Impact of HIV Antiretroviral Therapy on Depression and Mental Health

Among Clients With HIV in Uganda

GLENN J. WAGNER, PHD, BONNIE GHOSH-DASTIDAR, PHD, JEFFREY GARNETT, MPP, CISSY KITYO, MBCHB,
AND PETER MUGYENYI, MBCHB
Objective: With wide-reaching harmful effects of depression, and the absence of psychiatric treatment in most HIV care programs
in sub-Saharan Africa, we examined the effects of antiretroviral therapy (ART) on depression and other mental health indicators.
Methods: 602 patients (302 non-ART, 300 ART) were followed for the rst 12 months of HIV care in Uganda, with assessments
at entry into care and Months 6 and 12. Mental health was assessed with measures of depression, hopelessness, and internalized
HIV stigma; physical health functioning was assessed as an explanatory variable. Results: Thirteen percent had clinical depression,
57% had elevated depressive symptoms, and CD4 cell count was negatively correlated with measures of depression at baseline.
Signicant reductions in elevated depressive symptoms (time: odds ratio [95% condence interval] = 0.53 [0.43Y0.64]) and hopelessness (time: A = j0.12, p G .001) were observed in both the ART and non-ART groups, but the drop in depression was greater
among ART patients in intention-to-treat multivariate analysis (ART  time: p G .001). When added to the regression models, change
in physical health functioning predicted positive longitudinal change on measures of depression, hopelessness, and internalized stigma (all
p values G .001), yet ART status remained a signicant independent predictor of each (ART  time: p values ranged from G .05 to G .001).
Most mental health benets of ARTwere experienced in the rst 6 months of care. Conclusions: These ndings demonstrate the mental
health benets of HIV care and ART. However, in some people, mental health problems persist once physical health is stabilized, in
which case mental health treatment may be needed. Key words: depression, mental health, antiretroviral therapy, HIV, Uganda.

HIV = human immunodeciency virus; AIDS = acquired immunodeciency syndrome; ART = antiretroviral therapy; PHQ-9 = nineitem Patient Health Questionnaire; D-HSCL = depression subscale
of the Hopkins Symptom Checklist; PLHA = people living with HIV/
AIDS; ITT = intention to treat.

INTRODUCTION
esearch reveals that depression is equally as prevalent among
people living with human immunodeciency virus (HIV)/
acquired immunodeciency syndrome (AIDS) (PLHA) in subSaharan Africa as in their counterparts in the United States
(1,2), with rates ranging from 15% to 30% when assessed using
diagnostic interviews (3Y5) and 30% to 50% when using selfreports (3,4,6,7), including studies conducted in Uganda (8,9).
Although diagnostic criteria and measurement scales for depression have been developed in Western countries, cross-cultural
ethnographic research on depression among PLHA in Uganda
supports the validity of the Western concept of depression in this
context, with presenting symptoms being similar to classic major
depressive disorder (10,11), although patients typically emphasize somatic symptoms (e.g., fatigue, difculty sleeping, poor
appetite) during initial presentation.
The high prevalence of depression among PLHA is troubling not only because of its obvious effects on psychological
well-being and quality of life (12,13) but also because of the
evidence for its relationship to poor adherence to (14,15) and
outcomes of HIV treatment (16Y18), more rapid HIV disease
progression (19Y21), engagement in sexual risk behavior (22,23),
and difculties in work performance and functioning (24,25).

From the RAND Corporation (G.J.W., B.G.-D., J.G.), Santa Monica, California;
and Joint Clinical Research Centre (C.K., P.M.), Kampala, Uganda.
Address correspondence and reprint requests to Glenn J. Wagner, PhD, RAND
Corporation, 1776 Main St, Santa Monica, CA 90407. E-mail: gwagner@rand.org
This research is supported by a grant from The Rockefeller Foundation
(Grant No. HE 007; principal investigator: G.J.W.).
Received for publication November 20, 2011; revision received May 1, 2012.
DOI: 10.1097/PSY.0b013e31826629db

The wide reaching harmful effects of depression highlight the

need for effective clinical management of depression in PLHA,
yet in sub-Saharan Africa, depression and mental health disorders in general are rarely diagnosed let alone treated (26). Although billions of dollars in foreign aid have resulted in greatly
improved access to HIV antiretroviral therapy (ART) in the
region, it is strikingly rare for any ART programs to include
professional mental health services. Simultaneously, HIV medical care and ART are being scaled up and becoming more accessible to this population. In the absence of psychiatric care,
it is particularly important to understand the effects that ART
and HIV medical care can have on mental health and depression
in particular.
There is a scarcity of research on the impact of ART on
mental health and depression, and ndings are mixed from
published studies in sub-Saharan Africa. Studies in Uganda
(27,28), Zimbabwe (13), and South Africa (29) have reported
ART to be associated with reduced depression and improved
quality of life, whereas a study in Mozambique found that depression increased after 1 year of ART (30). It is plausible for
HIV treatment to have both positive and negative effects on
mental health. Treatment may alleviate depression and improve
overall outlook on life via improved physical health functioning, ability to work and provide for ones family, and related
hope and optimism for long-term survival and well-being. Better
physical health and associated improvements in social and economic functioning could also aid individuals in adjusting psychologically to an HIV diagnosis and reduce internalized HIV
stigma. Yet, ART may also contribute to chronic distress associated with having to adhere to strict lifelong regimens, the
constant reminder of illness from daily pill taking, and the interference in daily activities and stigmatization associated with
treatment and its potential adverse effects (31).
In this article, we report ndings from a study of the impact
of ART on mental health outcomes among new clinic patients
in Uganda who were followed up for the rst 12 months of care.
We examined the effects of ART with regard to depression, as

Psychosomatic Medicine 74:883Y890 (2012)

0033-3174/12/7409Y0883
Copyright * 2012 by the American Psychosomatic Society

Copyright 2012 by the American Psychosomatic Society. Unauthorized reproduction of this article is prohibited.

