Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2017 Jul 1.
Published in final edited form as: AIDS Behav. 2016 Jul;20(7):1527–1534. doi: 10.1007/s10461-016-1415-9

The effect of depressive symptoms on adherence to daily oral PrEP in men who have sex with men and transgender women: A marginal structural model analysis of the iPrEx OLE study

Megha L Mehrotra 1,2,#, David V Glidden 2, Vanessa McMahan 1, K Rivet Amico 3, Sybil Hosek 4, Patricia Defechereux 1, Kenneth H Mayer 5, Valdilea G Veloso 6, Linda-Gail Bekker 7, Vivian I Avelino-Silva 8, Mauro Schechter 9, Robert M Grant 1,2
PMCID: PMC4912836  NIHMSID: NIHMS792947  PMID: 27125241

Introduction

Pre-exposure prophylaxis (PrEP) with oral emtricitabine and tenofovir disoproxil fumarate (FTC/TDF) has proven to be one of the most promising new HIV prevention tools. Over the past 5 years, results from numerous studies have supported its efficacy and safety.(1–6) These studies had a wide range of efficacy estimates from intention-to-treat analyses, but when adherence was consistently high, estimated efficacy was found to be greater than 90%. (1,3,7) As a result, in 2012, the US Food and Drug Administration approved the use of daily, oral FTC/TDF for the prevention of HIV.(8)

Adherence

As with any medication, PrEP efficacy is highly dependent on adherence. Currently, FTC/TDF for PrEP is indicated for daily oral dosing. However, modeling studies estimate that 4 pills/week are likely sufficient to achieve complete protection from HIV infection.(7) While this makes the regimen forgiving of occasional missed doses, achieving and sustaining high enough levels of adherence to secure protection from HIV remains challenging for some.(6,9) For example, in the iPrEx blinded trial and its open-label extension (OLE), transgender women (TGW) had consistently lower concentrations of drug than men who have sex with men (MSM) and, therefore, less protection from PrEP.(1,10,11)

iPrEx OLE offered PrEP to HIV-negative participants who had been enrolled in three prior randomized placebo-controlled PrEP trials. Those who reported having condomless receptive anal intercourse (ncRAI) or had serological evidence of herpes were more likely to elect to receive PrEP in OLE.(10) Higher education, older age, baseline ncRAI, greater number of sexual partners and a history of syphilis or herpes were associated with higher drug concentrations throughout the study. This suggests that some study participants successfully identify and appropriately react to times during which PrEP is most needed for protection from HIV (e.g., prevention-effective adherence(12)). However, given that many iPrEx OLE participants who had their blood levels tested were below the limit of quantitation, the need to better understand the factors that may influence PrEP adherence remains important.

Depression in MSM/TGW

An area that has been well-represented in the antiretroviral (ART) for HIV treatment literature and less so in the PrEP literature is the interplay between depression and adherence. Gaining a better understanding of the influence of depression on adherence to PrEP is particularly relevant given the high prevalence of depression and anxiety in MSM and TGW globally (31% and 62%, respectively).(13–17) Social isolation, stigma, and childhood trauma are all potential risk factors for depression and are more commonly experienced by MSM and TGW.(18) Prior studies have found that depressive symptoms may contribute to decreased agency, lower ability for self-care, and subsequently lower adherence to medication regimens.(18,19) Taken together, we hypothesize that depression in MSM and TGW may play an important role in adherence to PrEP. Additionally, given the higher prevalence of depression in TGW than MSM, we anticipate that the role of depression in adherence to PrEP may differ between TGW and MSM.

Given that depression has been shown to influence sexual behavior (16,20,21) and baseline sexual behavior is associated with subsequent higher adherence (10), we anticipate that sexual exposure is both a confounder and mediator of the effect of depression on adherence to PrEP.

