Wk6 Assignment 1 Discussion

732

AIDS PATIENT CARE and STDs Volume 21, Number 10, 2007 © Mary Ann Liebert, Inc. DOI: 10.1089/apc.2007.0012

Efficacy of Group Psychotherapy to Reduce Depressive Symptoms among HIV-Infected Individuals:

A Systematic Review and Meta-Analysis

SETH HIMELHOCH, M.D., M.P.H., DEBORAH R. MEDOFF, Ph.D., and GLORIA OYENIYI, B.A.

ABSTRACT

Depressed mood is highly prevalent among HIV-infected individuals. Some but not all stud- ies have found group psychotherapy to be efficacious in this population. We performed a sys- tematic review and meta-analysis of double-blinded, randomized controlled trials to exam- ine efficacy of group psychotherapy treatment among HIV infected with depressive symptoms. We used PubMed, the Cochrane database, and a search of bibliographies to find controlled clinical trials with random assignment to group psychotherapy or control condi- tion among HIV infected patients with depressive symptoms. The principal measure of ef- fect size was the standard difference between means on validated depression inventories. We identified 8 studies that included 665 subjects: 5 used cognitive behavioral therapy (CBT), 2 used supportive therapy, and 1 used coping effectiveness training. Three of the 8 studies re- ported significant effects. The pooled effect size from the random effects model was 0.38 (95% confidence interval [CI]: 0.23–0.53) representing a moderate effect. Heterogeneity of effect was not found to be significant (p � 0.69; I2 � 0%). Studies reporting use of group CBT had a pooled effect size from the random effects model of 0.37 (95% CI: 0.18–0.56) and was signif- icant. Studies reporting the use of group supportive psychotherapy had a pooled effect size from the random effects model 0.58 (95% CI: �0.05–1.22) and was nonsignificant. The results of this study suggest that group psychotherapy is efficacious in reducing depressive symp- toms among, HIV-infected individuals. Of note, women were nearly absent from all studies. Future studies should be directed at addressing this disparity.

INTRODUCTION

DEPRESSED MOOD is highly prevalent amongindividuals receiving medical care for HIV.1 Individuals with HIV and depressive disorders, compared to those with HIV alone, have increased HIV related morbidity,2,3 and among women a higher mortality.4,5 Although highly active antiretroviral therapy (HAART) has led to substantial reductions in morbidity

and mortality associated with HIV, studies have shown that individuals with HIV and de- pressive disorders are more likely to encounter greater delays in being prescribed antiretro- viral therapy,6 and have worse adherence to taking antiretroviral medication.7 This is in keeping with research that has shown that de- pression itself is associated with poor adher- ence to medical treatment.8

Recent studies, however, suggest that men-

Department of Psychiatry, University of Maryland School of Medicine, Baltimore, Maryland.

 

 

GROUP PSYCHOTHERAPY TO REDUCE DEPRESSIVE SYMPTOMS 733

tal health interventions may lead to improved depressive and HIV-related outcomes.9,10 A recent systematic review and meta-analysis found that antidepressants are efficacious tar- geting depression among those with HIV.11 However, antidepressant treatment may be as- sociated with high dropout rates11 and may not be acceptable to all patients. Wk6 Assignment 1 Discussion

Psychotherapeutic interventions have also been used to alleviate psychosocial and inter- personal difficulties and distress associated with HIV. Several randomized control trial studies have investigated the efficacy of group therapy techniques to decrease psychological distress, decrease social isolation, and improve coping among HIV-infected people.12–19 Most of these studies used interventions based on cognitive behavioral theory and nearly all these studies were conducted among men. Because some, but not all, studies have found group therapy interventions to be efficacious in de- creasing distress among HIV-infected people, we undertook a meta-analysis of randomized controlled trials to examine whether depressive symptoms respond to group psychotherapy treatment among HIV-infected people.

MATERIALS AND METHODS

Search strategy and study inclusion criteria

Because the term AIDS was introduced in 1981 we searched MEDLINE, PSYCHINFO, and Cochrane databases from 1981–2006 using the key words: psychotherapy and adaptation, psychological with HIV or AIDS and limited to randomized control trials. In an effort to locate both published and unpublished studies the bibliographies of key reviews were examined. Studies were included if they met the follow- ing criteria: (1) prospective, double-blinded, controlled trials with random assignment; (2) report of outcomes of depressive symptoms; (3) report of use of a psychotherapeutic inter- ventions. The three authors independently screened the titles and abstracts of each citation. Wk6 Assignment 1 Discussion

Data extraction

Data were independently extracted from the studies by the three authors. Discrepancies

were resolved by formal review and then by consensus. Our outcome of interest was de- pressive symptoms. Depression inventories that were specific for depressive symptoms were abstracted. These inventories included the Hamilton Depression Inventory (Ham-D), Center for Epidemiogic Studies-Depression (CES-D), and Beck Depression Inventory (BDI).