883

G. J. WAGNER et al.
well as hopelessness and internalized HIV stigma, and the role
of physical health functioning as a key inuential variable in
determining the effects of HIV treatment.
METHODS
Study Design
Participants enrolled in a prospective cohort study of the effects of ART on
multiple health outcomes. Half the sample was starting ART, and half was not
yet eligible for ART but had signs of disease progression (CD4 cell count G400
cells/mm3). Assessments were made at study enrollment and Months 6 and
12 for all participants; however, non-ART patients who started ART during the
course of the study were administered the assessment at ART initiation and then
6 months later. Data collection took place between January 2008 and September
2009. Participants were paid 5000 Uganda shillings ($2.50) after each assessment. The protocol was approved by the institutional review board at
RAND Corporation and the Joint Clinical Research Centre.

Sample
The study was conducted at two HIV clinics operated by the Joint Clinical Research CentreVone in Kampala, the countrys capital, and the other in
Kakira, a small town approximately 100 km outside Kampala. Between January
and September 2008, consecutive new clinic clients who had just completed
evaluation for ART eligibility were approached to participate if they a) were
18 years or older, b) were about to start ART if eligible, c) or had CD4 cell count
less than 400 cells/mm3 if not ART eligible. The primary criterion for ART
eligibility was having either a CD4 cell count of 250 cells/mm3 or fewer or a
World Health Organization Stage 3 or 4 disease (AIDS diagnosis), but patients
also needed to demonstrate treatment readiness by identifying a treatment
supporter (typically a relative or a friend who would support the patients
access and adherence to treatment) and attending clinic regularly. Although
statistics on study refusal were not recorded, the study interviewers indicate that
nearly all clients who were eligible agreed to enroll in the study and provided
informed written consent. All participants received HIV primary medical care,
which includes monitoring and treatment of active infections, and prescription
of appropriate prophylactic medications. Clients on ART come to the clinic on
a monthly basis at rst and then bimonthly, whereas the interval between clinic
appointments for non-ART patients is between 2 and 6 months depending on
their CD4 cell count. Psychiatric care was not available at the study sites during
the study period, and antidepressants were not used to treat depression; however,
counseling services were available to clients when requested or recommended by
the provider, and generally consisted of pre- and post-HIV test counseling, ART
adherence counseling, and counseling related to HIV disclosure and sexual and
reproductive health issues.

Measures
All measures were translated into Luganda using standard translation and
back-translation methods and were interviewer administered. The interviewers
were research study coordinators on staff at the clinics and were not involved
in provision of HIV care. For all scales, higher scores represent greater levels
of the construct.

Background and Demographic Characteristics

These included age, sex, level of formal education, relationship status, and
work status over the past 7 days.

Physical Health
CD4 cell count and HIV disease stage at the time of entry into care were
abstracted manually from the patients medical chart. Physical health functioning was measured with the six-item subscale of the Medical Outcomes
Study HIV Health Survey (32), which assesses the extent to which the respondent feels limited by their health in being able to engage in a range of daily
activities (e.g., lifting heavy objects; moving a table or carrying groceries;
walking one block; eating, dressing, bathing, or using toilet); response options
include yes, limited a lot, yes, limited a little, and no, not limited at all;
884

and the subscale score is standardized on a scale of 0 to 100. Cronbach > for
this scale in our study was 0.87.

Mental Health
To assess mental health, we collected measures of depression, hopelessness,
and internalized HIV stigma. Two measures of depression were included in the
assessment battery, the nine-item Patient Health Questionnaire (PHQ-9 (33))
and the 15-item depression subscale of the Hopkins Symptom Checklist (DHSCL (34)). The PHQ-9 items correspond directly to the criteria included in
the DSM-IV (35) for diagnosing major depression; each item measures the
frequency of a symptom over the last 2 weeks using a rating scale from 0, not
at all, to 3, nearly everyday, and scores are summed (possible range for total
score is from 0 to 27). A total score greater than 9 indicates the presence of
clinical depression because this cutoff has been found to correspond highly to a
diagnosis of major depression (33); scores of 5 to 9 represent mild; 10 to 14,
moderate; 15 to 19, moderately severe; and 20 to 27, severedepression.
Unlike the PHQ-9, which provides a diagnostic type measure of a depressive
disorder, the D-HSCL is more of a global measure of depressive symptoms.
The D-HSCL elicits information about both cognitive (e.g., feeling blue, worry
too much about thing, thought of ending your life) and vegetative symptoms
(e.g., feeling low in energy, slowed down; poor appetite; difculty falling asleep
or staying asleep) of depression and uses a response format of 0, not at all, to
3, extremely, to reect the frequency of each symptom during the last week; a
mean item score is calculated, and normative data indicate that a mean of 0.75
or greater is reective of elevated depressive symptoms. Both the PHQ-9 (36)
and D-HSCL (8) have been used successfully with PLHA in sub-Saharan Africa.
Cronbach > values for these scales in our study were 0.77 for the PHQ-9 and 0.87
for the D-HSCL; the two scales were signicantly correlated (r = 0.73, p G .001).
Hopelessness was assessed with two items from the Beck Hopelessness
Scale (37): I look forward to the future with hope and enthusiasm and I
might as well give up because theres nothing I can do about making things
better for myself. Response options range from 1, strongly agree, to 4,
strongly disagree; scoring of the positively worded item is reversed, and then,
a mean item score is calculated. Internalized HIV stigma was assessed with
an eight-item scale developed by Kalichman et al. (38). Examples of items
include being HIV positive makes me feel damaged, I am ashamed that I
am HIV positive, and I hide my HIV status from others; response options
range from 1, disagree strongly, to 5, agree strongly, and a mean item score
is calculated. Cronbach > for the stigma scale in our study was 0.90.

