[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Figure.
Association Between Percentage of Days Spent With Depression and Risk of Missed HIV Primary Care Appointments, Risk of Detectable Viral Load, and Mortality Rate Among 5927 Adults Receiving HIV Primary Care in the United States
Association Between Percentage of Days Spent With Depression and Risk of Missed HIV Primary Care Appointments, Risk of Detectable Viral Load, and Mortality Rate Among 5927 Adults Receiving HIV Primary Care in the United States

A, Predicted probability of missed HIV primary care appointments, with 95% CIs from pooled Poisson regression models with robust variance, with the exposure (percentage of days with depression) modeled using a restricted cubic spline functional form. B, Predicted probability of detectable viral load, with 95% CIs from pooled Poisson regression models with robust variance, with the exposure (percentage of days with depression) modeled using a restricted cubic spline functional form. C, Hazard ratios of mortality rate, with 95% CIs from a Cox proportional hazards model, relative to a referent of 0% of days spent with depression, with the exposure modeled using a restricted cubic spline functional form; the blue horizontal line denotes the null value of 1.0. PHQ-9 indicates Patient Health Questionnaire-9. Dashed lines indicate the 95% CI.

Table 1.  
Characteristics of 5927 US Adults Receiving HIV Primary Care at Sites Participating in the Center for AIDS Research Network of Integrated Clinical Systems, 2005-2015
Characteristics of 5927 US Adults Receiving HIV Primary Care at Sites Participating in the Center for AIDS Research Network of Integrated Clinical Systems, 2005-2015
Table 2.  
Association of Increased Depression Burden With All-Cause Mortality, Missed Visits, and Lack of Viral Suppression Among 5927 Adults Receiving HIV Primary Care in the United States
Association of Increased Depression Burden With All-Cause Mortality, Missed Visits, and Lack of Viral Suppression Among 5927 Adults Receiving HIV Primary Care in the United States
1.
Pence  BW, Miller  WC, Whetten  K, Eron  JJ, Gaynes  BN.  Prevalence of DSM-IV-defined mood, anxiety, and substance use disorders in an HIV clinic in the Southeastern United States.  J Acquir Immune Defic Syndr. 2006;42(3):298-306.PubMedGoogle ScholarCrossref
2.
Bing  EG, Burnam  MA, Longshore  D,  et al.  Psychiatric disorders and drug use among human immunodeficiency virus-infected adults in the United States.  Arch Gen Psychiatry. 2001;58(8):721-728.PubMedGoogle ScholarCrossref
3.
Ciesla  JA, Roberts  JE.  Meta-analysis of the relationship between HIV infection and risk for depressive disorders.  Am J Psychiatry. 2001;158(5):725-730.PubMedGoogle ScholarCrossref
4.
Orlando  M, Burnam  MA, Beckman  R,  et al.  Re-estimating the prevalence of psychiatric disorders in a nationally representative sample of persons receiving care for HIV: results from the HIV Cost and Services Utilization Study.  Int J Methods Psychiatr Res. 2002;11(2):75-82.PubMedGoogle ScholarCrossref
5.
Schumacher  JE, McCullumsmith  C, Mugavero  MJ,  et al.  Routine depression screening in an HIV clinic cohort identifies patients with complex psychiatric co-morbidities who show significant response to treatment.  AIDS Behav. 2013;17(8):2781-2791.PubMedGoogle ScholarCrossref
6.
Tegger  MK, Crane  HM, Tapia  KA, Uldall  KK, Holte  SE, Kitahata  MM.  The effect of mental illness, substance use, and treatment for depression on the initiation of highly active antiretroviral therapy among HIV-infected individuals.  AIDS Patient Care STDS. 2008;22(3):233-243.PubMedGoogle ScholarCrossref
7.
O’Cleirigh  C, Magidson  JF, Skeer  MR, Mayer  KH, Safren  SA.  Prevalence of psychiatric and substance abuse symptomatology among HIV-infected gay and bisexual men in HIV primary care.  Psychosomatics. 2015;56(5):470-478.PubMedGoogle ScholarCrossref
8.
DiMatteo  MR, Lepper  HS, Croghan  TW.  Depression is a risk factor for noncompliance with medical treatment: meta-analysis of the effects of anxiety and depression on patient adherence.  Arch Intern Med. 2000;160(14):2101-2107.PubMedGoogle ScholarCrossref
9.
Zuniga  JA, Yoo-Jeong  M, Dai  T, Guo  Y, Waldrop-Valverde  D.  The role of depression in retention in care for persons living with HIV.  AIDS Patient Care STDS. 2016;30(1):34-38.PubMedGoogle ScholarCrossref
10.
Pence  BW, Miller  WC, Gaynes  BN, Eron  JJ  Jr.  Psychiatric illness and virologic response in patients initiating highly active antiretroviral therapy.  J Acquir Immune Defic Syndr. 2007;44(2):159-166.PubMedGoogle ScholarCrossref
11.
Mugavero  MJ, Pence  BW, Whetten  K,  et al.  Predictors of AIDS-related morbidity and mortality in a southern US cohort.  AIDS Patient Care STDS. 2007;21(9):681-690.PubMedGoogle ScholarCrossref
12.
Ironson  G, O’Cleirigh  C, Fletcher  MA,  et al.  Psychosocial factors predict CD4 and viral load change in men and women with human immunodeficiency virus in the era of highly active antiretroviral treatment.  Psychosom Med. 2005;67(6):1013-1021.PubMedGoogle ScholarCrossref
13.
Todd  JV, Cole  SR, Pence  BW,  et al.  Effects of antiretroviral therapy and depressive symptoms on all-cause mortality among HIV-infected women.  Am J Epidemiol. 2017;185(10):869-878.PubMedGoogle ScholarCrossref
14.
Ickovics  JR, Hamburger  ME, Vlahov  D,  et al; HIV Epidemiology Research Study Group.  Mortality, CD4 cell count decline, and depressive symptoms among HIV-seropositive women: longitudinal analysis from the HIV Epidemiology Research Study.  JAMA. 2001;285(11):1466-1474.PubMedGoogle ScholarCrossref
15.
Anastos  K, Schneider  MF, Gange  SJ,  et al; Women’s Interagency HIV Study Collaborative Group.  The association of race, sociodemographic, and behavioral characteristics with response to highly active antiretroviral therapy in women.  J Acquir Immune Defic Syndr. 2005;39(5):537-544.PubMedGoogle Scholar
16.
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(7):1133-1140.PubMedGoogle ScholarCrossref
17.
French  AL, Gawel  SH, Hershow  R,  et al.  Trends in mortality and causes of death among women with HIV in the United States: a 10-year study.  J Acquir Immune Defic Syndr. 2009;51(4):399-406.PubMedGoogle ScholarCrossref
18.
Hessol  NA, Kalinowski  A, Benning  L,  et al.  Mortality among participants in the Multicenter AIDS Cohort Study and the Women’s Interagency HIV Study.  Clin Infect Dis. 2007;44(2):287-294.PubMedGoogle ScholarCrossref
19.
Leserman  J, Pence  BW, Whetten  K,  et al.  Relation of lifetime trauma and depressive symptoms to mortality in HIV.  Am J Psychiatry. 2007;164(11):1707-1713.PubMedGoogle ScholarCrossref
20.
Lima  VD, Geller  J, Bangsberg  DR,  et al.  The effect of adherence on the association between depressive symptoms and mortality among HIV-infected individuals first initiating HAART.  AIDS. 2007;21(9):1175-1183.PubMedGoogle ScholarCrossref
21.
Murphy  K, Hoover  DR, Shi  Q,  et al.  Association of self-reported race with AIDS death in continuous HAART users in a cohort of HIV-infected women in the United States.  AIDS. 2013;27(15):2413-2423.PubMedGoogle ScholarCrossref
22.
Wada  N, Jacobson  LP, Cohen  M, French  A, Phair  J, Muñoz  A.  Cause-specific life expectancies after 35 years of age for human immunodeficiency syndrome–infected and human immunodeficiency syndrome–negative individuals followed simultaneously in long-term cohort studies, 1984-2008.  Am J Epidemiol. 2013;177(2):116-125.PubMedGoogle ScholarCrossref
23.
Cohen  MH, Hotton  AL, Hershow  RC,  et al.  Gender-related risk factors improve mortality predictive ability of VACS index among HIV-infected women.  J Acquir Immune Defic Syndr. 2015;70(5):538-544.PubMedGoogle ScholarCrossref
24.
Cohen  MH, French  AL, Benning  L,  et al.  Causes of death among women with human immunodeficiency virus infection in the era of combination antiretroviral therapy.  Am J Med. 2002;113(2):91-98.PubMedGoogle ScholarCrossref