Methods

Study design

iPrEx OLE was a cohort study that enrolled 1603 HIV-negative MSM/TGW at 11 study sites in 6 countries (Peru, Brazil, the United States, Ecuador, Thailand, and South Africa). The majority of study participants were from the Andean region (62%).(10) 1225 participants elected to take daily, oral FTC/TDF PrEP. All participants were followed for up to 72 weeks. Participants could elect to begin taking PrEP at any point during the first 48 weeks of the study.(10) Adherence was quantified by measuring drug levels in dried blood spots (DBS) in a nested case-cohort sample of the overall study (N=349)(22). Cases were all participants who became HIV infected during study follow-up, and controls were a sitestratified random sample of seronegative on-PrEP participants at baseline(10). All timepoints after drug dispensation were tested for drug concentrations. This analysis included all members of the original case-cohort sample for whom we had at least one assessment of depressive symptoms (N=334). Sampling weights were incorporated into the final model to reflect the case-cohort design.

Exposure, Outcome, and Other Covariates

Drug Concentration

Drug concentration measured in DBS at all time-points after PrEP dispensation (week 4, week 8, week 12, and every 12 weeks thereafter) reflects drug exposure over the prior month.(7) Using methods described previously,(23) drug concentrations were categorized according to dosing corresponding to the number of pills taken per week. The primary outcome for these analyses was a binary indicator of having DBS levels corresponding to taking ≥4 pills/week—a level that has been shown to be sufficient for maximum protection by PrEP.(7,10)

CES-D

Depressive symptoms were assessed at enrollment and every 24 weeks during follow-up using the 20-item Center for Epidemiologic Studies Depression Scale (CES-D). The CES-D is a valid and reliable screening tool for depression and anxiety symptoms. It has been validated in all 4 languages used in the iPrEx OLE study: Spanish, Portuguese, English and Thai.(24–29) The CES-D reflects depressive symptoms occurring in the prior 7 days,(30) and was categorized according to standard clinical cutoffs as follows: none-mild depression (CES-D 0-15); mild-moderate depression (CES-D 16-26); severe depression (CES-D 27+).(20) Pre-existing conditions of depression and anxiety were also recorded via interview at enrollment in the study.

Sexual Behavior

Sexual behavior was recorded at baseline and every 12 weeks thereafter via an interviewer-administered questionnaire. Questionnaires asked about behaviors occurring in the prior 12 weeks. Reported ncRAI and total number of male partners were the measures chosen to reflect sexual exposure in this analysis.

Baseline Confounders

Baseline confounders include study site, age at enrollment, highest level of education, gender identity (MSM or TGW), history of having a sexually transmitted infection (STI), pre-existing conditions of anxiety, and baseline sexual exposure measures (number of partners, history of having an HIV positive partner, and ncRAI).

Causal Model

Directed acyclic graphs (DAGs)(31) are non-parametric causal diagrams used to aid causal analysis. We hypothesized that sexual exposure functions as a time-varying confounder-mediator in the association between depressive symptoms and adherence to PrEP as described in DAG shown in Figure 1. The outcome is a binary indicator of DBS levels corresponding to ≥4 pills/week (DBS in the DAG). Time-fixed confounders (not shown in the DAG) include baseline sexual exposures, study site, age, highest level of education, and STI diagnosis s at baseline. To attempt to disentangle depressive symptoms from anxiety symptoms as measured by the CES-D, preexisting conditions of anxiety were also included in the model.

Figure 1.

Figure 1

Open in a new tab

Directed acyclic graph (DAG) illustrating the proposed causal model between depression, sexual behavior, and adherence to PrEP (as measured by drug concentrations in DBS) over time. Measures of sexual behavior reflect the prior 3 month interval, adherence measures (in DBS) reflect the prior 1 month, and CES-D captures depressive symptoms in the prior 7 days. Sexual behavior has been previously shown to predict subsequent adherence to PrEP in this cohort; depressive symptoms have been shown to influence subsequent sexual behavior in MSM; and sexual behavior may also influence subsequent depression. The model illustrates the need for a marginal structural model to control for timevarying confounding in which the confounder is also affected by prior exposure.