Data Analysis
Pearson (for the continuous measures of CD4 cell count and physical health
functioning) and biserial (for the binary AIDS diagnosis variable) correlation
coefcients were computed to examine relationships between the four continuous mental health variables and measures of physical health (CD4 cell count,
AIDS diagnosis, and physical health functioning). To compare the ART and
non-ART groups with regard to each of the mental health measures, we rst used
bivariate statistics (two-tailed t tests and W2 tests) to compare baseline characteristics among the ART and non-ART groups and to examine (unadjusted)
within-group change over time (paired t test and McNemar test).
A staged approach to performing multivariate longitudinal regression models
was used to examine the effects of ART and HIV care on outcomes measured
across the three assessments and the potential explanatory role of change in
physical health functioning. In these models, we used the generalized estimating
equation method for analysis of correlated data to model the repeated measurements, assuming a normal distribution for the mean hopelessness and
stigma scale scores, and a binomial distribution for the dichotomous outcomes
of clinical depression and elevated depressive symptoms. As an initial step, we
analyzed models in which the dependent variable was change in the mental
health measure across the three study assessments, and the independent variables included ART status (representing whether there is a group difference
in the dependent variable at baseline), time (ordinal variable representing the
change in the dependent variable for each additional unit of time [i.e., 6 months]
over the three periods, and which is attributed to HIV medical care), and the
interaction of ART status by time (represents the additional change in the
dependent variable with each additional unit of time among patients in the ART
group relative to the group receiving HIV medical care alone or the non-ART
Psychosomatic Medicine 74:883Y890 (2012)

Copyright 2012 by the American Psychosomatic Society. Unauthorized reproduction of this article is prohibited.

ART IMPACT ON MENTAL HEALTH

TABLE 1. Baseline Sample Characteristics by ART Status
Total
Non-ART
Sample

Variable
n

602

302

ART
300

Demographics
Age, y

35.7

35.7

35.7

Men, %
At least some secondary
education, %

32
41

30
45

33
37

Working (past 7 d), %

61

71***

50***

Married or in a committed
relationship, %

46

47

44

Physical health
CD4 cell count

216

301***

130***

AIDS diagnosis, %

45

31***

60***

Physical health functioning

71.8

78.5***

65.0***

Mental health
Depression (D-HSCL)

0.85

0.71***

Depression (PHQ-9)

5.22

4.36***

1.00***
6.09***

Clinical depression (PHQ-9 99), %

13

9**

17**

Elevated depressive symptoms

(D-HSCL Q0.75), %

56

42***

70**

Internalized HIV stigma

2.42

2.31**

2.53**

Hopelessness

1.35

1.37

1.32

ART = antiretroviral therapy; AIDS = acquired immunodeciency syndrome;

D-HSCL = depression subscale of the Hopkins Symptom Checklist; PHQ-9 =
nine-item Patient Health Questionnaire; HIV = human immunodeciency virus.
All statistics are mean values, except for the percentages, which represent
proportions of the sample.
** p G .01, *** p G .001.
group). Covariates that were added to the model included patient characteristics
that controlled for differences (excluding the mental health variables) in the
ART and non-ART groups at baseline (CD4 cell, physical health functioning,
and work status), as well as age and sex. (Note that AIDS diagnosis was not
included as a covariate because of its overlap with CD4 cell count, given that
CD4 cell count is a large determinant of an AIDS diagnosis denition.) In the
second step, the same models were reexamined but with the addition of change
in physical health functioning from baseline in separate analyses. All analyses
included attrition weights to account for dropout; these were derived using
logistic regression with completion status as the outcome and baseline measures
associated with ART and completion status as independent variables.

Sensitivity Analyses
The primary analysis used an intention-to-treat (ITT) approach, which included all participants in the ART and non-ART groups at baseline, thus
resulting in a conservative estimate of the effects of ART given that some nonART patients (n = 50) would start ART during the study period. We augmented
the ITT analysis with two sensitivity analyses. In sensitivity Analysis 1, we excluded the 50 non-ART patients who started ART during the course of the study.
In sensitivity Analysis 2, the non-ART group was restricted to the most comparable 122 patients with signs of disease progression that would normally
signify ART eligibility (CD4 cell count G250 cells/mm3 or AIDS diagnosis),
although in these cases, ART had been deferred for other medical or psychosocial reasons (e.g., poor clinic attendance, active tuberculosis that was being
treated rst, patient refusal of ART). Both sensitivity analyses were performed
for each of the four mental health outcomes. In each case, the ndings remained
unchanged from the original models; therefore, we have chosen not to present
the data from the sensitivity analyses. For the models that included the change in
physical health functioning, the sensitivity analyses for hopelessness resulted
in one change, and this is noted in the Results section.