25.
Kroenke  K, Spitzer  RL.  The PHQ-9: a new depression diagnostic and severity measure.  Psychiatr Ann. 2002;32(9):509-515. doi:10.3928/0048-5713-20020901-06Google ScholarCrossref
26.
Kitahata  MM, Rodriguez  B, Haubrich  R,  et al.  Cohort profile: the Centers for AIDS Research Network of Integrated Clinical Systems.  Int J Epidemiol. 2008;37(5):948-955.PubMedGoogle ScholarCrossref
27.
Spitzer  RL, Kroenke  K, Williams  JB.  Validation and utility of a self-report version of PRIME-MD: the PHQ primary care study: Primary Care Evaluation of Mental Disorders: Patient Health Questionnaire.  JAMA. 1999;282(18):1737-1744.PubMedGoogle ScholarCrossref
28.
Saunders  JB, Aasland  OG, Babor  TF, de la Fuente  JR, Grant  M.  Development of the Alcohol Use Disorders Identification Test (AUDIT): WHO Collaborative Project on Early Detection of Persons with Harmful Alcohol Consumption—II.  Addiction. 1993;88(6):791-804.PubMedGoogle ScholarCrossref
29.
Humeniuk  R, Ali  R, Babor  TF,  et al.  Validation of the Alcohol, Smoking And Substance Involvement Screening Test (ASSIST).  Addiction. 2008;103(6):1039-1047.PubMedGoogle ScholarCrossref
30.
WHO ASSIST Working Group.  The Alcohol, Smoking and Substance Involvement Screening Test (ASSIST): development, reliability and feasibility.  Addiction. 2002;97(9):1183-1194.PubMedGoogle ScholarCrossref
31.
Chesney  MA, Ickovics  JR, Chambers  DB,  et al.  Self-reported adherence to antiretroviral medications among participants in HIV clinical trials: the AACTG adherence instruments. Patient Care Committee & Adherence Working Group of the Outcomes Committee of the Adult AIDS Clinical Trials Group (AACTG).  AIDS Care. 2000;12(3):255-266.PubMedGoogle ScholarCrossref
32.
Cholera  R, Pence  BW, Bengtson  AM,  et al.  Mind the gap: gaps in antidepressant treatment, treatment adjustments, and outcomes among patients in routine HIV care in a multisite US clinical cohort.  PLoS One. 2017;12(1):e0166435.PubMedGoogle ScholarCrossref
33.
Pence  BW, O’Donnell  JK, Gaynes  BN.  Falling through the cracks: the gaps between depression prevalence, diagnosis, treatment, and response in HIV care.  AIDS. 2012;26(5):656-658.PubMedGoogle ScholarCrossref
34.
Pence  BW, O’Donnell  JK, Gaynes  BN.  The depression treatment cascade in primary care: a public health perspective.  Curr Psychiatry Rep. 2012;14(4):328-335.PubMedGoogle ScholarCrossref
35.
Lave  JR, Frank  RG, Schulberg  HC, Kamlet  MS.  Cost-effectiveness of treatments for major depression in primary care practice.  Arch Gen Psychiatry. 1998;55(7):645-651.PubMedGoogle ScholarCrossref
36.
Mallick  R, Chen  J, Entsuah  AR, Schatzberg  AF.  Depression-free days as a summary measure of the temporal pattern of response and remission in the treatment of major depression: a comparison of venlafaxine, selective serotonin reuptake inhibitors, and placebo.  J Clin Psychiatry. 2003;64(3):321-330.PubMedGoogle ScholarCrossref
37.
Pyne  JM, Tripathi  S, Williams  DK, Fortney  J.  Depression-free day to utility-weighted score: is it valid?  Med Care. 2007;45(4):357-362.PubMedGoogle ScholarCrossref
38.
Vannoy  SD, Arean  P, Unützer  J.  Advantages of using estimated depression-free days for evaluating treatment efficacy.  Psychiatr Serv. 2010;61(2):160-163.PubMedGoogle ScholarCrossref
39.
Cole  SR, Napravnik  S, Mugavero  MJ, Lau  B, Eron  JJ  Jr, Saag  MS.  Copy-years viremia as a measure of cumulative human immunodeficiency virus viral burden.  Am J Epidemiol. 2010;171(2):198-205.PubMedGoogle ScholarCrossref
40.
Mugavero  MJ, Amico  KR, Horn  T, Thompson  MA.  The state of engagement in HIV care in the United States: from cascade to continuum to control.  Clin Infect Dis. 2013;57(8):1164-1171.PubMedGoogle ScholarCrossref
41.
Mugavero  MJ, Lin  HY, Willig  JH,  et al.  Missed visits and mortality among patients establishing initial outpatient HIV treatment.  Clin Infect Dis. 2009;48(2):248-256.PubMedGoogle ScholarCrossref
42.
Hernán  MA, McAdams  M, McGrath  N, Lanoy  E, Costagliola  D.  Observation plans in longitudinal studies with time-varying treatments.  Stat Methods Med Res. 2009;18(1):27-52.PubMedGoogle ScholarCrossref
43.
Cole  SR, Hernán  MA, Robins  JM,  et al.  Effect of highly active antiretroviral therapy on time to acquired immunodeficiency syndrome or death using marginal structural models.  Am J Epidemiol. 2003;158(7):687-694.PubMedGoogle ScholarCrossref
44.
Baron  RM, Kenny  DA.  The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations.  J Pers Soc Psychol. 1986;51(6):1173-1182.PubMedGoogle ScholarCrossref
45.
Pence  BW, Mugavero  MJ, Carter  TJ,  et al.  Childhood trauma and health outcomes in HIV-infected patients: an exploration of causal pathways.  J Acquir Immune Defic Syndr. 2012;59(4):409-416.PubMedGoogle ScholarCrossref
46.
Panel on Antiretroviral Guidelines for Adults and Adolescents, Department of Health and Human Services. Guidelines for the use of antiretroviral agents in adults and adolescents living with HIV. https://aidsinfo.nih.gov/contentfiles/lvguidelines/adultandadolescentgl.pdf. Updated October 17, 2017. Accessed December 11, 2017.
47.
Aberg  JA, Gallant  JE, Ghanem  KG, Emmanuel  P, Zingman  BS, Horberg  MA; Infectious Diseases Society of America.  Primary care guidelines for the management of persons infected with HIV: 2013 update by the HIV medicine association of the Infectious Diseases Society of America.  Clin Infect Dis. 2014;58(1):e1-e34.PubMedGoogle ScholarCrossref
48.
Adams  JL, Gaynes  BN, McGuinness  T, Modi  R, Willig  J, Pence  BW.  Treating depression within the HIV “medical home”: a guided algorithm for antidepressant management by HIV clinicians.  AIDS Patient Care STDS. 2012;26(11):647-654.PubMedGoogle ScholarCrossref
49.
Gaynes  BN, Rush  AJ, Trivedi  MH,  et al.  Primary versus specialty care outcomes for depressed outpatients managed with measurement-based care: results from STAR*D.  J Gen Intern Med. 2008;23(5):551-560.PubMedGoogle ScholarCrossref
50.
Kaplan  RC, Kingsley  LA, Sharrett  AR,  et al.  Ten-year predicted coronary heart disease risk in HIV-infected men and women.  Clin Infect Dis. 2007;45(8):1074-1081.PubMedGoogle ScholarCrossref
51.
White  JR, Chang  CC, So-Armah  KA,  et al.  Depression and human immunodeficiency virus infection are risk factors for incident heart failure among veterans: Veterans Aging Cohort Study.  Circulation. 2015;132(17):1630-1638.PubMedGoogle ScholarCrossref
52.
Khambaty  T, Stewart  JC, Gupta  SK,  et al.  Association between depressive disorders and incident acute myocardial infarction in human immunodeficiency virus–infected adults: Veterans Aging Cohort Study.  JAMA Cardiol. 2016;1(8):929-937.PubMedGoogle ScholarCrossref
53.
Evans  DL, Ten Have  TR, Douglas  SD,  et al.  Association of depression with viral load, CD8 T lymphocytes, and natural killer cells in women with HIV infection.  Am J Psychiatry. 2002;159(10):1752-1759.PubMedGoogle ScholarCrossref
54.
Gaynes  BN, O’Donnell  J, Nelson  E,  et al.  Psychiatric comorbidity in depressed HIV-infected individuals: common and clinically consequential.  Gen Hosp Psychiatry. 2015;37(4):277-282.PubMedGoogle ScholarCrossref
55.
Gaynes  BN, Pence  BW, Eron  JJ  Jr, Miller  WC.  Prevalence and comorbidity of psychiatric diagnoses based on reference standard in an HIV+ patient population.  Psychosom Med. 2008;70(4):505-511.PubMedGoogle ScholarCrossref
56.
Israelski  DM, Prentiss  DE, Lubega  S,  et al.  Psychiatric co-morbidity in vulnerable populations receiving primary care for HIV/AIDS.  AIDS Care. 2007;19(2):220-225.PubMedGoogle ScholarCrossref
57.
Sowa  NA, Bengtson  A, Gaynes  BN, Pence  BW.  Predictors of depression recovery in HIV-infected individuals managed through measurement-based care in infectious disease clinics.  J Affect Disord. 2016;192:153-161.PubMedGoogle ScholarCrossref
Original Investigation
April 2018

