Hays RD, Shaul JA, Williams VSL, et al. Psychometric Properties
of the CAHPS(TM) 1.0 Survey Measures. Med Care 1999;37:
 Muthén LK, Muthén BO. Mplus user's guide. 6th ed. Los Angeles:
Muthén & Muthén; 2010. Available at: http://www.statmodel.com/
 Curran P, Muthén B. The application of latent curve analysis to testing
developmental theories in intervention research. Am J Community
 Hu LT, Bentler P. Cutoff criteria for fit indexes in covariance structure
analysis: Conventional criteria versus new alternatives. Struct Equ
 Simon GE. Long-term prognosis of depression in primary care. Bull
World Health Organ 2000;78:439–45.
 Lin EH, Katon WJ, VonKorff M, et al. Relapse of depression in
primary care. Rate and clinical predictors. Arch Fam Med 1998;7:
 Segal ZV, Pearson JL, Thase ME. Challenges in preventing relapse in
major depression: Report of a National Institute of Mental Health
Workshop on state of the science of relapse prevention in major
depression. J Affect Disord 2003;77:97–108.
7A. Interian et al. / General Hospital Psychiatry xx (2011) xxx–xxx
 González HM, Tarraf W, Whitfield KE, Vega WA. The epidemiology Download full-text
of major depression and ethnicity in the United States. J Psychiatr Res
 Link BG, Struening EL, Neese-Todd S, et al. Stigma as a barrier to
recovery: The consequences of stigma for the self-esteem of people
with mental illnesses. Psychiatr Serv 2001;52:1621–6.
 Perlick DA, Rosenheck R, Clarkin JF, et al. Stigma as a barrier to
recovery: adverse effects of perceived stigma on social adaptation of
persons diagnosedwith bipolaraffectivedisorder.PsychiatrServ2001;
 Pyne JM, Kuc EJ, Schroeder PJ, et al. Relationship between
perceived stigma and depression severity. J Nerv Ment Dis 2004;
 Alegria M, Polo A, Gao S, et al. Evaluation of a patient activation and
empowerment intervention in mental health care. Med Care 2008;46:
 Loh A, Simon D, Wills CE, et al. The effects of a shared decision-
making intervention in primary care of depression: a cluster-
randomized controlled trial. Patient Educ Couns 2007;67:324–32.
 Interian A, Martinez I, Iglesias Rios L, et al. Adaptation of a
motivational interviewing intervention to improve antidepressant
adherence among Latinos. Cult Div Ethnic Min Psychol 2010;16:
 Cabassa L, Molina G, Baron M. Depression Fotonovela: development
of a depression literacy tool for Latinos with limited English
proficiency. Health Promot Pract 2010 Nov 4 [Epub ahead of print].
 Lillie-Blanton M, Hoffman C. The role of health insurance coverage in
reducing racial/ethnic disparities in health care. Health Aff 2005;24:
 Lagomasino IT, Dwight-Johnson M, Miranda J, et al. Disparities in
depression treatment for latinos and site of care. Psychiatr Serv 2005;
8 A. Interian et al. / General Hospital Psychiatry xx (2011) xxx–xxx