RESULTS
Sample Characteristics at Baseline
The sample consists of 602 participants, including 300 ART
and 302 non-ART patients: 302 participants were from Kampala
and 300 were from Kakira, with even distribution of ART and
non-ART patients at both sites. Baseline characteristics of the
total sample, and the ART and non-ART subgroups, are listed
in Table 1. Compared with the non-ART group, the ART patients
had worse physical health (lower CD4 cell count, greater proportion having an AIDS diagnosis, and lower physical health
functioning) and were less likely to be currently working.
Other demographic characteristics did not differ between the
two groups.
Attrition was very low, with 92% and 94% of the ART and
non-ART groups, respectively, completing Month 12. Compared
with completers, study dropouts were less likely to be working
(45% versus 62%, p G .05) at baseline; there were no other differences between completers and dropouts. Fifty (17%) nonART patients started ART before the Month 12 assessment.
Mental Health Characteristics at Baseline
The mean (M) (standard deviation) PHQ-9 score for the total
sample was 5.22 (3.93) (range, 0Y24); 290 (48%) had none
or minimal depressive symptoms, 235 (39%) had mild depression, and 78 (13%) had clinical depression (PHQ-9 99). Rates
of clinical depression did not differ between men (12%) and
women (13%). Among those with clinical depression, 64 (10.6%
of the total sample), 10 (1.7%), and 4 (0.7%) had moderate,
moderately severe, and severe depression, respectively. A higher
proportion of ART patients had clinical depression compared
with the non-ART group (17% versus 9%, p G .001). The most
common depressive symptoms (reported on the PHQ-9) present at least more than half the days during the past 2 weeks
were poor appetite (23%); feeling down, depressed, or hopeless
(19%); and feeling tired or having little energy (18%); having
any thoughts of being better off dead or of hurting yourself was
reported by 9%.
A much higher proportion of the sample (56%) had elevated depressive symptoms as measured by the D-HSCL
TABLE 2. Bivariate Relationships Between Measures of Mental
and Physical Health
Mental Health
Measures

CD4 Cell
Count

AIDS
Diagnosis

Physical Health
Function

Depression (PHQ-9)

j0.260***

0.275***

j0.536***

Depression
(D-HSCL)

j0.304***

0.278***

j0.530***

Hopelessness

j0.038

0.205***

j0.207***

Internalized HIV
stigma

j0.112

0.230***

j0.239***

AIDS = acquired immunodeciency syndrome; PHQ-9 = nine-item Patient

Health Questionnaire; D-HSCL = depression subscale of the Hopkins Symptom
Checklist; HIV = human immunodeciency virus.
*** p G .001.
p G .10.

Psychosomatic Medicine 74:883Y890 (2012)

Copyright 2012 by the American Psychosomatic Society. Unauthorized reproduction of this article is prohibited.

885

G. J. WAGNER et al.

Figure 1. Clinical depression and elevated depressive symptoms by antiretroviral therapy (ART) status.

(score Q0.75), including 70% of ART patients compared with

42% of non-ART patients ( p G .001). Elevated depressive
symptoms were also much higher among women (62%) compared with men (44%) ( p G .001). Mean levels of depression
(as measured by both the PHQ-9 and D-HSCL) and internalized HIV stigma were also greater in the ART group compared
with the non-ART group (Table 1); levels of hopelessness did
not differ between the two groups.
Table 2 lists the correlates between each of the four continuous mental health variables (PHQ-9 total score and mean
item scores for the D-HSCL, hopelessness, and stigma scales)
and the measures of physical health (CD4 cell count, AIDS
diagnosis, and physical health functioning). CD4 cell count
was negatively correlated with both measures of depression, as
well as internalized HIV stigma, but was not associated with
hopelessness, whereas having an AIDS diagnosis and physical
health functioning were positively correlated with each of the
mental health measures.
Change in Mental Health Over 12 Months of
HIV Treatment
Clinical Depression
As depicted in Figure 1, bivariate analysis shows that the
proportion of ART patients with clinical depression decreased
signicantly from baseline (15.0%) to Month 6 (5.1%, p G .001)
but remained statistically unchanged at Month 12 (3.6%) compared with Month 6. In the non-ART group, the proportion of
clinically depressed patients decreased from 8.1% at baseline
to 3.0% at Month 6 (p G .01) and then remained at an equivalently low 4.9% at Month 12. In the sample as a whole, 24
patients (4%) were clinically depressed at Month 12 (down
from 13% at baseline), which included 10 (15%) of the 69
patients who were depressed at baseline and completed the
study. The reduction in rates of clinical depression from baseline to Month 12 was marginal among men (5% was clinically
depressed at Month 12, p G .1) but statistically signicant
among women (4% was clinically depressed at Month 12,
p G .05). The multivariate ITT analysis of clinical depression
over the three study assessments revealed no difference between the ART and non-ART groups at baseline, nor was the
overall time trend signicant; however, the interaction of time
and ART status was signicant (p G .05), indicating that the
886

reduction in clinical depression in the ART group was significantly greater than that in the non-ART group (see Model 1
in Table 3).
Elevated Depressive Symptoms
The proportion of patients with elevated depressive symptoms dramatically declined from 69% at baseline to 18% at
Month 6 (p G .001) in the ART group and from 41% to 23%
in the non-ART group (p G .001); in both groups, there was
no signicant change from Months 6 to 12 (Fig. 1). The proportion of the total sample with elevated depressive symptoms
dropped from 56% (n = 338) at baseline to 18% of study
completers (n = 104) at Month 12, including 26% (82/319) of
study completers with elevated symptoms at baseline. Both men
and women experienced a signicant reduction in the rate of
elevated depressive symptoms (p G .001), with 13% and 21%
of men and women, respectively, having elevated symptoms at
Month 12. The multivariate ITT analysis of elevated depressive
symptoms revealed no difference between the ART and nonART groups at baseline; however, time (p G .001) and the interaction of time and ART status (p G .001) were signicant,
TABLE 3. Multivariate Analysis of Effects of ART on Change in
Depression Outcomes Over 12 Months of HIV Care
Clinical Depression
OR

95% CI

Elevated Depressive
Symptoms
OR

95% CI

Model 1
ART

1.03

0.51Y2.07

1.20

0.77Y1.88

Time
Time by ART

0.72
0.51*

0.50Y1.06
0.29Y0.87

0.53**
0.34**

0.43Y0.64
0.25Y0.48

ART

0.92

0.43Y1.96

1.04

0.62Y1.73

Time

1.00

0.61Y1.65

0.62**

0.50Y0.76

Time by ART

0.75

0.36Y1.57

0.48**

0.34Y0.68

$Physical functioning

0.96**

0.94Y0.98

0.96**

0.96Y0.97

Model 2

ART = antiretroviral therapy; HIV = human immunodeciency virus; OR =

odds ratio; CI = condent interval.
All models included the following baseline covariates: age, sex, CD4 cell count,
physical functioning, and work status.
* p G .05, ** p G .001.
Psychosomatic Medicine 74:883Y890 (2012)

Copyright 2012 by the American Psychosomatic Society. Unauthorized reproduction of this article is prohibited.