Statistical Analyses

A marginal structural logistic regression model (32) was used to estimate the association between depressive symptoms and protective adherence to daily, oral FTC/TDF PrEP in iPrEx OLE. In order to control for bias from time-varying confounding variables that may also be affected by prior depression status, we used inverse probability weights for the ordinal exposure categories (no depressive symptoms, mild depressive symptoms, and severe depressive symptoms) and censoring.(32,33) Stabilization with baseline variables provided a reasonable distribution of exposure weights (Mean: 1.0; Standard Deviation: 0.083; Range [0.21–3.12]).(34) The final weights were not truncated and included sampling weights, the exposure weights, and censoring weights. The structural model was estimated using logistic regression, robust standard errors, and included adjustment for all baseline covariates.(33) Multiplicative and additive interaction between CES-D scores and gender identity was also assessed using the final structural model.(35) Population attributable fraction (PAF) was calculated to estimate the absolute impact of treating depression on adherence to PrEP in both MSM and TGW. All analyses were conducted in STATA 13.1.(36)

Results

Study population

The final analysis included 334 participants. At baseline, most participants had a CES-D score less than 16 (72%), were from the Andean region (53%), were 25 years old or younger (32%), and had a maximum of a high school education (50%). 11% identified as a transgender woman, and 14% and 7% had reported pre-existing conditions of depression and anxiety respectively (Table I).

Table 1.

Baseline Characteristics

Frequency Percent
Baseline CESD Score
  CESD less than 16 239 72
  CESD 16–26 67 20
  CESD greater than 26 28 8
Study region
  Andes 176 53
  Brazil 70 21
  United States 55 16
  Thailand 18 5
  South Africa 15 4
Age
  <= 25 106 32
  > 25 – 30 77 23
  > 30 – 39 80 24
  > 40 71 21
Gender Identity
  MSM 296 89
  TGW 38 11
Education
  Less than HS 69 21
  High School 167 50
  College 98 29
Pre-Existing Conditions
  Pre-Existing Depression 46 14
  Pre-Existing Anxiety 24 7
  Total Number of Participants: 334
Open in a new tab

Depressive Symptoms and Adherence

The final structural model found a non-linear relationship between CES-D scores and odds of not having drug levels that conferred >90% protection from HIV infection based on the Wald test of linearity (chi2=16.71, p<0.001) in MSM. MSM with CES-D scores between 16 and 26 were more likely to have protective drug levels (OR 1.66 95%CI [1.05, 2.60]) compared to those in the reference lowest CES-D category. MSM with CES-D scores >26 were less likely to have protective drug levels compared to those in the lowest CES-D category (OR 0.41 95%CI [0.22, 0.77]). Older age, higher education, and pre-existing conditions of anxiety were all associated with greater odds of protective levels of drug detection (Table II).

Table II.

Odds of Protective Drug Levels

Odds ratio 95% Cl
CESD scores
  CESD less than 16 Ref –
  CESD 16–26 1.66 [1.05, 2.60]
  CESD greater than 26 0.41 [0.22, 2.39]
  Trans identity 1.02 [0.44, 2.39]
CESD Trans interaction
  CESD 16–26* Trans 0.24 [0.05, 1.11]
  CESD greater than 26* Trans 1.34 [0.09, 20.96]
  Anxiety Pre-existing condition 3.88 [1.62, 9.27]
  Clusters (Number of Participants) 334
Open in a new tab

Exponentiated coefficients; 95% confidence intervals in brackets

Model also controlled for study site; age; education; and sexual behavior

The model shows modest evidence in support of effect heterogeneity of CES-D on adherence between MSM and TGW, particularly in the middle category of CES-D scores (OR: 0.24 95%CI [0.05, 1.11]). Among TGW, participants with CES-D scores between 16 and 26 or CES-D scores >26 were less likely to have protective drug levels (OR 0.39 95%CI [0.09, 1.63]) and OR 0.55 95%CI [0.04, 8.04] respectively) than MSM. The marginal probability of protective adherence between transgender women and MSM by CES-D category shows the greatest discrepancy in adherence among those with CES-D scores between 16 and 26. (Table II)

Population Attributable Fraction

Given the baseline prevalence of mild-moderate depressive symptoms and severe depressive symptoms in MSM and TGW, the percent of inadequate adherence to PrEP that can be attributed to elevated CES-D is up to 1% and 6% respectively. (Table III)

Table III.