Association of Increased Chronicity of Depression With HIV Appointment Attendance, Treatment Failure, and Mortality Among HIV-Infected Adults in the United States

Author Affiliations
  • 1Department of Epidemiology, Gillings School of Global Public Health, The University of North Carolina at Chapel Hill
  • 2Department of Epidemiology, Brown University, Providence, Rhode Island
  • 3Department of Psychiatry, School of Medicine, the University of North Carolina at Chapel Hill
  • 4Department of Epidemiology, University of Florida, Gainesville
  • 5Department of Medicine, University of Florida, Gainesville
  • 6Department of Medicine, School of Medicine, Johns Hopkins University, Baltimore, Maryland
  • 7Department of Psychiatry, School of Medicine, University of California, San Diego
  • 8The Fenway Institute, Fenway Community Health, Boston, Massachusetts
  • 9Department of Psychiatry, Harvard Medical School, Boston, Massachusetts
  • 10Department of Psychiatry, Massachusetts General Hospital, Boston
  • 11Department of Medicine, University of Alabama at Birmingham
  • 12UAB Center for AIDS Research, University of Alabama at Birmingham
JAMA Psychiatry. 2018;75(4):379-385. doi:10.1001/jamapsychiatry.2017.4726
Key Points

Question  For adults living with HIV, is the percentage of time spent with depression associated with appointment attendance, treatment failure, and mortality?

Finding  In this large, multisite clinical cohort study, a greater time spent with depression was associated in a dose-response fashion with higher risk of missing appointments for HIV primary care, higher risk of detectable viral load, and higher mortality rates.

Meaning  Although entirely eliminating depression is impractical, even shortening the duration of depressive episodes (eg, through integration of improved screening and evidence-based depression treatment into HIV care) may have important HIV-related benefits.

Abstract

Importance  Depression commonly affects adults with HIV and complicates the management of HIV. Depression among individuals with HIV tends to be chronic and cyclical, but the association of this chronicity with HIV outcomes (and the related potential for screening and intervention to shorten depressive episodes) has received little attention.

Objective  To examine the association between increased chronicity of depression and multiple HIV care continuum indicators (HIV appointment attendance, treatment failure, and mortality).

Design, Setting, and Participants  The study comprised an observational clinical cohort of 5927 patients with 2 or more assessments of depressive severity who were receiving HIV primary care at 6 geographically dispersed US academic medical centers from September 22, 2005, to August 6, 2015.

Main Outcomes and Measures  Missing a scheduled HIV primary care visit, detectable HIV RNA viral load (≥75 copies/mL), and all-cause mortality. Consecutive depressive severity measures were converted into a time-updated measure: percentage of days with depression (PDD), following established methods for determining depression-free days.

Results  During 10 767 person-years of follow-up, the 5927 participants (5000 men, 926 women, and 1 intersex individual; median age, 44 years [range, 35-50 years]) had a median PDD of 14% (interquartile range, 0%-48%). During follow-up, 10 361 of 55 040 scheduled visits (18.8%) were missed, 6191 of 28 455 viral loads (21.8%) were detectable, and the mortality rate was 1.5 deaths per 100 person-years. Percentage of days with depression showed a dose-response relationship with each outcome. Each 25% increase in PDD led to an 8% increase in the risk of missing a scheduled appointment (risk ratio, 1.08; 95% CI, 1.05-1.11), a 5% increase in the risk of a detectable viral load (risk ratio, 1.05; 95% CI, 1.01-1.09), and a 19% increase in the mortality hazard (hazard ratio, 1.19; 95% CI, 1.05-1.36). These estimates imply that, compared with patients who spent no follow-up time with depression (PDD, 0%), those who spent the entire follow-up time with depression (PDD, 100%) faced a 37% increased risk of missing appointments (risk ratio, 1.37; 95% CI, 1.22-1.53), a 23% increased risk of a detectable viral load (risk ratio, 1.23; 95% CI, 1.06-1.43), and a doubled mortality rate (hazard ratio, 2.02; 95% CI, 1.20-3.42).