ART IMPACT ON MENTAL HEALTH

indicating that rates of elevated depressive symptoms reduced
signicantly over the study period in the sample overall but
even more so in the ART group (see Model 1 in Table 3).
Hopelessness
In the ART group, mean levels of hopelessness decreased
from baseline (1.33) to Month 6 (1.21, p G .001) and further
decreased at Month 12 (1.10, p G .001); in the non-ART group,
hopelessness also decreased from 1.37 at baseline to 1.14 at
Month 6 (p G .001), but there was no further change at Month 12
(M = 1.15) (Fig. 2). In the multivariate ITT analysis, the ART
group had signicantly higher hopelessness at baseline (p G .01),
and levels of hopelessness decreased signicantly over time in the
sample as a whole (time was a signicant predictor; p G .001),
with no difference in this reduction between ART and non-ART
groups (see Model 1 in Table 4).
Internalized HIV Stigma
Levels of internalized HIV stigma decreased signicantly
in the ART group from a mean of 2.57 at baseline to 1.92 at
Month 6 (p G .001), but then, there was no further change
at Month 12 (M = 1.85); in the non-ART group, stigma also
reduced from baseline (2.31) to Month 6 (1.85, p G .001) but
then returned to baseline levels at Month 12 (2.23, p G .001)
(Fig. 3). In the multivariate ITT analysis of stigma, the ART
and non-ART groups did not differ at baseline, and the overall time trend was only marginally signicant ( p G .10), but the
interaction of time and ART was signicant ( p G .001); this
nding indicates that the reduction in stigma over time was
signicantly greater in the ART group compared with the nonART group, with the latter group having no signicant change in
stigma over the 12 months (see Model 1 in Table 4).
The Role of Physical Health Functioning in the
Effects of ART on Mental Health
To examine whether physical health functioning helps to
explain the effects of HIV care and ART on the mental health
outcomes, we rst assessed whether physical health functioning

Figure 2. Hopelessness by antiretroviral therapy (ART) status over 12 months

of human immunodeciency virus care.

TABLE 4. Multivariate Analysis of Effects of ART on Change in

Hopelessness and Internalized HIV Stigma Over 12 Months
of HIV Care
Hopelessness
B

SE

Stigma
B

Model 1
ART

j0.12**

0.04

0.07

Time

j0.12***

0.02

j0.05

0.02

j0.30***

Time by ART

0.00

SE

0.08
j0.03
0.04

Model 2
ART

j0.14***

0.04

0.04

0.08

Time

j0.09***

0.02

j0.01

0.03

0.05*

0.02

j0.21***

Time by ART

$Physical functioning j0.005*** 0.0005 j0.008***

0.04
0.001

ART = antiretroviral therapy; HIV = human immunodeciency virus; B =

adjusted A; SE = standard error.
All models included the following baseline covariates: age, sex, CD4 cell count,
physical functioning, and work status.
* p G .05, ** p G .01, *** p G .001.
p G .10.

improved signicantly over the course of the study and whether

such improvement was associated with ART status. Note that,
as reported previously, the ART group had lower physical health
functioning compared with the non-ART group at baseline.
In the ART group, physical health functioning improved signicantly from baseline (M = 66.6) to Month 6 (M = 85.7,
p G .001), and then, even further improvement was observed
at Month 12 (M = 92.5, p G .001). The non-ART group experienced a smaller increase in physical health functioning from
baseline (M = 78.8) to Month 6 (M = 87.7, p G .001), and then
no further change was observed at Month 12 (M = 86.9).
When change in physical health functioning was added to
the previously described multivariate models for the four mental
health outcomes, change in physical health functioning was a
strong predictor (p values G .001) in each model, with improved
physical health functioning being associated with reduced levels
of both depression measures, hopelessness, and internalized HIV
stigma (see Model 2 in Tables 3 and 4). In the two depression
models, the addition of change in physical health functioning

Figure 3. Internalized human immunodeciency virus (HIV) stigma by ART

status over 12 months of HIV care. ART = antiretroviral therapy.

Psychosomatic Medicine 74:883Y890 (2012)

Copyright 2012 by the American Psychosomatic Society. Unauthorized reproduction of this article is prohibited.