Population Attributable Fraction (PAF) of Depressive Symptoms in MSM and Transgender Women.

Population level impact of severe depression on adherence to PrEP.

Baseline
Prevalence		 Marginal
Probability of
Insufficient
Adherence% 		Odds
Ratio* 		Population
Attributable
Fraction#
CES-D < 16 0.73 0.6 -- --
MSM 16 ≤ CES-D < 27 0.2 0.5 0.6 −0.03+
27 ≤ CES-D 0.07 0.74 2.44 0.01
CES-D < 16 0.61 0.59 -- --
TGW 16 ≤ CES-D < 27 0.24 0.74 2.56 0.07
27 ≤ CES-D 0.16 0.69 1.82 0.03
Open in a new tab
%

Marginal probability of insufficient adherence

*

Odds of not having sufficient drug levels for full protection

#

Population attributable fraction represents the proportion of non-adherence that would be resolved if an intervention were to move all CES-D scores to the lowest category while holding all else constant.

+

Negative value reflects the fact that adherence was highest among those with CES-D scores corresponding to mild-moderate depressive symptoms

Discussion

Overall, depressive symptoms played a modest role in adherence to PrEP, and the effect of depressive symptoms on adherence likely differs between MSM and TGW. In MSM, the relationship between depressive symptoms (as measured by the CES-D) and adherence is non-monotonic. Compared to CES-D scores below 16, we found that scores between 16 and 26 were associated with enhanced adherence, while CES-D scores of 27 or higher impaired adherence to PrEP regimens. In TGW, we find that any elevated CES-D scores above 16 are associated with decreased adherence to PrEP.

Severe depressive symptoms were substantially less common than mild-moderate depressive symptoms. It is possible that in cases of severe depression, because of psychomotor retardation, capacity for self-care and agency may be impaired such that adherence to medication becomes more challenging, as has been cited in other areas of depression research.(15,18,19) However, the absolute population impact of severe depressive symptoms on adherence to PrEP is relatively low. Based on the PAF, intervening on severe depressive symptoms would resolve at most 1% of the non-adherence in MSM and 3% in TGW. Thus, it is clear that severe depressive symptoms, as measured by the CES-D, do not play a major direct role in non-adherence to PrEP and should not preclude prescription of PrEP to those who need it.

In understanding the effect of moderate depressive symptoms (CES-D scores between 16 and 26) on adherence in MSM, it is important to consider the potential role of anxiety. The CES-D has been shown to be sensitive to both depressive and anxiety symptoms, but specific to neither.(30,37) Thus, it is possible that anxiety and depression may have opposite effects on adherence to PrEP. Controlling for baseline pre-existing conditions of anxiety was intended to help separate depression from anxiety in the CES-D measures, however misclassification of anxiety, frequent co-morbidity of the two conditions, and differential reporting of anxiety between sites may allow for there to be residual confounding by anxiety (Figure 2). The fact that pre-existing anxiety is such a strong predictor of increased adherence in MSM suggests that this hypothesis may be reasonable (Table II), but the overall findings regarding depressive symptoms and adherence were not significantly changed when removing preexisting anxiety from the model (data not shown) suggesting that anxiety likely does not fully explain the observed pattern.

Figure 2.

Figure 2

Open in a new tab

A directed acyclic graph (DAG) illustrating the potential role of anxiety in confounding the relationship between CES-D and adherence to PrEP. This analysis adjusted for reported anxiety at baseline to block this confounding pathway. However, given the potential misclassification of anxiety in the study population, it is possible that residual confounding by anxiety may contribute to the observed non-monotonic relationship between CES-D and adherence.

While severe anxiety has been shown to decrease adherence to HIV treatment,(38,39) mild to moderate anxiety may make some PrEP users more worried about potential HIV infections and more likely to adhere than those without any anxiety symptoms. Further work designed to disentangle the two constructs should be conducted to elucidate the individual effect of each on adherence. For example, the Beck depression inventory is a screening tool that is more specific to depressive symptoms alone and may be better suited for future studies to isolate the effect of depression.(40)

These analyses, while limited by a small sample of TGW, show a striking difference in the effect of depressive symptoms on adherence between MSM and TGW. Based on the marginal probabilities of protective adherence by CES-D category, it is evident that the discrepancy in adherence to PrEP between MSM and TGW that has been reported in prior studies(1,10,11) is concentrated among those with CES-D scores between 16 and 26 (Figure 3). Importantly, 24% of TGW in the current sample had CES-D scores that fell within this range, and thus the PAF for non-adherence is highest in this group (7%) (Table III).