Conclusions and Relevance  Greater chronicity of depression increased the likelihood of failure at multiple points along the HIV care continuum. Even modest increases in the proportion of time spent with depression led to clinically meaningful increases in negative outcomes. Clinic-level trials of protocols to promptly identify and appropriately treat depression among adults living with HIV should be conducted to understand the effect of such protocols on shortening the course and preventing the recurrence of depressive illness and improving clinical outcomes.

Introduction

Depression is one of the most common comorbidities affecting adults living with HIV, with prevalence estimates ranging from 20% to 40%.1-7 In addition to being a serious clinical concern, depression complicates the management of HIV: it has been linked to low adherence to antiretroviral therapy (ART),8 poor attendance at HIV primary care appointments,9 ART failure,10 accelerated clinical progression of HIV,11,12 and higher mortality rates.13-24

Much of the literature associating depression with HIV outcomes has focused on categorizing scores on depression severity scales such as the Center for Epidemiologic Studies–Depression or the Patient Health Questionnaire-9 (PHQ-9)25 into “depressed” or “not depressed” at a particular point in time based on standard cutpoints. This dichotomous approach fails to capture the variation in severity and chronicity of depressive episodes and the cumulative burden of exposure to depression. The perspective of the cumulative burden has important implications for quantifying the potential benefit of efforts to shorten (rather than universally prevent) depressive episodes through evidence-based treatment. In addition, much of the current literature examines depression in association with a single HIV-related outcome rather than considering the association of depression with the continuum from engagement in HIV care, through response to ART treatment, to arguably the most important outcome, survival.

This article adapts a measure commonly used in the depression treatment trial field, depression-free days, to capture the cumulative burden of depression over time among a large, multisite cohort of adults receiving HIV primary care in the United States, and determines whether a greater cumulative burden of depression is associated with HIV care engagement, treatment response, and mortality.

Methods
Study Population

Participants came from the Center for AIDS Research Network of Integrated Clinical Systems (CNICS),26 a collaboration of 8 geographically dispersed academic medical centers in the United States, each of which has established a database capturing clinical information on their patients receiving HIV primary care. Sites upload data quarterly to a central CNICS repository, where quality checks are performed. Sites capture data on basic demographics, appointment attendance, diagnoses, medications, and laboratory test results. Race and ethnicity are captured as recorded in each site’s medical record system. Most CNICS sites have integrated patient-reported outcomes (PROs) into the routine clinical workflow. Patient-reported outcomes are targeted for completion at routine clinical visits approximately every 6 months and include measures of depressive symptoms (PHQ-9), panic symptoms (PHQ-5),27 alcohol use (Alcohol Use Disorders Identification Test–Clinical),28 substance use (The Alcohol, Smoking and Substance Involvement Screening Test),29,30 and ART adherence (AIDS Clinical Trials Group assessment31). Deaths are ascertained regularly through site reports and queries of the National Death Index. Patients provide written informed consent for their information to be collected; participation rates are greater than 90% at all sites. Data collection is approved by the institutional review boards at each CNICS site (Case Western Reserve University; Fenway Community Health Center; Johns Hopkins University; the University of Alabama at Birmingham; the University of California at San Diego; the University of California at San Francisco; the University of North Carolina at Chapel Hill; and the University of Washington); this analysis was approved by the University of North Carolina–Chapel Hill institutional review board.

This analysis included all CNICS participants between September 22, 2005, and August 6, 2015, with 2 or more consecutive PHQ-9 measures, defined as 2 PHQ-9 measures separated by less than 365 days. Two sites without PHQ-9 measures available in the CNICS database during this time period (Case Western Reserve University and Johns Hopkins University) were not represented in the analysis dataset. The median gap between PHQ-9 assessments in the final sample was 182 days (interquartile range, 126-226 days). For this study, patients entered the analysis on the date of the second such PHQ-9 assessment and remained in the analysis until the earliest of the following: death, censoring owing to a lapse in PHQ-9 measures (12 months after the last consecutive PHQ-9 measure), censoring owing to loss to care (≥12 months without an HIV primary care appointment), 6 years’ follow-up, or administrative censoring. Administrative censoring was defined based on each site’s most recent upload at the time of the data query for this analysis and ranged from October 1, 2014, to October 1, 2015, depending on the site.

The standard of care for depression screening and treatment varies across CNICS sites. Some sites provide PHQ-9 results back to the medical or social work team, in real time or with a delay, while other sites treat the PHQ-9 results as research data. Further assessment, referral, and treatment plans are at the discretion of the clinician and specific to local resources available. Prior research in CNICS has documented large gaps in the initiation of treatment for depression, guideline-concordant adjustments to the depression treatment plan, and achievement of remission across all sites32,33; these gaps are similar to those identified in primary care and other medical settings.34

Measures: Cumulative Burden of Depression

Cumulative burden of depression was measured using an adaptation of depression-free days, a metric commonly used in trials of depression treatment to quantify the relative benefits of different treatments in reducing the amount of time spent with depression during follow-up.35-38 In the depression-free days approach, consecutive scores on a standardized depressive severity scale, such as the PHQ-9, are each converted into a scale from 0 to 1, with 1 corresponding to a region below a threshold indicating full remission of depression (fully depression free; eg, PHQ-9 score <5), 0 corresponding to a region above a second threshold indicating fully symptomatic depression (fully not depression free; eg, PHQ-9 score ≥15), and intermediate scores being assigned a prorated value between 0 and 1. The values at the beginning and end of each time interval are averaged to indicate the proportion of the time interval spent depression free, and can be multiplied by the length of the interval to indicate the number of days in the interval spent depression free. To characterize cumulative exposure to a condition, depression-free days can be thought of as analogous to other cumulative measures, such as pack-years of smoking or viremia copy-years.39

Depression-free days correlate well with other measures of response to treatment of depression.38 The measure of depression-free days is commonly calculated from depression assessments spaced at 3- to 6-month intervals.35-38 Prior validation research has indicated that 6-month spacing, as used in our analysis, functions well, and little additional information is gained by conducting assessments more frequently than every 6 months.38

As we sought in this analysis to examine the negative outcomes of a longer duration of depression rather than the benefits of depression treatments, we calculated days with depression, the converse of depression-free days, by setting PHQ-9 scores of 15 or more to a value of 1 (fully depressed), PHQ-9 scores less than 5 to a value of 0 (fully not depressed), and PHQ-9 scores between 5 and 14 to a prorated value between 0 and 1. We then multiplied the mean of each pair of consecutive measures by the number of days between those measures to calculate days with depression in that interval. We aggregated days with depression over time and divided it by total days of follow-up through a given time point to obtain a time-updated measure of the mean cumulative burden of depression during follow-up, referred to as percentage of days with depression (PDD).