887

G. J. WAGNER et al.
to the model did not alter the predictive value of time or the
interaction of time and ART, with the exception of the model
predicting clinical depression; the time-by-ART interaction was
no longer a signicant predictor of change in clinical depression, suggesting that restored physical health functioning is a
key driver for alleviating clinical depression, more so than receipt of ART. These ndings were replicated in both sensitivity
analyses (data not shown).
In the model predicting hopelessness, the interaction of time
and ART became signicant (p G .05) once change in physical
health functioning was added, suggesting a benet of ART over
and above HIV care in reducing hopelessness when physical
health functioning is controlled for. Last, in the model for internalized HIV stigma, the addition of change in physical health
functioning did not change the other primary parameters, with
the interaction of time by ART remaining signicant (p G .001).
In the sensitivity analyses, the only nding that was different
was with regard to hopelessness in the models where non-ART
patients who switched to being on ART were removed from the
analysis; unlike the other analyses, the interaction of time and
ART was now signicantly associated (p G .05) with reduced
hopelessness over time when change in physical health functioning was controlled for (data not shown).
DISCUSSION
The ndings of this study strongly support a benecial effect of HIV medical care, and ART in particular, on the mental
health of HIV clients during the initial year of treatment. The
sample as a whole, including both ART and non-ART patients,
experienced signicant reductions on measures of depression
and hopelessness in multivariate analysis. However, although
HIV medical care demonstrated benecial effects on mental
health even in the absence of ART, the reduction in depression
was greater among ART patients compared with non-ART
patients. Benets of ART on reduced internalized HIV stigma
were also evident.
The rate of clinical depression in the study sample at baseline was 13%, whereas a much higher proportion of the sample
(56%) had elevated depressive symptoms, both of which are
generally consistent with other studies of PLHA in sub-Saharan
Africa (3Y9). Most patients who were clinically depressed or had
elevated depressive symptoms at entry into HIV care were no
longer depressed after 12 months of treatment, although roughly
one quarter of the patients who struggled with depression at
baseline continued to do so at Month 12. The bivariate analysis, as depicted in Figures 1 to 3, highlights how most of the
mental health benets of HIV care and ART occurred during
the rst 6 months of treatment, with only small, if any, further improvement reported by participants after Month 6. These
ndings suggest that the onset of HIV medical care and ART,
and associated stabilization of physical health and improved
functional status, may be sufcient for helping to lift the depression and improve the mental health of many patients who
are distressed at entry into care; however, depression continues
to persist for some patients, for whom mental health treat888

ment is needed. Unfortunately, like most ART programs in

sub-Saharan Africa (4), mental health treatment is not available, and antidepressants are not used to treat depression at the
study sites.
Physical health functioning was shown to have strong predictive value for positive longitudinal change on all mental
health outcomes examined in the studys multivariate analyses.
This consistent nding highlights the role of improved physical
health and functioning that accompanies HIV medical care,
regardless of the inclusion of ART, and it helps to explain the
signicant (albeit smaller) improvement in mental health observed in the non-ART group, as well as the ART patients. It
is not surprising that restored physical health and functioning
translate to improved health in other aspects of a patients quality
of life, including mental health. Improved physical functioning
better equips people to work and provide and care for their
themselves and their family, which undoubtedly inuences
mental outlook and well-being, because other studies conducted
in Uganda have found that depression among PLHA is often
associated with socioeconomic stressors (11). Improved physical health functioning contributes to improved mental health,
but better mental health and alleviated depression can also result in improved physical functioning via heightened motivation and energy level, so the relationship between physical and
mental health is likely bidirectional.
Improved physical health functioning may be the primary
mechanism by which HIV care and ART inuence mental
health. Yet, the fact that ART patients experienced greater improvement on most mental health outcomes compared with
non-ART patients, even after change in physical health functioning was controlled for, suggests that there are other mechanisms by which ART affects mental health. ART patients are
required to attend clinic more frequently than non-ART patients
(often monthly compared with every 6 months), which exposes
these patients to more opportunities for receiving support from
peers and providers and forms of counseling related to ART
management and general psychosocial support, which may also
serve to improve psychological well-being.
Internalized HIV stigma is an indicator of an individuals
psychological adjustment to living with HIV, and it has been
found to be associated with depression and general mental
health (30), as well as maladaptive health behaviors including
increased sexual risk behavior (39) and lower HIV medical adherence (40) among PLHA. Levels of stigma were not associated with time or receipt of HIV medical care, in part, because
stigma returned to baseline levels at Month 12 in the non-ART
group, after an initial reduction at Month 6. In contrast, the signicant reduction in stigma achieved by the ART group during
the rst 6 months of treatment was maintained at Month 12. This
is consistent with other studies that have reported a reduction in
stigma associated with ART (41,42) and the potential for improved health and functioning to enable patients to feel better
about themselves and their ability to cope with HIV. Conversely,
some studies suggest that stigma from others can increase toward ART patients who are functioning well because these
patients are viewed as functional vectors of disease transmission
Psychosomatic Medicine 74:883Y890 (2012)

Copyright 2012 by the American Psychosomatic Society. Unauthorized reproduction of this article is prohibited.

ART IMPACT ON MENTAL HEALTH

(43), but our data do not support the possibility that such
stigma, if evident, is internalized by patients.
Study limitations include the inability to randomly assign
ART to matching groups. With widespread access to ART in
Uganda at the time of study enrollment, ART randomization
was not ethical. Although the non-ART group was relatively
comparable (many had CD4 cell G250 cells/mm3 or an AIDS
diagnosis), it had clear indicators of better physical and mental health at baseline. Yet, our sensitivity analyses resulted in
equivalent ndings to that of the main ITT analysis, indicating
that our results are robust. In addition, although the HIV care
comparison group enabled us to control for time trends in the
context of receipt of HIV care, the lack of a comparison group
of PLHA not in HIV care prevented us from being able to account for natural changes in the outcomes that may occur over
time in the absence of HIV medical care and to isolate the effect
of HIV care. Furthermore, our evaluation of clinical depression
would have been strengthened with the use of a structured diagnostic clinical interview, rather than reliance on subjective
self-reports; however, the PHQ-9 cutoff of scores greater than
9 has been shown to have high correspondence with major
depression diagnosed by the Mini International Neuropsychiatric Interview (44), which is widely used in sub-Saharan Africa.
In conclusion, this study demonstrates the benets of HIV
care and ART on mental health, including depression. Although
both ART and non-ART patients experienced mental health
benets, improvement was even greater among those on ART.
Most of the mental health benets were experienced in the rst
6 months of care, coinciding with the timing of most of the
increase in physical health functioning. However, for those who
continue to struggle with depression or other mental health
problems after their physical health has been stabilized, mental
health support and treatment are needed. With the harmful effects of depression on ART adherence (14,15), clinical outcomes
(16Y18), and sexual risk behavior (22,23), these ndings argues for the importance of integrating mental health services
into HIV care programs.