Figure 3.

Figure 3

Open in a new tab

The marginal probability of protective adherence by CES-D category and by MSM/TGW Table III: Population level impact of severe depression on adherence to PrEP.

Gender is known to play a role in the optimal CES-D cut-off to reflect clinical symptoms of depression, and studies have indicated that the cutoff should be higher for women than men (29). Post-hoc sensitivity analyses using higher CES-D cutoffs for categorizing TGW depressive symptoms did not impact the findings in the primary analyses (data not shown).

TGW in the iPrEx OLE study were concentrated in only a few study sites of the overall cohort, and differed from MSM in a number of important demographic characteristics (11). Thus, the standard regression adjustment methods employed here cannot fully disentangle the precise reasons for the observed differences in adherence between TGW and MSM. Future analyses designed specifically for addressing this question are required.

While the findings in TGW did not reach statistical significance, the difference in effect estimates of CES-D and in marginal probability of adherence between MSM and TGW is substantial and merits closer review. These results highlight the fact that MSM and TGW are two distinct populations with different cultural contexts and structural barriers to PrEP access and adherence. Studies that treat these populations as one homogenous group may inadvertently obfuscate important predictors of PrEP uptake and adherence that vary between the two populations. Additional studies with larger transgender populations need to be conducted in order to fully understand the barriers and facilitators of effective PrEP adherence.

Haberer and colleagues recently outlined a new paradigm for defining successful PrEP adherence, coined “prevention-effective adherence.”(12) This strategy defines successful adherence not just in terms of absolute drug-detection, but rather as drug-detection in the context of changing HIV-acquisition risk. Our study aimed to move closer to this outcome of interest by focusing on drug detection commensurate with full protection from HIV rather than absolute drug levels and by controlling for time-dependent confounding by sexual behavior. While a helpful step in the right direction, further work needs to be done to better define a more specific outcome that reflects true prevention-effective adherence.

In conclusion, it is unlikely that depressive symptoms alone contribute greatly to decreased adherence to PrEP in either MSM or TGW on a population level. Given this, and the fact that PrEP has not been shown to increase depression in MSM/TGW(41), depressive symptoms should not contraindicate the prescription of PrEP. Instead, PrEP clinics should be used as another avenue for screening and treating mental health issues in this often vulnerable population. Further efforts should be made in both the research and clinical setting to explore and address the differences between TGW and MSM that may contribute to successful prevention-effective PrEP adherence.

Acknowledgments

Funding Source: We would like to thank and acknowledge the iPrEx and iPrEx OLE study participants. The iPrEx OLE study was funded by the US National Institutes of Health (5U01AI064002); study medication was donated by Gilead Sciences, which also supported travel expenses for non-US investigators to attend study meetings.

Footnotes

Conflict of Interest: DVG has received fees from ViiV, a manufacturer of an investigational compound being investigated for use as PrEP. RA has an educational grant from Gilead Sciences through the University of Michigan. KM has received unrestricted research and education grants from Gilead Sciences. RMG has received honoraria from Clinical Care Options. RMG served as a consultant for Siemens Healthcare on their guidelines panel. The remaining authors have no conflicts of interest to disclose.

Ethical Statement:

Ethical approval: All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Informed consent: Informed consent was obtained from all individual participants included in the study.