Measures: Outcomes

We considered the following 3 outcomes capturing 3 key markers of the continuum of HIV care engagement and management40: missed HIV primary care appointments, detectable HIV viral load, and all-cause mortality. Missed HIV primary care appointments were measured from administrative appointment data. This outcome was defined as whether a particular scheduled appointment was kept or missed. All scheduled appointments during each participant’s follow-up were included, so that participants could have multiple scheduled appointments considered in the analysis. Consistent with related literature, scheduled appointments that were canceled by the patient or rescheduled by the clinic were excluded.41

Detectable HIV viral load was defined as a HIV RNA viral load measure of 75 copies/mL or more based on the highest limit of detection among the viral load assays used during the time period covered by this analysis. All viral loads during each participant’s follow-up were included. Mortality was defined as date of death from any cause.

Measures: Covariates

Time-fixed covariates included baseline site, age, sex, race/ethnicity (white non-Hispanic, black non-Hispanic, Hispanic, or other), and diagnoses in the medical record of cardiovascular disease, type 1 or type 2 diabetes, depression, anxiety, and other mental health disorders. Time-varying covariates included self-reported heavy use of alcohol (Alcohol Use Disorders Identification Test–Clinical score ≥4 for men and ≥3 for women), self-reported drug use (any past or current use of illegal drugs other than marijuana), ART use and adherence (not using ART, using ART and no self-reported missed doses in the past week, or using ART and ≥1 self-reported missed dose in the past week), and laboratory test reports of HIV viral load and CD4 T-cell count.

Statistical Analysis

The goal of the analysis was to estimate the effect of increased PDD on the risk of missed appointments, the risk of detectable viral load, and the rate of all-cause mortality, using causal inference methods to estimate effects from observational data.42,43 Time-varying covariates such as ART use and adherence, viral load, CD4 T-cell count, and alcohol and drug use may both confound and mediate the effect of depression burden with outcomes. As such, adjustment for these covariates through traditional multivariable regression models would be expected to bias effect estimates.42,43 Therefore, we used inverse probability of exposure weights to fit marginal structural regression models.43 We combined the inverse probability of exposure weights with inverse probability of censoring weights to address informative censoring and inverse probability of visit weights to address differing frequencies of measurement of appointments and viral loads.42 Detail of the construction of the weights is provided in the eAppendix in the Supplement.

We used the combined weights to fit a weighted, marginal structural Cox proportional hazards regression model to estimate the effect of PDD with mortality, and to fit weighted pooled repeated-measures marginal structural Poisson regression models with robust variance to estimate the effect of PDD with the risk of missed appointments and detectable viral load. We assessed the assumption of a linear association between PDD and outcomes by comparing linear, quadratic, and restricted cubic spline specifications, both statistically (by comparing the Akaike information criterion) and graphically. For the Cox proportional hazards regression model, we assessed the validity of the proportional hazards assumption using Schoenfeld residuals, which indicated that the proportional hazards assumption was upheld.

In an exploratory analysis, we examined whether time-updated missed appointments mediated the effect of PDD on viral load, and whether time-updated missed appointments and viral load mediated the effect of PDD on mortality by comparing point estimates in our primary models with point estimates after adjustment for the hypothesized mediator. If the hypothesized variables act as partial or full mediators, we would expect the point estimates to shift toward or to the null on adjustment.44,45

Results

A total of 5927 participants contributed 10 767 person-years to the analysis (Table 1). Participants were a median age of 44 years (interquartile range, 35-50 years) and most were male (5000 [84.4%]). Of the 5874 participants with data on race/ethnicity, approximately half were white non-Hispanic (2918 [49.7%]), 1914 were black non-Hispanic (32.6%), 823 were Hispanic (14.0%), and 219 were other races/ethnicities (3.7%). At entry into the analysis, most participants had CD4 T-cell counts of 350 cells/mm3 or more (4111 [69.4%]), were receiving ART (4440 of 5736 [77.4%]), and had a suppressed viral load (3759 [63.4%]).

During follow-up, participants had a median percentage of days with depression (PDD) of 14% (interquartile range, 0%-48%). Approximately one-third of participants (1880 [31.7%]) had 0% PDD whereas 214 (3.6%) had 100% PDD. During the follow-up period, 158 deaths were observed (mortality rate, 1.5 deaths per 100 person-years). Participants had 55 040 scheduled appointments, of which 10 361 were missed (18.8%), and 28 455 viral loads, of which 6191 were detectable (21.8%) (Table 1).

In regression models, to facilitate meaningful interpretation of coefficients, PDD was scaled so that a 1-unit change corresponded with a 25–percentage point increase in PDD (eg, 25% vs 0% PDD, or 50% vs 25% PDD). A PDD value of 25% could reflect, for example, full depression for 1 quarter of follow-up or mild depressive symptoms (eg, PHQ-9 score of 7) for the entirety of follow-up.

In the weighted marginal structural models, increasing PDD was associated in a dose-response fashion with an increased risk of missed appointments, increased risk of detectable viral load, and accelerated mortality. Each 25% increase in PDD led to an 8% increase in the risk that a particular scheduled appointment would be missed (risk ratio, 1.08; 95% CI, 1.05-1.11), a 5% increase in the risk that a particular viral load would be detectable (risk ratio, 1.05; 95% CI, 1.01-1.09), and a 19% increase in the mortality hazard (hazard ratio, 1.19; 95% CI, 1.05-1.36) (Table 2). These estimates imply that, compared with those who spent no follow-up time with depression (PDD, 0%), those who spent the entire follow-up time with depression (PDD, 100%) faced a 37% increased risk of missing appointments (risk ratio, 1.37; 95% CI, 1.22-1.53), a 23% increased risk of a detectable viral load (risk ratio, 1.23; 95% CI, 1.06-1.43), and a doubled mortality rate (hazard ratio, 2.02; 95% CI, 1.20-3.42).

Assessment of the assumption of a linear, dose-response relationship between PDD and each outcome indicated that, for the risk of missed appointments and detectable viral load, the dose-response assumption was reasonable. For mortality, there was some suggestion of a threshold rather than a dose-response relationship, with all levels of PDD from 25% to 100% having approximately a doubled hazard of death compared with those with a PDD of 0% (Figure).

There was limited support for the hypothesized mediation. Time-updated missed appointments were independently associated with both detectable viral load and mortality. The addition of missed appointments to the models attenuated the estimated association of PDD with viral load by 20% and the estimated association of PDD with mortality by 25%. Time-updated viral loads were also independently associated with mortality, but their addition to the mortality model did not further change the estimated association of PDD with mortality.

Discussion

In this large, multisite cohort of patients receiving HIV primary care in the United States, we found that a greater proportion of time spent with depression increased the risk of missed HIV primary care appointments and lack of viral suppression in a dose-response manner. Mortality rates also increased with greater depression burden, with a suggestion of a threshold association such that even modest increases in time spent with depression, amounting to a mean of 1 in 4 days during follow-up, approximately doubled the mortality rate.

These findings, which suggest that systematic screening and enhanced treatment to shorten the duration of depressive illness may have multiple benefits, stand in contrast with the trend in HIV treatment guidelines toward less frequent monitoring of patients whose disease is stable with treatment.46,47 Protocolized, time-efficient strategies to enhance depression treatment in nonpsychiatric settings, such as measurement-based care,48 have been shown to be as effective as psychiatric care in achieving remission of depression.49 However, these protocols require several months of short-term follow-up to ensure that treatment is adjusted until remission is reached. Integration of strategies such as measurement-based care would require more intensive follow-up of some patients, potentially through task-shifting strategies involving social workers or other clinic personnel as depression care managers, in an effort to maximize depression-free days.48

These results are consistent with those in prior literature that has reported associations between depression and HIV outcomes.10,13-24 Our findings are an important extension of prior work in several ways. First, we used a unified analytical approach to consider the association of depression with 3 key indicators of the HIV treatment continuum.40 Second, we used a novel adaptation of a metric commonly used in trials of treatment for depression to capture the varying levels of chronicity and severity of depressive illness over time, rather than relying on a binary classification of depressed or not depressed. Our results indicate that even short-term or mild depression, such as 1 in 4 days spent fully depressed or persistent mild depressive symptoms, can have meaningful negative outcomes on HIV treatment and survival. Third, we applied advanced causal inference statistical methods to comprehensively address time-varying confounding, informative censoring, and varying frequency of assessments, strengthening the interpretation of our results. Finally, we leveraged a large, demographically diverse, and geographically dispersed cohort of patients to generate precise effect estimates that are likely applicable to many patients receiving HIV primary care in the United States.

The mechanisms through which depression may influence care engagement, treatment response, and mortality remain a matter of some debate. Depression has been strongly and consistently linked to reduced adherence to ART,8 which can lead to ART-resistant virus, treatment failure including elevated viral load, and increased mortality. Depression is also one of the strongest risk factors for suicide attempts and completions. Depression negatively affects the incidence and management of other chronic diseases such as cardiovascular disease and diabetes, which are increasingly common causes of death in populations with HIV.24,50-52 Furthermore, depression can have a direct suppressing effect on the immune system.53 Finally, depression commonly co-occurs with other psychiatric conditions such as posttraumatic stress disorder, generalized anxiety disorder, and alcohol and substance misuse. These comorbidities are themselves recognized as barriers to care engagement and markers of increased risk of mortality.1,5,6,54-57 Although our analysis controlled for psychiatric comorbidities using appropriate methods for time-varying confounding, the association between depression and other mental health conditions is likely complex, bidirectional, and difficult to fully disentangle statistically. Our exploratory mediation analysis provided partial support to appointment adherence as a mediator of the association between depression, viral load, and mortality, but indicated that other pathways are likely to account for the majority of the association.

Limitations

We quantified cumulative burden of depression by converting repeated PHQ-9 measures into an estimate of PDD during follow-up. Although this approach likely has some measurement error relative to repeated clinical diagnostic evaluations, the method has been used for nearly 20 years in the literature on trials of depression treatment to compare the relative benefits of different interventions.35-38 The measure correlates well with other measures of depression burden35-38 and is relatively insensitive to the spacing between PHQ-9 assessments in the range seen in this sample.38

Our analysis included only individuals who had consented to participate in CNICS and had completed at least 2 PHQ-9s. Participation in CNICS is >90% at all sites, and the demographic and clinical characteristics of our sample were very similar to those of the overall CNICS cohort, suggesting that selection bias was unlikely to have had a major influence on our results. Although the CNICS cohort represents a diverse and geographically dispersed network of academic medical centers, these results may be less generalizable to smaller or nonspecialist care settings.

Conclusions

In this large sample of patients engaged in HIV primary care in the United States, greater chronicity of depression elevated the risk of missed appointments and detectable viral load as well as mortality rates. Even modest increases in the proportion of time spent with depression led to clinically meaningful increases in negative outcomes. These findings suggest the importance of promptly identifying and treating depression among adults living with HIV to shorten the course and prevent the return of their depressive illness and ultimately improve their clinical outcomes. These findings further suggest the need for clinic-randomized, implementation science–focused trials of the mental health and HIV-related benefits of integration of systematic screening and enhanced treatment protocols into routine HIV care.

Back to top
Article Information

Accepted for Publication: December 26, 2017.

Correction: This article was corrected on April 4, 2018, to fix an error in the short title.

Corresponding Author: Brian W. Pence, PhD, Department of Epidemiology, Gillings School of Global Public Health, The University of North Carolina at Chapel Hill, 2103C McGavran-Greenberg Hall, Chapel Hill, NC 27599 (bpence@unc.edu).

Published Online: February 21, 2018. doi:10.1001/jamapsychiatry.2017.4726

Author Contributions: Dr Pence had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Pence, Mills, Breger, Mugavero.

Acquisition, analysis, or interpretation of data: Pence, Mills, Bengtson, Gaynes, Cook, Moore, Grelotti, O’Cleirigh, Mugavero.

Drafting of the manuscript: Pence, Mills.

Critical revision of the manuscript for important intellectual content: Mills, Bengtson, Gaynes, Breger, Cook, Moore, Grelotti, O’Cleirigh, Mugavero.

Statistical analysis: Pence.

Obtained funding: Pence, Mugavero.

Administrative, technical, or material support: Bengtson, Gaynes, Breger, Moore, Grelotti, Mugavero.

Study supervision: Gaynes.

Conflict of Interest Disclosures: None reported.

Funding/Support: This work was supported by grant R01MH100970 from the National Institutes of Mental Health (Dr Pence) and by grants R24AI067039 and P30AI50410 from the National Institute of Allergy and Infectious Disease.

Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

References
1.
Pence  BW, Miller  WC, Whetten  K, Eron  JJ, Gaynes  BN.  Prevalence of DSM-IV-defined mood, anxiety, and substance use disorders in an HIV clinic in the Southeastern United States.  J Acquir Immune Defic Syndr. 2006;42(3):298-306.PubMedGoogle ScholarCrossref
2.
Bing  EG, Burnam  MA, Longshore  D,  et al.  Psychiatric disorders and drug use among human immunodeficiency virus-infected adults in the United States.  Arch Gen Psychiatry. 2001;58(8):721-728.PubMedGoogle ScholarCrossref
3.
Ciesla  JA, Roberts  JE.  Meta-analysis of the relationship between HIV infection and risk for depressive disorders.  Am J Psychiatry. 2001;158(5):725-730.PubMedGoogle ScholarCrossref
4.
Orlando  M, Burnam  MA, Beckman  R,  et al.  Re-estimating the prevalence of psychiatric disorders in a nationally representative sample of persons receiving care for HIV: results from the HIV Cost and Services Utilization Study.  Int J Methods Psychiatr Res. 2002;11(2):75-82.PubMedGoogle ScholarCrossref
5.
Schumacher  JE, McCullumsmith  C, Mugavero  MJ,  et al.  Routine depression screening in an HIV clinic cohort identifies patients with complex psychiatric co-morbidities who show significant response to treatment.  AIDS Behav. 2013;17(8):2781-2791.PubMedGoogle ScholarCrossref
6.
Tegger  MK, Crane  HM, Tapia  KA, Uldall  KK, Holte  SE, Kitahata  MM.  The effect of mental illness, substance use, and treatment for depression on the initiation of highly active antiretroviral therapy among HIV-infected individuals.  AIDS Patient Care STDS. 2008;22(3):233-243.PubMedGoogle ScholarCrossref
7.
O’Cleirigh  C, Magidson  JF, Skeer  MR, Mayer  KH, Safren  SA.  Prevalence of psychiatric and substance abuse symptomatology among HIV-infected gay and bisexual men in HIV primary care.  Psychosomatics. 2015;56(5):470-478.PubMedGoogle ScholarCrossref
8.
DiMatteo  MR, Lepper  HS, Croghan  TW.  Depression is a risk factor for noncompliance with medical treatment: meta-analysis of the effects of anxiety and depression on patient adherence.  Arch Intern Med. 2000;160(14):2101-2107.PubMedGoogle ScholarCrossref
9.
Zuniga  JA, Yoo-Jeong  M, Dai  T, Guo  Y, Waldrop-Valverde  D.  The role of depression in retention in care for persons living with HIV.  AIDS Patient Care STDS. 2016;30(1):34-38.PubMedGoogle ScholarCrossref
10.
Pence  BW, Miller  WC, Gaynes  BN, Eron  JJ  Jr.  Psychiatric illness and virologic response in patients initiating highly active antiretroviral therapy.  J Acquir Immune Defic Syndr. 2007;44(2):159-166.PubMedGoogle ScholarCrossref
11.
Mugavero  MJ, Pence  BW, Whetten  K,  et al.  Predictors of AIDS-related morbidity and mortality in a southern US cohort.  AIDS Patient Care STDS. 2007;21(9):681-690.PubMedGoogle ScholarCrossref
12.
Ironson  G, O’Cleirigh  C, Fletcher  MA,  et al.  Psychosocial factors predict CD4 and viral load change in men and women with human immunodeficiency virus in the era of highly active antiretroviral treatment.  Psychosom Med. 2005;67(6):1013-1021.PubMedGoogle ScholarCrossref
13.
Todd  JV, Cole  SR, Pence  BW,  et al.  Effects of antiretroviral therapy and depressive symptoms on all-cause mortality among HIV-infected women.  Am J Epidemiol. 2017;185(10):869-878.PubMedGoogle ScholarCrossref
14.
Ickovics  JR, Hamburger  ME, Vlahov  D,  et al; HIV Epidemiology Research Study Group.  Mortality, CD4 cell count decline, and depressive symptoms among HIV-seropositive women: longitudinal analysis from the HIV Epidemiology Research Study.  JAMA. 2001;285(11):1466-1474.PubMedGoogle ScholarCrossref
15.
Anastos  K, Schneider  MF, Gange  SJ,  et al; Women’s Interagency HIV Study Collaborative Group.  The association of race, sociodemographic, and behavioral characteristics with response to highly active antiretroviral therapy in women.  J Acquir Immune Defic Syndr. 2005;39(5):537-544.PubMedGoogle Scholar
16.
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(7):1133-1140.PubMedGoogle ScholarCrossref
17.
French  AL, Gawel  SH, Hershow  R,  et al.  Trends in mortality and causes of death among women with HIV in the United States: a 10-year study.  J Acquir Immune Defic Syndr. 2009;51(4):399-406.PubMedGoogle ScholarCrossref
18.
Hessol  NA, Kalinowski  A, Benning  L,  et al.  Mortality among participants in the Multicenter AIDS Cohort Study and the Women’s Interagency HIV Study.  Clin Infect Dis. 2007;44(2):287-294.PubMedGoogle ScholarCrossref
19.
Leserman  J, Pence  BW, Whetten  K,  et al.  Relation of lifetime trauma and depressive symptoms to mortality in HIV.  Am J Psychiatry. 2007;164(11):1707-1713.PubMedGoogle ScholarCrossref
20.
Lima  VD, Geller  J, Bangsberg  DR,  et al.  The effect of adherence on the association between depressive symptoms and mortality among HIV-infected individuals first initiating HAART.  AIDS. 2007;21(9):1175-1183.PubMedGoogle ScholarCrossref
21.
Murphy  K, Hoover  DR, Shi  Q,  et al.  Association of self-reported race with AIDS death in continuous HAART users in a cohort of HIV-infected women in the United States.  AIDS. 2013;27(15):2413-2423.PubMedGoogle ScholarCrossref
22.
Wada  N, Jacobson  LP, Cohen  M, French  A, Phair  J, Muñoz  A.  Cause-specific life expectancies after 35 years of age for human immunodeficiency syndrome–infected and human immunodeficiency syndrome–negative individuals followed simultaneously in long-term cohort studies, 1984-2008.  Am J Epidemiol. 2013;177(2):116-125.PubMedGoogle ScholarCrossref
23.
Cohen  MH, Hotton  AL, Hershow  RC,  et al.  Gender-related risk factors improve mortality predictive ability of VACS index among HIV-infected women.  J Acquir Immune Defic Syndr. 2015;70(5):538-544.PubMedGoogle ScholarCrossref
24.
Cohen  MH, French  AL, Benning  L,  et al.  Causes of death among women with human immunodeficiency virus infection in the era of combination antiretroviral therapy.  Am J Med. 2002;113(2):91-98.PubMedGoogle ScholarCrossref
25.
Kroenke  K, Spitzer  RL.  The PHQ-9: a new depression diagnostic and severity measure.  Psychiatr Ann. 2002;32(9):509-515. doi:10.3928/0048-5713-20020901-06Google ScholarCrossref
26.
Kitahata  MM, Rodriguez  B, Haubrich  R,  et al.  Cohort profile: the Centers for AIDS Research Network of Integrated Clinical Systems.  Int J Epidemiol. 2008;37(5):948-955.PubMedGoogle ScholarCrossref
27.
Spitzer  RL, Kroenke  K, Williams  JB.  Validation and utility of a self-report version of PRIME-MD: the PHQ primary care study: Primary Care Evaluation of Mental Disorders: Patient Health Questionnaire.  JAMA. 1999;282(18):1737-1744.PubMedGoogle ScholarCrossref
28.
Saunders  JB, Aasland  OG, Babor  TF, de la Fuente  JR, Grant  M.  Development of the Alcohol Use Disorders Identification Test (AUDIT): WHO Collaborative Project on Early Detection of Persons with Harmful Alcohol Consumption—II.  Addiction. 1993;88(6):791-804.PubMedGoogle ScholarCrossref
29.
Humeniuk  R, Ali  R, Babor  TF,  et al.  Validation of the Alcohol, Smoking And Substance Involvement Screening Test (ASSIST).  Addiction. 2008;103(6):1039-1047.PubMedGoogle ScholarCrossref
30.
WHO ASSIST Working Group.  The Alcohol, Smoking and Substance Involvement Screening Test (ASSIST): development, reliability and feasibility.  Addiction. 2002;97(9):1183-1194.PubMedGoogle ScholarCrossref
31.
Chesney  MA, Ickovics  JR, Chambers  DB,  et al.  Self-reported adherence to antiretroviral medications among participants in HIV clinical trials: the AACTG adherence instruments. Patient Care Committee & Adherence Working Group of the Outcomes Committee of the Adult AIDS Clinical Trials Group (AACTG).  AIDS Care. 2000;12(3):255-266.PubMedGoogle ScholarCrossref
32.
Cholera  R, Pence  BW, Bengtson  AM,  et al.  Mind the gap: gaps in antidepressant treatment, treatment adjustments, and outcomes among patients in routine HIV care in a multisite US clinical cohort.  PLoS One. 2017;12(1):e0166435.PubMedGoogle ScholarCrossref
33.
Pence  BW, O’Donnell  JK, Gaynes  BN.  Falling through the cracks: the gaps between depression prevalence, diagnosis, treatment, and response in HIV care.  AIDS. 2012;26(5):656-658.PubMedGoogle ScholarCrossref
34.
Pence  BW, O’Donnell  JK, Gaynes  BN.  The depression treatment cascade in primary care: a public health perspective.  Curr Psychiatry Rep. 2012;14(4):328-335.PubMedGoogle ScholarCrossref
35.
Lave  JR, Frank  RG, Schulberg  HC, Kamlet  MS.  Cost-effectiveness of treatments for major depression in primary care practice.  Arch Gen Psychiatry. 1998;55(7):645-651.PubMedGoogle ScholarCrossref
36.
Mallick  R, Chen  J, Entsuah  AR, Schatzberg  AF.  Depression-free days as a summary measure of the temporal pattern of response and remission in the treatment of major depression: a comparison of venlafaxine, selective serotonin reuptake inhibitors, and placebo.  J Clin Psychiatry. 2003;64(3):321-330.PubMedGoogle ScholarCrossref
37.
Pyne  JM, Tripathi  S, Williams  DK, Fortney  J.  Depression-free day to utility-weighted score: is it valid?  Med Care. 2007;45(4):357-362.PubMedGoogle ScholarCrossref
38.
Vannoy  SD, Arean  P, Unützer  J.  Advantages of using estimated depression-free days for evaluating treatment efficacy.  Psychiatr Serv. 2010;61(2):160-163.PubMedGoogle ScholarCrossref
39.
Cole  SR, Napravnik  S, Mugavero  MJ, Lau  B, Eron  JJ  Jr, Saag  MS.  Copy-years viremia as a measure of cumulative human immunodeficiency virus viral burden.  Am J Epidemiol. 2010;171(2):198-205.PubMedGoogle ScholarCrossref
40.
Mugavero  MJ, Amico  KR, Horn  T, Thompson  MA.  The state of engagement in HIV care in the United States: from cascade to continuum to control.  Clin Infect Dis. 2013;57(8):1164-1171.PubMedGoogle ScholarCrossref
41.
Mugavero  MJ, Lin  HY, Willig  JH,  et al.  Missed visits and mortality among patients establishing initial outpatient HIV treatment.  Clin Infect Dis. 2009;48(2):248-256.PubMedGoogle ScholarCrossref
42.
Hernán  MA, McAdams  M, McGrath  N, Lanoy  E, Costagliola  D.  Observation plans in longitudinal studies with time-varying treatments.  Stat Methods Med Res. 2009;18(1):27-52.PubMedGoogle ScholarCrossref
43.
Cole  SR, Hernán  MA, Robins  JM,  et al.  Effect of highly active antiretroviral therapy on time to acquired immunodeficiency syndrome or death using marginal structural models.  Am J Epidemiol. 2003;158(7):687-694.PubMedGoogle ScholarCrossref
44.
Baron  RM, Kenny  DA.  The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations.  J Pers Soc Psychol. 1986;51(6):1173-1182.PubMedGoogle ScholarCrossref
45.
Pence  BW, Mugavero  MJ, Carter  TJ,  et al.  Childhood trauma and health outcomes in HIV-infected patients: an exploration of causal pathways.  J Acquir Immune Defic Syndr. 2012;59(4):409-416.PubMedGoogle ScholarCrossref
46.
Panel on Antiretroviral Guidelines for Adults and Adolescents, Department of Health and Human Services. Guidelines for the use of antiretroviral agents in adults and adolescents living with HIV. https://aidsinfo.nih.gov/contentfiles/lvguidelines/adultandadolescentgl.pdf. Updated October 17, 2017. Accessed December 11, 2017.
47.
Aberg  JA, Gallant  JE, Ghanem  KG, Emmanuel  P, Zingman  BS, Horberg  MA; Infectious Diseases Society of America.  Primary care guidelines for the management of persons infected with HIV: 2013 update by the HIV medicine association of the Infectious Diseases Society of America.  Clin Infect Dis. 2014;58(1):e1-e34.PubMedGoogle ScholarCrossref
48.
Adams  JL, Gaynes  BN, McGuinness  T, Modi  R, Willig  J, Pence  BW.  Treating depression within the HIV “medical home”: a guided algorithm for antidepressant management by HIV clinicians.  AIDS Patient Care STDS. 2012;26(11):647-654.PubMedGoogle ScholarCrossref
49.
Gaynes  BN, Rush  AJ, Trivedi  MH,  et al.  Primary versus specialty care outcomes for depressed outpatients managed with measurement-based care: results from STAR*D.  J Gen Intern Med. 2008;23(5):551-560.PubMedGoogle ScholarCrossref
50.
Kaplan  RC, Kingsley  LA, Sharrett  AR,  et al.  Ten-year predicted coronary heart disease risk in HIV-infected men and women.  Clin Infect Dis. 2007;45(8):1074-1081.PubMedGoogle ScholarCrossref
51.
White  JR, Chang  CC, So-Armah  KA,  et al.  Depression and human immunodeficiency virus infection are risk factors for incident heart failure among veterans: Veterans Aging Cohort Study.  Circulation. 2015;132(17):1630-1638.PubMedGoogle ScholarCrossref
52.
Khambaty  T, Stewart  JC, Gupta  SK,  et al.  Association between depressive disorders and incident acute myocardial infarction in human immunodeficiency virus–infected adults: Veterans Aging Cohort Study.  JAMA Cardiol. 2016;1(8):929-937.PubMedGoogle ScholarCrossref
53.
Evans  DL, Ten Have  TR, Douglas  SD,  et al.  Association of depression with viral load, CD8 T lymphocytes, and natural killer cells in women with HIV infection.  Am J Psychiatry. 2002;159(10):1752-1759.PubMedGoogle ScholarCrossref
54.
Gaynes  BN, O’Donnell  J, Nelson  E,  et al.  Psychiatric comorbidity in depressed HIV-infected individuals: common and clinically consequential.  Gen Hosp Psychiatry. 2015;37(4):277-282.PubMedGoogle ScholarCrossref
55.
Gaynes  BN, Pence  BW, Eron  JJ  Jr, Miller  WC.  Prevalence and comorbidity of psychiatric diagnoses based on reference standard in an HIV+ patient population.  Psychosom Med. 2008;70(4):505-511.PubMedGoogle ScholarCrossref
56.
Israelski  DM, Prentiss  DE, Lubega  S,  et al.  Psychiatric co-morbidity in vulnerable populations receiving primary care for HIV/AIDS.  AIDS Care. 2007;19(2):220-225.PubMedGoogle ScholarCrossref
57.
Sowa  NA, Bengtson  A, Gaynes  BN, Pence  BW.  Predictors of depression recovery in HIV-infected individuals managed through measurement-based care in infectious disease clinics.  J Affect Disord. 2016;192:153-161.PubMedGoogle ScholarCrossref
×