The standardized difference in means (Co- hen d), the effect size, was calculated from means and standard deviations from these scales. When data on means or standard devi- ations were lacking we contacted the authors of the manuscripts. The one author contacted did not respond to our inquiry for requested information. We also compiled information re- garding demographics, study characteristics, and type of psychotherapy intervention re- ported. Wk6 Assignment 1 Discussion

Quality of clinical trials

As variation in quality of clinical trials can result in biased estimates of reported interven- tion effectiveness, we evaluated the quality of the clinical trials using a 15-item scale devel- oped by Detsky et al.20 Each author indepen- dently rated the quality of the clinical studies. Discrepancies were resolved by formal review and then by consensus.

Statistical analysis

We calculated effect sizes and pooled esti- mates of effect across studies (Stat 8.0: metan command) using analysis of variance models for standardized mean differences (Cohen d). A random effects model was used. We chose to use a random effects model because it takes into account both within and between-study variation leading to a more conservative weighting estimates. Heterogeneity, or the be- tween study variation in outcomes, was mea- sured using the Q statistic.21 The Q statistic is considered to have a low power as a test of het- erogeneity; therefore, heterogeneity was con- sidered present with a p � 0.10. If heterogene- ity was found to be present the I2 statistic was used to describe the percentage of variation due to heterogeneity across studies. In the ab- sence of heterogeneity (i.e., Q statistic, p � 0.10), pooled results were reported. Publication

 

 

bias was evaluated using a funnel plot as well as Eggers and Beggs tests.21

RESULTS

Search findings

We identified 18 randomized clinical tri- als.12–19,22–31 Of these, 8 trials12–19 met inclusion criteria (Fig. 1). These 8 trials included 665 pa- tients randomly assigned to psychotherapy or a parallel control arm (Table 1). Depression was required at baseline for only one study14 and two studies excluded those with major depres- sion.15,17 With respect to the type of psycho- therapeutic treatment all of the studies used a group format. One study had two intervention arms—a CBT group intervention and a sup- portive therapy group intervention.14 Five of the treatment interventions were described as cognitive behavioral therapy (CBT),12–16 one was described as coping effectiveness training (CET),17 and two were described as supportive psychotherapy.14,18 Finally one study reported

results that combined two treatment arms (emotional expressive and CBT therapy) to- gether.19 Length of treatment ranged between 7–15 sessions. The length of the intervention ranged between 90 and 150 minutes. All inter- ventions were directed at improving psycho- logical distress and improving mood. Two interventions were also directed at reducing grief.16,18 Six trials occurred in the United States, one trial occurred in Amsterdam19 and one occurred in Hong Kong.13 With respect to demographics all but one16 study was con- ducted on men (Table 1). All studies were rated as reflecting good quality.

Depressive symptom outcome

Three of the 8 studies reported significant ef- fects. Of the 3 studies that found significant ef- fects, one used cognitive behavioral treatment intervention,16 one used supportive psycho- therapy,14 and one reported the results of a combination of emotional expressive and CBT therapy.19 The pooled effect size from the ran- dom effects model was 0.38 (95% CI: 0.23–0.53;

HIMELHOCH ET AL.734

FIG. 1. Flow diagram of randomized control trials included and excluded in meta-analysis.

 

 

GROUP PSYCHOTHERAPY TO REDUCE DEPRESSIVE SYMPTOMS 735

Fig. 2) representing a small-moderate effect size. Heterogeneity of effect was not found to be significant (p � 0.69; I2 � 0% of variability in effect sizes due to heterogeneity). Wk6 Assignment 1 Discussion

We were interested in investigating whether intervention type (i.e., CBT versus non-CBT group therapy interventions) moderated the ef- fect between psychotherapy and depressive symptoms. Studies reporting use of group CBT had a pooled effect size from the random ef- fects model of 0.37 (95% CI: 0.18–0.56]) and was significant representing a moderate effect size. Studies reporting the use of group supportive psychotherapy had a pooled effect size from the random effects model 0.58 (95% CI: �0.05–1.22]) and was nonsignificant. In the one study that used CET, the effect size from the random effects model was 0.16 (95% CI: �0.27–0.59]) and was not significant.

We were also interested in investigating whether the focus of treatment (i.e., grief and depressive symptoms versus depressive symp- toms) moderated the effect between psy- chotherapy and depressive symptoms. Studies focusing on grief and depressive symptoms had a pooled effect size from the random ef- fects model of 0.34 (95% CI: 0.12–0.56]) and was significant, representing a small to moderate ef- fect size. Studies focusing on depressive symp-

toms had a pooled effect size from the random effects model 0.42 (95% CI: 0.21–0.63) and was significant representing a moderate effect size.

Finally we were interested in investigating whether the exclusion of depression moderated the effect between psychotherapy and depres- sive symptoms. The two studies that excluded participants with major depression were found to have a pooled effect size from the random effects model of 0.26 (95% CI: �0.10–0.61) and was not significant. In contrast, those studies that included participants with major depres- sion had a pooled effect size from the random effects model of 0.41 (95% CI: 0.24–0.48) and was significant, representing a moderate effect size. Wk6 Assignment 1 Discussion

Assessment of publication bias

The funnel plot was roughly symmetric. Eg- ger’s test and Begg’s test were both nonsignif- icant. Taken together these findings suggest the relative absence of publication bias.

DISCUSSION

Our meta-analysis of randomized double- blinded controlled trials of group psychother-

TABLE 1. CHARACTERISTICS OF THE GROUP THERAPY STUDIES

Baseline Number Group Depression Number Age Male Caucasian depression group meetings: outcome Type of control

Study randomized (mean) (%) (%) required meetings min/wk measurea group

Goodkin 97 36.5 100 52.6 No 10 90 Hamilton Usual care Sikkema 235 40.3 64 28.0 No 12 90 Hamilton Usual care Kellyb 68 34.0 100 62.0 Yes 8 90 CES-D Usual care Chanc 13 38.1 100 — No 7 120 CES-D Wait list Chesney 84 39.0 100 82.0 Noe 10 90 CES-D HIV info/wait list Mulderc,d 27 40.4 100 — No 15 150 BDI Wait list Lutgendorf 40 36.7 100 62.5 Noe 10 135 BDI Wait list Antoni 101 41.6 100 52.0 No 10 135 BDI Med adherence

aThe Ham-D is a 17-item scale clinician-rated depression scale with a response range from 0–54. The CES-D is a 20-item subject-rated depression scale with a response range from 0–60. The BDI is a 21-item subject-rated depres- sion scale with a response range from 5–63.

bThis study had a CBT arm and a supportive therapy arm. cThe Chan study was from Hong Kong and did not report on race. The Mulder sample was from Amsterdam and

did not report on race. dThe Mulder study had a CBT and an emotional expressive therapy arm. However, the intervention results were

presented as a combination of both CBT and emotional expressive therapy. eThe Chesney study excluded participants with major depression. The Lutgendorf study excluded participants with

Hamilton Depression Rating Scale for Depression in the “moderate or greater severity level.” CBT, cognitive behavioral therapy. Wk6 Assignment 1 Discussion

 

 

apy targeting depressive symptoms among HIV-infected individuals found that group psychotherapy is efficacious. The combined ef- fect size was 0.38 (95% CI: 0.23–0.53) repre- senting a small to moderate effect size. We did not find any heterogeneity among the studies and there did not appear to be publication bias. A meta-analysis of group psychotherapy for unipolar depression found that among 15 stud- ies in which participants in the group psy- chotherapy intervention were compared to un- treated controls the pooled effect size was 1.03.32 The greater effect size found in the meta- analysis among those treated for unipolar de- pression may not be surprising. Those with unipolar depression, in contrast to those with depressive symptoms, are, on average more likely to have a greater burden of depressive symptoms and therefore have a greater proba- bility of depressive symptom reduction which would be reflected in a larger effect size.

In our meta-analysis, most studies used a cognitive behavior group therapy intervention to target depressive symptoms. The combined effect size for cognitive behavior was 0.37 (95% CI: 0.18–0.56) representing a moderate effect size. Thus, cognitive behavioral therapy ap-

pears to be efficacious in targeting depressive symptoms among HIV-infected individuals.

Less can be said about the other forms of therapy used. For example, although support- ive therapy seems to have a positive effect on reducing distress and depression among HIV- infected individuals, the limited number of studies and the large variability in the results of these studies makes it difficult to draw a clear conclusion. Whether the focus of the in- tervention was on grief and depressive symp- toms or depressive symptoms alone, did not appear to moderate the effect of the interven- tion with respect to depressive symptoms. Wk6 Assignment 1 Discussion

Finally, the pooled results of the studies that included participants with major depression appeared to have a significant effect while those that excluded participants with major de- pression did not. As those with major depres- sion, on average, are likely to have a greater probability of depressive symptom reduction than those without major depression, the dif- ference we found may in fact reflect a floor ef- fect.

Although the theoretical underpinnings of the group therapy interventions included in the meta-analysis were diverse they did share sev-

HIMELHOCH ET AL.736

FIG. 2. Forrest plot: Effect of group psychotherapy on depressive symptom outcome stratified by type of group intervention.

 

 

GROUP PSYCHOTHERAPY TO REDUCE DEPRESSIVE SYMPTOMS 737

eral features in common. First, all used a group therapy format. Second, all sessions were at least 90 minutes and occurred on average for 10 sessions. Third, each study used techniques specifically tailored to improve coping strate- gies and improve social support. Most, but not all, also provided some form of relaxation train- ing. These elements may represent common components of successful group psychother- apy for HIV-infected individuals with distress. Wk6 Assignment 1 Discussion

With respect to demographics it is interest- ing to note that all but one of the studies was conducted among men. These findings may in part be result of the demographic nature of the epidemic over time. In the late 1980s and early 1990s HIV was considered primarily a disease of men.33 However, the emerging population at risk for HV are now non-white and Hispanic women. Providing effective interventions that target depressive symptoms among women is especially important as two prospective stud- ies demonstrate that compared to nonde- pressed women with HIV, women with depres- sive symptoms are significantly at increased risk of mortality.4,5 Furthermore, being a woman is considered an independent risk fac- tor for depression.34,35 Because some studies suggest that mental health interventions may in fact be protective 9 it is important to ensure that women are accessing appropriate mental health treatment. As the results of the meta- analysis may not generalize to women, future studies may be needed to address this dis- parity. Wk6 Assignment 1 Discussion

Minorities appeared to be well represented in most of the studies evaluated. Among the 5 studies that occurred in the United States, mi- norities represented, on average, about half of the participant sample.

There are several limitations to this study. First, many of the studies occurred prior to the HAART era and as such we were unable to ad- dress whether or not adherence to HAART was an important moderator of response. Studies have shown that individuals with HIV and de- pressive disorders, compared to those with HIV alone, have worse adherence to taking an- tiretroviral medication.6,36,37 However, studies have also found that mental health treatment increases the probability that individuals with depression receive and utilize HAART.9,38,39

Thus, it is possible that interventions that re- duce depressive symptoms may in fact im- prove access to and adherence with HAART. Future meta-analyses may be able to better ad- dress this outcome.

Second, only a couple of studies provided in- formation of CD4 counts or HIV disease sever- ity and therefore we were unable to determine the impact this may have had on treatment re- sponse. As there did not appear to be any sig- nificant heterogeneity in the studies investi- gated, it is unclear whether severity of illness would be important moderators to consider in a meta-regression. Third, we acknowledge that individuals enrolled in clinical trials may be more adherent to interventions and may be dif- ferent then patients seen in actual clinical prac- tice. This may then limit the generalizability of the findings of this meta-analysis.

Finally, we used a unit-free, standardized score, the effect size, in order to combine the results from several depression instruments. By combining the results of the depression instru- ments in this way we avoided the possibility of selection bias (i.e., not including results in the meta-analysis because they contained different depression outcome measures) and increased the overall power of our analysis. This method, though, assumes that the different instruments used in the meta-analysis, in fact, measure the same construct (i.e., depression) and are simi- larly responsive to symptom change. If these assumptions are not met, there is a potential for increased heterogeneity in the study results. As our study used instruments that are frequently used to measure depression and as we did not find any heterogeneity in our study results we believe that combing results from different de- pression instruments did not violate the above assumptions. Wk6 Assignment 1 Discussion

CONCLUSION

This study suggests that group therapy, and particularly group cognitive behavioral ther- apy may be efficacious in treating depressive symptoms among those infected with HIV. However, the underrepresentation of women limits the generalizability of these findings. Be- cause women may be at risk for depression and

 

 

are an emerging population at risk for HIV fu- ture studies should be directed to remedy this disparity.

REFERENCES

1. Bing EG, Burnam MA, Longshore D, et al. Psychiatric disorders and drug use among human immunodefi- ciency virus-infected adults in the United States. Arch Gen Psychiatry 2001;58:721–728.

2. Leserman J, Petitto JM, Gu H, et al. Progression to AIDS, a clinical AIDS condition and mortality: Psy- chosocial and physiological predictors. Psychol Med 2002;32:1059–1073.

3. McDaniel JS, Fowlie E, Summerville MB, Farber EW, Cohen-Cole SA. An assessment of rates of psychiatric morbidity and functioning in HIV disease. Gen Hosp Psychiatry 1995;17:346–352.

4. Ickovics JR, Hamburger ME, Vlahov D, et al. Mortal- ity, CD4 cell count decline, and depressive symptoms among HIV-seropositive women: Longitudinal anal- ysis from the HIV Epidemiology Research Study. JAMA 2001;285:1466–1474.

5. Cook JA, Grey D, Burke J, et al. Depressive symptoms and AIDS-related mortality among a multisite cohort of HIV-positive women. Am J Public Health 2004;94: 1133–1140.

6. Gordillo V, del Amo J, Soriano V, Gonzalez-Lahoz J. Sociodemographic and psychological variables influ- encing adherence to antiretroviral therapy. AIDS 1999;13:1763–1769.

7. Fairfield KM, Libman H, Davis RB, Eisenberg DM. Delays in protease inhibitor use in clinical practice. J Gen Intern Med 1999;14:395–401.

8. DiMatteo MR, Lepper HS, Croghan TW. Depression is a risk factor for noncompliance with medical treat- ment: Meta-analysis of the effects of anxiety and de- pression on patient adherence. Arch Intern Med 2000;160:2101.

9. Himelhoch S, Treisman G, Moore RA, Gebo K. Does presence of a mental disorder in AIDS patients affect the initiation of anitretroviral treatment and duration of therapy? J Acquir Immune Defic Syndr 2004;37: 1457–1463.

10. Yun LWH, Maravi M, Koayashi JS, Barton PL, David- son AJ. Antidepressant treatment improves adhernce to antiretorviral therapy among depressed HIV-in- fected patients. J Acquir Immune Defic Syndr 2005; 38:432–438.

11. Himelhoch S, Medoff DR. Efficacy of antidepressant medication among HIV� individuals with depres- sion: A systematic review and meta-analysis. AIDS Patient Care STDs 2005;19:813–822.

12. Antoni MH, Carrico AW, Duran RE, et al. Ran- domized clinical trial of cognitive behavioral stress management on human immunodeficiency virus viral load in gay men treated with highly active antiretroviral therapy. Psychosom Med 2006;68: 143–151.

13. Chan I, Kong P, Leung P, Cognitive-behavioral group program for Chinese heterosexual HIV-infected men in Hong Kong. Patient Educ Couns 2005;56:78–84.

14. Kelly JA, Murphy DA, Bahr GR, et al. Outcome of cognitive-behavioral and support group brief thera- pies for depressed, HIV-infected persons. Am J Psy- chiatry 1993;150:1679–1686.

15. Lutgendorf SK, Antoni MH, Ironson G, et al. Cogni- tive-behavioral stress management decreases dys- phoric mood and herpes simplex virus-type 2 anti- body titers in symptomatic HIV-seropositive gay men. J Consult Clin Psychol 1997;65:31–43.

16. Sikkema KJ, Hansen NB, Kochman A, Tate DC, Difranceisco W. Outcomes from a randomized con- trolled trial of a group intervention for HIV positive men and women coping with AIDS-related loss and bereavement. Death Stud 2004;28:187–209.

17. Chesney MA, Chambers DB, Taylor JM, Johnson LM, Folkman S. Coping effectiveness training for men liv- ing with HIV: Results from a randomized clinical trial testing a group-based intervention. Psychosom Med 2003;65:1038–1046.

18. Goodkin K, Blaney NT, Feaster DJ, Baldewicz T, Burkhalter JE, Leeds B. A randomized controlled clin- ical trial of a bereavement support group intervention in human immunodeficiency virus type 1-seroposi- tive and -seronegative homosexual men. Arch Gen Psychiatry 1999;56:52–59.

19. Mulder CL, Emmelkamp PM, Antoni MH, Mulder JW, Sandfort TG, de Vries MJ. Cognitive-behavioral and experiential group psychotherapy for HIV-in- fected homosexual men: a comparative study. Psy- chosom Med 1994;56:423–431.

20. Detsky AS, Naylor CD, O’Rourke K, McGeer AJ, L’Abbe KA. Incorporating variations in the quality of individual randomized trials into meta-analysis. J Clin Epidemiol 1992;45:255–265.

21. Egger M, Smith GD, Altman DG, eds. Systematic Re- views in Health Care Meta-Analysis in Context. Lon- don: BMJ Books, 2001:3-475.

22. Markowitz JC, Kocsis JH, Fishman B, et al. Treatment of depressive symptoms in human immunodeficiency virus-positive patients. Arch Gen Psychiatry 1998;55: 452–457.

23. Targ EF, Karasic DH, Diefenbach PN, Anderson DA, Bystritsky A, Fawzy FI. Structured group therapy and fluoxetine to treat depression in HIV-positive per- sons. Psychosomatics 1994;35:132–137.

24. Weber R, Christen L, Loy M, et al. Randomized, placebo-controlled trial of Chinese herb therapy for HIV-1-infected individuals. J Acquir Immune Defic Syndr 1999;22:56–64.

25. Zisook S, Peterkin J, Goggin KJ, Sledge P, Atkinson JH, Grant I. Treatment of major depression in HIV- seropositive men. HIV Neurobehavioral Research Center Group. J Clin Psychiatry 1998;59:217–224.

26. Ironson G, Weiss S, Lydston D, et al. The impact of improved self-efficacy on HIV viral load and distress in culturally diverse women living with AIDS: The SMART/EST Women’s Project. AIDS Care 2005;17: 222–236.

HIMELHOCH ET AL.738

 

 

GROUP PSYCHOTHERAPY TO REDUCE DEPRESSIVE SYMPTOMS 739

27. Koopman C, Gore-Felton C, Marouf F, et al. Rela- tionships of perceived stress to coping, attachment and social support among HIV-positive persons. AIDS Care 2000;12:663–672.

28. Inouye J, Flannelly L, Flannelly KJ. The effectiveness of self-management training for individuals with HIV/AIDS. J Assoc Nurses AIDS Care 2001;12:71–82.

29. Carrico AW, Antoni MH, Duran RE, et al. Reductions in depressed mood and denial coping during cogni- tive behavioral stress management with HIV-positive gay men treated with HAART. Ann Behav Med 2006; 31:155–164.

30. Lechner SC, Antoni MH, Lydston D, et al. Cognitive- behavioral interventions improve quality of life in women with AIDS. J Psychosom Res 2003;54:253–261.

31. Antoni MH, Baggett L, Ironson G, et al. Cognitive-be- havioral stress management intervention buffers dis- tress responses and immunologic changes following notification of HIV-1 seropositivity. J Consult Clin Psychol 1991;59:906–915.

32. McDermut W, Miller IW, Brown RA. The efficacy of group psychotherapy for depression: A meta-analy- sis and review of the empirical literature. Clin Psy- chol 2001;8:98.

33. HIV and AIDS—United States, 1981–2000. MMWR Morb Mortal Wkly Rep 2001;50:430–434.

34. Regier DA, Narrow WE, Rae DS, Manderscheid RW, Locke BZ, Goodwin FK. The de facto US mental and addictive disorders service system. Epidemiologic catchment area prospective 1-year prevalence rates of disorders and services. Arch Gen Psychiatry 1993;50: 85–94.

35. Weissman MM, Bland RC, Canino GJ, et al. Cross-na- tional epidemiology of major depression and bipolar disorder. JAMA 1996;276:293–299.

36. Mugavero M, Ostermann J, Whetten K, et al. Barriers to antiretroviral adherence: The importance of de- pression, abuse, and other traumatic events. AIDS Pa- tient Care STDs 2006;20:418–28.

37. Wagner GJ. Predictors of antiretroviral adherence as measured by self-report, electronic monitoring, and medication diaries. AIDS Patient Care STDs 2002;16: 599–608.

38. Cook JA, Cohen MH, Burke J, et al. Effects of de- pressive symptoms and mental health quality of life on use of highly active antiretroviral therapy among HIV-seropositive women. J Acquir Immune Defic Syndr 2002;30:401–409.

39. Turner BJ, Fleishman JA, Wenger N, et al. Effects of drug abuse and mental disorders on use and type of antiretroviral therapy in HIV-infected persons. J Gen Intern Med 2001;16:625–633.

Address reprint requests to: Seth Himelhoch, M.D., M.P.H.

Department of Psychiatry Division of Services Research

737 Lombard Street, Room 516 Baltimore, MD 21201

E-mail: shimelho@psych.umaryland.edu