References

  • 1.Grant RM, Lama JR, Anderson PL, McMahan V, Liu AY, Vargas L, et al. Preexposure chemoprophylaxis for HIV prevention in men who have sex with men. N Engl J Med. 2010 Dec 30;363(27):2587–2599. doi: 10.1056/NEJMoa1011205. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Baeten JM, Donnell D, Ndase P, Mugo NR, Campbell JD, Wangisi J, et al. Antiretroviral Prophylaxis for HIV-1 Prevention among Heterosexual Men and Women. N Engl J Med. 2012 Aug 2;367(5):399–410. doi: 10.1056/NEJMoa1108524. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Choopanya K, Martin M, Suntharasamai P, Sangkum U, Mock PA, Leethochawalit M, et al. Antiretroviral prophylaxis for HIV infection in injecting drug users in Bangkok, Thailand (the Bangkok Tenofovir Study): a randomised, double-blind, placebo-controlled phase 3 trial. The Lancet. 2013 Jun 21;381(9883):2083–2090. doi: 10.1016/S0140-6736(13)61127-7. [DOI] [PubMed] [Google Scholar]
  • 4.Marrazzo JM, Ramjee G, Richardson BA, Gomez K, Mgodi N, Nair G, et al. Tenofovir-Based Preexposure Prophylaxis for HIV Infection among African Women. N Engl J Med. 2015 Feb 5;372(6):509–518. doi: 10.1056/NEJMoa1402269. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Thigpen MC, Kebaabetswe PM, Paxton LA, Smith DK, Rose CE, Segolodi TM, et al. Antiretroviral Preexposure Prophylaxis for Heterosexual HIV Transmission in Botswana. N Engl J Med. 2012 Aug 2;367(5):423–434. doi: 10.1056/NEJMoa1110711. [DOI] [PubMed] [Google Scholar]
  • 6.Van Damme L, Corneli A, Ahmed K, Agot K, Lombaard J, Kapiga S, et al. Preexposure Prophylaxis for HIV Infection among African Women. N Engl J Med. 2012 Aug 2;367(5):411–422. doi: 10.1056/NEJMoa1202614. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7.Anderson PL, Glidden DV, Liu A, Buchbinder S, Lama JR, Guanira JV, et al. Emtricitabine-Tenofovir Concentrations and Pre-Exposure Prophylaxis Efficacy in Men Who Have Sex with Men. Sci Transl Med. 2012 Sep 12;4(151) doi: 10.1126/scitranslmed.3004006. 151ra125–51ra125. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8. [cited 2015 Aug 17];Commissioner O of the. Press Announcements - FDA approves first drug for reducing the risk of sexually acquired HIV infection [Internet] Available from: http://www.fda.gov/NewsEvents/Newsroom/PressAnnouncements/ucm312210.htm.
  • 9.Corneli AL, Deese J, Wang M, Taylor D, Ahmed K, Agot K, et al. FEM-PrEP: adherence patterns and factors associated with adherence to a daily oral study product for pre-exposure prophylaxis. J Acquir Immune Defic Syndr. 1999. 2014 Jul 1;66(3):324–331. doi: 10.1097/QAI.0000000000000158. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Grant RM, Anderson PL, McMahan V, Liu A, Amico KR, Mehrotra M, et al. Uptake of pre-exposure prophylaxis, sexual practices, and HIV incidence in men and transgender women who have sex with men: a cohort study. Lancet Infect Dis. 2014 Sep;14(9):820–829. doi: 10.1016/S1473-3099(14)70847-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Deutsch MB, Glidden DV, Sevelius J, Keatley J, McMahan V, Guanira J, et al. HIV pre-exposure prophylaxis in transgender women: a subgroup analysis of the iPrEx trial. [cited 2015 Nov 9];Lancet HIV [Internet] 2015 Nov; doi: 10.1016/S2352-3018(15)00206-4. Available from: http://linkinghub.elsevier.com/retrieve/pii/S2352301815002064. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Haberer JE, Bangsberg DR, Baeten JM, Curran K, Koechlin F, Amico KR, et al. Defining success with HIV pre-exposure prophylaxis: a prevention-effective adherence paradigm. AIDS Lond Engl. 2015 Jul 17;29(11):1277–1285. doi: 10.1097/QAD.0000000000000647. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Cochran SD, Mays VM, Sullivan JG. Prevalence of mental disorders, psychological distress, and mental health services use among lesbian, gay, and bisexual adults in the United States. J Consult Clin Psychol. 2003 Feb;71(1):53–61. doi: 10.1037//0022-006x.71.1.53. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14.Hoffman B. An Overview of Depression among Transgender Women. Depress Res Treat. 2014 Mar 13;2014:e394283. doi: 10.1155/2014/394283. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Safren SA, O’Cleirigh C, Tan JY, Raminani SR, Reilly LC, Otto MW, et al. A randomized controlled trial of cognitive behavioral therapy for adherence and depression (CBT-AD) in HIV-infected individuals. Health Psychol Off J Div Health Psychol Am Psychol Assoc. 2009 Jan;28(1):1–10. doi: 10.1037/a0012715. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Hatzenbuehler ML, O’Cleirigh C, Mayer KH, Mimiaga MJ, Safren SA. Prospective associations between HIV-related stigma, transmission risk behaviors, and adverse mental health outcomes in men who have sex with men. Ann Behav Med Publ Soc Behav Med. 2011 Oct;42(2):227–234. doi: 10.1007/s12160-011-9275-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.O’Cleirigh C, Skeer M, Mayer KH, Safren SA. Functional impairment and health care utilization among HIV-infected men who have sex with men: the relationship with depression and post-traumatic stress. J Behav Med. 2009 Oct;32(5):466–477. doi: 10.1007/s10865-009-9217-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Mays VM, Cochran SD. Mental health correlates of perceived discrimination among lesbian, gay, and bisexual adults in the United States. Am J Public Health. 2001 Nov;91(11):1869–1876. doi: 10.2105/ajph.91.11.1869. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Lin EHB, Katon W, Von Korff M, Rutter C, Simon GE, Oliver M, et al. Relationship of depression and diabetes self-care, medication adherence, and preventive care. Diabetes Care. 2004 Sep;27(9):2154–2160. doi: 10.2337/diacare.27.9.2154. [DOI] [PubMed] [Google Scholar]
  • 20.Koblin BA, Husnik MJ, Colfax G, Huang Y, Madison M, Mayer K, et al. Risk factors for HIV infection among men who have sex with men: AIDS. 2006 Mar;20(5):731–739. doi: 10.1097/01.aids.0000216374.61442.55. [DOI] [PubMed] [Google Scholar]
  • 21.Salomon EA, Mimiaga MJ, Husnik MJ, Welles SL, Manseau MW, Montenegro AB, et al. Depressive symptoms, utilization of mental health care, substance use and sexual risk among young men who have sex with men in EXPLORE: implications for age-specific interventions. AIDS Behav. 2009 Aug;13(4):811–821. doi: 10.1007/s10461-008-9439-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Prentice RL. A case-cohort design for epidemiologic cohort studies and disease prevention trials. Biometrika. 1986 Apr 1;73(1):1–11. [Google Scholar]
  • 23.Castillo-Mancilla JR, Zheng J-H, Rower JE, Meditz A, Gardner EM, Predhomme J, et al. Tenofovir, emtricitabine, and tenofovir diphosphate in dried blood spots for determining recent and cumulative drug exposure. AIDS Res Hum Retroviruses. 2013 Feb;29(2):384–390. doi: 10.1089/aid.2012.0089. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Crockett LJ, Randall BA, Shen Y-L, Russell ST, Driscoll AK. Measurement equivalence of the center for epidemiological studies depression scale for Latino and Anglo adolescents: a national study. J Consult Clin Psychol. 2005 Feb;73(1):47–58. doi: 10.1037/0022-006X.73.1.47. [DOI] [PubMed] [Google Scholar]
  • 25.Radloff LS. The CES-D Scale A Self-Report Depression Scale for Research in the General Population. Appl Psychol Meas. 1977 Jun 1;1(3):385–401. [Google Scholar]