8.
9.
10.
11.
12.
13.

14.

15.

16.

17.
18.
19.

20.

21.

REFERENCES
1. Rabkin J. Prevalence of psychiatric disorders in HIV illness. Int Rev
Psychiatry 1996;8:157Y66.
2. Bing EG, Burnman MA, Longshore D, Fleishman JA, Sherbourne CD,
London AS, Turner BJ, Eggan F, Beckman R, Vitiello B, Morton SC,
Orlando M, Bozzette SA, Ortiz-Barron L, Shapiro M. Psychiatric disorders
and drug use among HIV-infected adults in the U.S. Arch Gen Psychiatry
2001;58:721Y8.
3. Brandt R. The mental health of people living with HIV/AIDS in Africa:
a systematic review. Afr J AIDS Res 2009;8:123Y33.
4. Collins PY, Homan AR, Freeman MC, Patel V. What is the relevance of
mental health to HIV/AIDS care and treatment programs in developing
countries? A systematic review. AIDS 2006;20:1571Y82.
5. Myer L, Smit J, Le Roux L, Parker S, Stein DJ, Seedat S. Common mental
disorders among HIV-infected individuals in South Africa: prevalence,
predictors, and validation of brief psychiatric rating scales. AIDS Patient
Care STDS 2008;22:147Y57.
6. Olleya BO, Seedat S, Stein DJ. Persistence of psychiatric disorders in a
cohort of HIV/AIDS patients in South Africa: a 6-month follow-up study.
J Psychosom Res 2006;61:479Y84.
7. Sebit MB, Tombe M, Siziya S, Balus S, Nkomo SDA, Maramba P.
Prevalence of HIV/AIDS and psychiatric disorders and their related risk

22.

23.

24.

25.

26.
27.

factors among adults in Epworth, Zimbabwe. East Afr Med J 2003;80:

503Y12.
Bolton P, Wilk CM, Ndogoni L. Assessment of depression prevalence
in rural Uganda using symptom and function criteria. Soc Psychiatry
Psychiatr Epidemiol 2004;39:442Y7.
Kinyanda E. Frequency with which psychiatric disorder is associated with
positive HIV-1 serostatus as seen in persons attending a TASO clinic in
Mulago Hospital, Kampala, Uganda. S Afr Med J 1998;88:1178.
Wilk CM, Bolton P. Local perceptions of the mental health effects of the
Uganda acquired immunodeficiency syndrome epidemic. J Nerv Ment Dis
2002;190:394Y7.
Okello E, Neema S. Explanatory models and help seeking behavior: pathways
to psychiatric care among patients admitted for depression in Mulago Hospital, Kampala, Uganda. Qual Health Res 2007;17:14Y25.
Briongos-Figuero LS, Bachiller-Luque P, Palacios-Martin T, de Luis-Roman
D, Eiros-Bouza JM. Depression and health related quality of life among
HIV-infected people. Eur Rev Med Pharmacol Sci 2011;15:855Y62.
Patel R, Kassaye S, Gore-Felton C, Wyshak G, Kadzirange G, Woelk G,
Katzenstein D. Quality of life, psychosocial health, and antiretroviral
therapy among HIV-positive women in Zimbabwe. AIDS Care 2009;21:
1517Y27.
Starace F, Ammassari A, Trotta MP, Murri R, De Longis P, Izzo C, Scalzini
A, DArminio Monforte A, Wu AW, Antinori A; AdlCoNA Study Group,
NeuroICoNA Study Group. Depression is a risk factor for suboptimal
adherence to antiretroviral therapy. J Acquir Immune Defic Syndr 2002;
31:S136Y9.
Ammassari A, Antinori A, Aloisi MS, Trotta MP, Murri, R, Bartoli L,
DArminio Monforte A, Wu AW, Starace F. Depressive symptoms, neurocognitive impairment, and adherence to highly active antiretroviral therapy
among HIV-infected persons. Psychosomatics 2004;45:394Y402.
Cook JA, Grey D, Burke J, Cohen MH, Gurtman AC, Richardson JL,
Wilson TE, Young MA, Hessol NA. Depressive symptoms and AIDS-related
mortality among a multisite cohort of HIV-positive women. Am J Public
Health 2004;94:1133Y40.
Pence BW, Miller WC, Gaynes BN, Eron JJ. Psychiatric illness and virologic response in patients initiating highly active antiretroviral therapy.
J Acquir Immune Defic Syndr 2007;44:159Y66.
Paterson DL, Swindells S, Mohr J, Brester M, Vergis EN, Squier C,
Wagener MM, Singh N. Adherence to protease inhibitor therapy and outcomes in patients with HIV infection. Ann Intern Med 2000;133:21Y30.
Ickovics JR, Hamburger ME, Vlhaov D, Schoenbaum EE, Schuman P,
Boland RJ, Moore J. Mortality, CD4 cell count decline, and depressive
symptoms among HIV-seropositive women: longitudinal analysis from the
HIV Epidemiology Research Study. JAMA 2001;285:1466Y74.
Evans DL, Ten Have TR, Douglas SD, Gettes DR, Morrison M, Chiappini
MS, Brinker-Spence P, Job C, Mercer DE, Wang YL, Cruess D, Dube B,
Dalen EA, Brown T, Bauer R, Petitto JM. Association of depression with
viral load, CD8 T lymphocytes, and natural killer cells in women with HIV
infection. Am J Psychiatry 2002;159:1752Y9.
Kalichman SC, Difonzo K, Austin J, Luke W, Rompa D. Prospective study
of emotional reactions to changes in HIV viral load. AIDS Patient Care
STDS 2002;16:113Y20.
Kennedy CA, Skurnick J, Wan JY, Quattrone G, Sheffet A, Quinones M,
Wang W, Louria D. Psychological distress, drug and alcohol use as correlates of condom use in HIV-serodiscordant heterosexual couples. AIDS
1993;7:1493Y9.
Wagner GJ, Ghosh-Dastidar B, Holloway I, Kityo C, Mugyenyi P. Depression in the pathway of HIV antiretroviral effects on sexual risk behavior
among patients in Uganda. AIDS Behav October 11, 2011. Epub ahead of
print, doi: 10.1007/s10461-011-0051-7.
Preau M, Bonnet A, Bouhnik AD, Fernandez L, Obadia Y, Spire B, VESPA.
Anhedonia and depressive symptomatology among HIV-infected patients
with highly active antiretroviral therapies (ANRS-EN12-VESPA). Encephale
2008;34:385Y93.
Wagner GJ, Holloway I, Ghosh-Dastidar B, Kityo C, Mugyenyi P. Understanding the influence of depression on self-efficacy, work status and
condom use among HIV clients in Uganda. J Psychosom Res 2011;70:
440Y8.
Patel V, Araya R, Bolton P. Treating depression in the developing world.
Trop Med Int Health 2004;9:539Y41.
Mast TC, Kigozi G, Wabwire-Mangen F, Black R, Sewankambo N,
Swerwadda D, Gray R, Wawer M, Wu AW. Measuring quality of life
among HIV-infected women using a culturally adapted questionnaire in
Rakai district, Uganda. AIDS Care 2004;16:81Y94.

Psychosomatic Medicine 74:883Y890 (2012)

Copyright 2012 by the American Psychosomatic Society. Unauthorized reproduction of this article is prohibited.

889

G. J. WAGNER et al.
28. Stangl AL, Wamai N, Mermin J, Awor AC, Bunnell RE. Trends and predictors of quality of life among HIV-infected adults taking highly active
antiretroviral therapy in rural Uganda. AIDS Care 2007;19:626Y36.
29. Jelsma J, Maclean E, Hughes J, Tinise X, Darder M. An investigation into
the health-related quality of life of individuals living with HIV who are
receiving HAART. AIDS Care 2005;17:579Y88.
30. Pearson CR, Micek MA, Pfeiffer J, Montoya P, Matediane E, Jonasse T,
Cunguara A, Rao D, Gloyd SS. One year after ART initiation: psychosocial
factors associated with stigma among HIV-positive Mozambicans. AIDS
Behav 2009;13:1189Y96.
31. Berg CJ, Michelson SE, Safren SA. Behavioral aspects of HIV care: adherence, depression, substance use, and HIV-transmission behaviors. Infect
Dis Clin North Am 2007;21:181Y200.
32. Wu AW, Revicki DA, Jacobson D, Malitz FE. Evidence for reliability,
validity and usefulness of the Medical Outcomes Study HIV Health Survey
(MOS-HIV). Qual Life Res 1997;6:481Y93.
33. Kroenke K, Spitzer RL, Williams JB. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med 2001;16:606Y13.
34. Derogatis LR, Lipman RS, Rickels K, Uhlenhuth EH, Covi L. The Hopkins
Symptom Checklist (HSCL): a self-report symptom inventory. Behav Sci
1974;19:115.
35. American Psychiatric Association. Diagnostic Statistical Manual of Mental Disorders. 4th ed. Washington, DC: American Psychiatric Association;
1994.
36. Monahan PO, Shacham E, Reece M, Kroenke K, Ongor WO, Omollo O,
Yebel VN, Ojwang C. Validity/reliability of PHQ-9 and PHQ-2 depres-

890

37.
38.
39.
40.

41.
42.

43.
44.

sion scales among adults living with HIV/AIDS in western Kenya. J Gen
Intern Med 2009;24:189Y97.
Beck AT. Beck Hopelessness Scale. San Antonio, Texas: Psychological
Corp; 1988.
Kalichman SC, Simbayi LC, Jooste S, Toefy Y, Cain D, Cherry C, Kagee A.
Development of a brief scale to measure AIDS-related stigma in South
Africa. AIDS Behav 2005;9:135Y43.
Clum G, Chung S, Ellen JM. Mediators of HIV-related stigma and risk
behavior in HIV infected young women. AIDS Care 2009;21:1455Y62.
Nam SL, Fielding K, Avalos A, Dickinson D, Gaolathe T, Geissler PW.
The relationship of acceptance or denial of HIV status to antiretroviral
adherence among adult HIV patients in urban Botswana. Soc Sci Med
2008;67:301Y10.
Abadia-Barrero CE, Castro A. Experiences of stigma and access to HAART
in children and adolescents living with HIV/AIDS in Brazil. Soc Sci Med
2006;62:1219Y28.
Makoae LN, Portillo CJ, Uys LR, Dlamini PS, Greeff M, Chirwa M,
Kohi TW, Naidoo J, Mullan J, Wantland D, Durrheim K, Holzemer WL.
The impact of taking or not taking ARVs on HIV stigma as reported
by persons living with HIV infection in five African countries. AIDS Care
2009;21:1357Y62.
Roura M, Urassa M, Busza J, Mbata D, Wringe A, Zaba B. Scaling up
stigma? The effects of antiretroviral roll-out on stigma and HIV testing.
Early evidence from rural Tanzania. Sex Transm Infect 2009;85:308Y12.
Sheehan D, Lecrubier Y, Sheehan K, Janavs J, Weiller E, Keskiner A. The
validity of the Mini International Neuropsychiatric Interview (MINI)
according to the SCID-P and its reliability. Eur Psychiatry 1997;12:232Y41.

Psychosomatic Medicine 74:883Y890 (2012)

Copyright 2012 by the American Psychosomatic Society. Unauthorized reproduction of this article is prohibited.