  • 26.Ruiz-Grosso P, Loret de Mola C, Vega-Dienstmaier JM, Arevalo JM, Chavez K, Vilela A, et al. Validation of the Spanish Center for Epidemiological Studies Depression and Zung Self-Rating Depression Scales: A Comparative Validation Study. PLoS ONE. 2012 Oct 8;7(10):e45413. doi: 10.1371/journal.pone.0045413. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Reuland DS, Cherrington A, Watkins GS, Bradford DW, Blanco RA, Gaynes BN. Diagnostic Accuracy of Spanish Language Depression-Screening Instruments. Ann Fam Med. 2009 Sep 1;7(5):455–462. doi: 10.1370/afm.981. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Vatanasin D, Thapinta D, Thompson EA, Thungjaroenkul P. Testing a Model of Depression Among Thai Adolescents. J Child Adolesc Psychiatr Nurs. 2012 Nov 1;25(4):195–206. doi: 10.1111/jcap.12012. [DOI] [PubMed] [Google Scholar]
  • 29.Lewinsohn PM, Rohde P, Seeley JR. Adolescent Suicidal Ideation and Attempts: Prevalence, Risk Factors, and Clinical Implications. Clin Psychol Sci Pract. 1996 Mar 1;3(1):25–46. [Google Scholar]
  • 30.Orme JG, Reis J, Herz EJ. Factorial and discriminant validity of the Center for Epidemiological Studies Depression (CES-D) scale. J Clin Psychol. 1986 Jan;42(1):28–33. doi: 10.1002/1097-4679(198601)42:1<28::aid-jclp2270420104>3.0.co;2-t. [DOI] [PubMed] [Google Scholar]
  • 31.Greenland S, Pearl J, Robins JM. Causal Diagrams for Epidemiologic Research. Epidemiology. 1999 Jan 1;10(1):37–48. [PubMed] [Google Scholar]
  • 32.Robins JM, Hernán MA, Brumback B. Marginal structural models and causal inference in epidemiology. Epidemiol Camb Mass. 2000 Sep;11(5):550–560. doi: 10.1097/00001648-200009000-00011. [DOI] [PubMed] [Google Scholar]
  • 33.Fewell Z, Hernán MA, Wolfe F, Tilling K, Choi H, Sterne JA. Controlling for time-dependent confounding using marginal structural models. Stata J. 2004;4(4):402–420. [Google Scholar]
  • 34.Cole SR, Hernán MA. Constructing inverse probability weights for marginal structural models. Am J Epidemiol. 2008 Sep 15;168(6):656–664. doi: 10.1093/aje/kwn164. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35.VanderWeele TJ, Knol MJ. A Tutorial on Interaction. [cited 2015 Aug 18];Epidemiol Methods [Internet] 2014 Jan 1;3(1) Available from: http://www.degruyter.com/view/j/em.2014.3.issue-1/em-2013-0005/em-2013-0005.xml. [Google Scholar]
  • 36.STATA Statistical Software: Release 13. College Station, TX: StataCorp LP; [Google Scholar]
  • 37.Roberts RE, Rhoades HM, Vernon SW. Using the CES-D scale to screen for depression and anxiety: Effects of language and ethnic status. Psychiatry Res. 1990 Jan;31(1):69–83. doi: 10.1016/0165-1781(90)90110-q. [DOI] [PubMed] [Google Scholar]
  • 38.Campos LN, Guimarães MDC, Remien RH. Anxiety and depression symptoms as risk factors for non-adherence to antiretroviral therapy in Brazil. AIDS Behav. 2010 Apr;14(2):289–299. doi: 10.1007/s10461-008-9435-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Nel A, Kagee A. The relationship between depression, anxiety and medication adherence among patients receiving antiretroviral treatment in South Africa. AIDS Care. 2013 Aug;25(8):948–955. doi: 10.1080/09540121.2012.748867. [DOI] [PubMed] [Google Scholar]
  • 40.Beck AT, Steer RA, Carbin MG. Psychometric properties of the Beck Depression Inventory: Twenty-five years of evaluation. Clin Psychol Rev. 1988;8(1):77–100. [Google Scholar]
  • 41.Defechereux P, Mehrotra M, Liu A, McMahan V, Glidden D, Mayer K, et al. Depression and Oral FTC/TDF Pre-exposure Prophylaxis (PrEP) Among Men and Transgender Women Who Have Sex With Men (MSM/TGW) AIDS Behav. 2015 Jun 16;:1–11. doi: 10.1007/s10461-015-1082-2. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES