Abstract
Objectives
To characterize antiretroviral therapy (ART) adherence across different programmes and examine the relationship between individual and programme characteristics and incomplete adherence among ART clients in sub-Saharan Africa.
Design
A cross-sectional study.
Methods
Systematically selected ART clients (≥18 years; on ART ≥6 months) attending 18 facilities in three countries (250 clients/facility) were interviewed. Client self-reports (3-day, 30-day, Case Index ≥48 consecutive hours of missed ART), healthcare provider estimates and the pharmacy medication possession ratio (MPR) were used to estimate ART adherence. Participants from two facilities per country underwent HIV RNA testing. Optimal adherence measures were selected on the basis of degree of association with concurrent HIV RNA dichotomized at less than or greater/equal to 1000 copies/ml. Multivariate regression analysis, adjusted for site-level clustering, assessed associations between incomplete adherence and individual and programme factors.
Results
A total of 4489 participants were included, of whom 1498 underwent HIV RNA testing. Nonadherence ranged from 3.2% missing at least 48 consecutive hours to 40.1% having an MPR of less than 90%. The percentage with HIV RNA at least 1000 copies/ml ranged from 7.2 to 17.2% across study sites (mean = 9.9%). Having at least 48 consecutive hours of missed ART was the adherence measure most strongly related to virologic failure. Factors significantly related to incomplete adherence included visiting a traditional healer, screening positive for alcohol abuse, experiencing more HIV symptoms, having an ART regimen without nevirapine and greater levels of internalized stigma.
Conclusion
Results support more in-depth investigations of the role of traditional healers, and the development of interventions to address alcohol abuse and internalized stigma among treatment-experienced adult ART patients.
Keywords: antiretroviral therapy, HIV/AIDS, medication adherence, sub-Saharan Africa, Tanzania, Uganda, Zambia
Introduction
In low-income and middle-income countries, the number of people receiving antiretroviral therapy (ART) grew from 300 000 in 2002 to 9.7 million by the end of 2012, with the greatest increase in coverage occurring in sub-Saharan Africa with 7.5 million ART recipients [1]. This rapid scale-up of ART provision has contributed to a 30% decrease in the global number of HIV-related deaths between 2005 and 2012 [2].
Although these gains are impressive, they also highlight the continued importance of adherence to ART. Delayed initiation of therapy, incomplete adherence and early treatment discontinuation remain among the strongest predictors of incomplete viral suppression, disease progression and mortality among individuals living with HIV [3–8]. Although a 2006 meta-analysis found that adherence levels in sub-Saharan Africa exceeded levels found among patients in developed countries [9], data from both low-income and high-income settings have raised concerns about declining adherence rates over time [10–14]. Factors often associated with incomplete adherence among adults in sub-Saharan Africa include patient-level characteristics (e.g. depression, alcohol use [15]), poor patient–provider interactions [16,17] and structural factors, including the cost of transport and distance to the health facility [18,19]. Less is known about how programme-level characteristics, for example, where ART clinics dispense drugs and if they require treatment support buddies prior to ART initiation, influence ART adherence.
This manuscript presents findings from the adherence component of a two-part study to examine retention in ART services [20] and adherence to ART among treatment-experienced adults in Tanzania, Uganda and Zambia. The objectives of the adherence study were to characterize patient ART adherence levels across multiple programme settings and to examine the relationship between individual and programme-level characteristics and incomplete adherence among ART clients who were on ART for at least 6 months.
Ethical review
The study was reviewed and approved by the ethics review board of the US Centers for Disease Control and Prevention (CDC) and the six partner and national ethical review committees. The Partners Healthcare Institutional Review Board ceded review to FHI 360.
Materials and methods
Design and study setting
A cross-sectional study was conducted among ART patients from 18 facilities from Tanzania, Uganda and Zambia. In each country, six sites with a minimum cohort size of 300 patients at the time of the protocol development (2006) were purposively chosen in order to fulfil the sample size requirement of 250 patients per study site and to explore the impact that different programme characteristics may have on retention and adherence outcomes (see supplemental Table 1, http://links.lww.com/QAD/A616). During study start-up, the study team found that one facility in Tanzania consisted of two adult ART clinics and recruited 125 patients from both the adult-only clinic and the family clinic wherein adults sought care and treatment with their families.
Study participants and data collection
Patients attending the study sites who were at least 18 years of age at ART initiation, had initiated ART at least 6 months prior to data collection and spoke one of the study languages were eligible. Participants were systematically sampled (see supplemental Table 1, http://links.lww.com/QAD/A616), and if a patient was ineligible, unwilling or unavailable, then the study teams selected the next ART patient attending the clinic. All selected patients underwent a screening and consent process by trained research interviewers and, if they consented, were interviewed. After the interview and clinic visit, the study teams collected adherence estimates from the patient’s healthcare providers (doctors, clinicians, nurses, lay workers, pharmacists) and abstracted data from the patient’s medical, pharmacy and laboratory records using structured data abstraction forms. Interviews were also held with the ART clinics managers regarding the site-level characteristics. At two sites in each country, the study team collected blood samples from all participants for HIV RNA testing. Data collection took place from May to October 2011.
Measures
Adherence measures
Participant adherence was assessed using self-report, provider report and pharmacy refill data. Participants reported their 3-day adherence using the Adult AIDS Clinical Trials Group (AACTG) measure specifying the numbers of tablets for each drug missed each day [21], followed by an aggregate question of how many ART tablets missed in the previous 30 days. Both of these measures were calculated by dividing the number of missed tablets by the total number of ART pills over the specified time period. The interviewer also asked the three-question CASE Adherence Index Scale [22], with a composite score of more than 10 indicating good adherence, and the 30-day visual analogue scale (VAS) [23], with zero indicating complete nonadherence and 100 indicating perfect adherence. To understand patterns of missed ART, we constructed a missed at least 48 consecutive hours measure from two questions about missed tablets in the past 3 months. Following the participant’s clinical visit, the interviewer asked all healthcare providers who interact with patients on adherence to estimate, based on the provider’s knowledge of the patient and their medical record, the proportion of time the participant had taken his/her drugs in the past month. In cases wherein multiple providers were interviewed, the lowest adherence estimate was taken for analysis. A Pharmacy Medication Possession Ratio (MPR) was also constructed from the pharmacy refill data by summarizing the number of pills dispensed to participants in the 6 months prior to the interview divided by the total number of pills the participants should have received during that time.
Independent variables
Basic demographic variables collected included age, sex, marital status and the demographic and health survey’s wealth index (see supplemental Table 2, http://links.lww.com/QAD/A616) [24–27].
For psychosocial factors, we assessed stigma using five yes/no questions from the Internalized AIDS Stigma Scale (IA-RSS) [28,29] that resulted in two factors retained with very good fit [with a root mean squared error of approximation (RMSEA) of 0.03, 95% confidence interval (95% CI) 0.02–0.05] and perfect reliability (Tucker and Lewis’ reliability coefficient = 1). One factor consisted of three questions on internalized stigma (Chronbach’s alpha = 0.78) and the second factor consisted of two questions on disclosure stigma (Chronbach’s alpha = 0.72). Country-specific medians were used as the cutoff points to dichotomize responses into high (>median) versus low stigma.
The Hopkins Symptoms Checklist depression subscale (HSCL-15) was used to assess symptoms of depression, using the standard cut-off of 1.75 [30] with good fit (RMSEA 0.055, 95% CI 0.053–0.058), reliability (Tucker and Lewis’s reliability coefficient = 0.90) and internal consistency (Chronbach’s alpha = 0.84).
Social support was assessed using nine questions from the Duke University, University of North Carolina Functional Social Support Questionnaire [31,32] and an added tenth question on receiving help to remember to take one’s ART. Exploratory and confirmatory factor analysis identified a two-factor model with adequate fit (RMSEA 0.097, 95% CI 0.093–0.10) and reliability (Tucker and Lewis’s reliability coefficient = 0.83). The first factor, Social Support Care, consisted of seven questions about social support (Chronbach’s alpha = 0.76) and the second factor, Social Support Help, consisted of four questions about instrumental help (Chronbach’s alpha = 0.78). Scores in the lowest tenth percentile were categorized as having low levels of social support (see supplemental Table 2, http://links.lww.com/QAD/A616).
The four yes/no questions regarding alcohol use [cutting down, annoyance by criticism, guilty feeling and eye-openers (CAGE)] were used to assess alcohol abuse or dependency. Summed scores at least 2 were considered positive for alcohol abuse/dependency [33,34]. Other psychosocial variables assessed included having ever visited a traditional healer/herbalist because of HIV (yes/no), having ever disclosed one’s HIV status and the average cost and time it took participants to reach the clinic.
The ART-related variables abstracted from the medical charts included current ART regimen, date of ART initiation and pre-ART initiation WHO stage and CD4+ cell count. In addition, we analysed participants’ self-reported daily pill burden and the 20-item HIV Symptom Index (see supplemental Table 2, http://links.lww.com/QAD/A616) [35].
The programme characteristics assessed during the ART clinic manager interviews included background information on the level, type, size and location of the HIV facility; and how the clinic dispensed ART and supported medication adherence (e.g. frequency of ART refills).
Analysis
Analysis was performed using SAS 9.3 and STATA 11.2. The analysis plan, prepared before the database was locked, included predetermined cut-offs for outcome and predictor variables with a primary focus on perfect (100%) versus incomplete adherence. These predetermined cutoffs were used to summarize the prevalence of adherence across the different measures.
To identify individual modifiable risk factors and programme characteristics associated with incomplete ART adherence, a multiple mixed effects logistic regression model was used for the adherence measure that correlated best with the HIV RNA measurement. The model contained fixed effects for all factors of interest and a random intercept term effect for the mean adherence at each site. Cut-offs for the adherence measures were selected on the basis of receiver operating characteristic (ROC) analysis with HIV RNA at least 1000 copies/ml as the reference standard. The model was constructed using a hierarchical stepwise procedure, with individual-level factors associated at the 0.10 level added first, followed by the programme characteristics significant at the 0.20 level. The model was simplified using step-wise deletion retaining only significant factors and interactions at 0.05. The estimates were corrected for predictor data missing using multiple imputations (see supplemental Table 1, http://links.lww.com/QAD/A616) [36].
Results
A total of 6825 ART clients were screened for eligibility at the participating sites. Of these, 1848 were ineligible (1523 initiated ART <6 months ago, 729 were <18 years of age and 783 did not speak one of the study languages; these categories were not mutually exclusive). Out of the 4977 eligible, 482 (9.7%) declined to participate, leaving 4495 participants who consented and completed the interview. During data cleaning, the team found six participants with ART initiation dates that were less than 6 months prior to the interview, leaving 4489 participants for the final analysis.
Programme characteristics
Nineteen ART clinics from 18 facilities were included in the analysis, seven sites in Tanzania (the facility consisting of two different models of care was considered as two separate sites), six sites in Uganda and six sites in Zambia (see supplemental Table 3, http://links.lww.com/QAD/A616). Slightly more than half of the health facilities were government facilities. Nongovernmental facilities were either faith-based or run by a nonreligious nongovernmental organization (NGO) with more NGO-supported facilities in Uganda compared with Tanzania and Zambia. More than two-thirds of the sites were located in an urban setting and eight sites had less than 2000 ART patients (range: 350–1967), seven sites between 2000 and 4000 patients (range: 2095–3989) and four had more than 4000 patients on ART (range: 4807–7471).
Characteristics of the analysis population
Of the 4489 participants, 68% were female, 56% were married/cohabitating and the average age was 40 years (Table 1). The majority of participants had started ART more than 2.2 years ago (75%), had a pre-ART CD4+ cell count of 250 cells/μl or less (67%) and were currently taking zidovudine (ZDV), lamivudine (3TC) and nevirapine (NVP)/efavirenz (EFV) (49%).
Table 1.
Participant characteristics among adults in selected antiretroviral treatment programs in Tanzania, Uganda and Zambia, 2011.
| Characteristic | Tanzania (n=1498) |
Uganda (n=1495) |
Zambia (n=1496) |
Total number of participants (N=4489) |
Sample from sites with VL measurements (n=1497) |
|---|---|---|---|---|---|
| Demographics | |||||
| Age (years) median (IQR) | 41 (35–47) | 39 (34–46) | 40 (34–47) | 40 (34–47) | 39 (33–46) |
| Age: n (%) | |||||
| <35 years | 350 (23.4) | 406 (27.2) | 395 (26.4) | 1151 (25.6) | 443 (29.6)* |
| ≥35 years | 1133 (75.6) | 1028 (68.8) | 1093 (73.1) | 3254 (72.5) | 1029 (68.7) |
| Missing | 15 (1.0) | 61 (4.1) | 8 (0.5) | 84 (1.9) | 25 (1.7) |
| Sex: n (%) | |||||
| Female | 1096 (73.2) | 982 (65.6) | 968 (64.7) | 3045 (67.8) | 999 (66.7) |
| Male | 402 (26.8) | 514 (34.4) | 528 (35.3) | 1444 (32.2) | 498 (33.3) |
| Marital status: n (%) | |||||
| Single | 204 (13.6) | 116 (7.8) | 122 (8.2) | 442 (9.8) | 149 (9.95) |
| Separate/divorced/widowed | 425 (28.4) | 637 (42.6) | 476 (31.8) | 1538 (34.3) | 496 (33.1) |
| Married/cohabitating | 863 (57.6) | 734 (49.1) | 898 (60.0) | 2495 (55.6) | 850 (56.8) |
| Missing | 6 (0.4) | 8 (0.5) | 0 (0.0) | 14 (0.3) | 2 (0.1) |
| ART characteristics | |||||
| Length of time on ART: n (%) | |||||
| <2.2 years | 425 (28.4) | 373 (24.9) | 304 (20.3) | 1102 (24.5) | 316 (21.1)* |
| 2.2–5.3 years | 739 (52.9) | 735 (49.2) | 721 (48.2) | 2249 (50.1) | 774 (51.7) |
| >5.3 years | 280 (18.7) | 387 (25.9) | 471 (31.5) | 1138 (25.4) | 407 (27.2) |
| Current ART regimen: n (%) | |||||
| D4T, 3TC, NVP | 561 (37.4) | 11 (0.7) | 159 (10.6) | 731 (16.3) | 321 (21.4)* |
| TDF, 3TC/FTC, EFV | 116 (7.7) | 86 (5.8) | 498 (33.3) | 700 (15.6) | 263 (17.5) |
| ZDV, 3TC, EFV | 396 (26.4) | 300 (20.1) | 78 (5.2) | 774 (17.2) | 271 (18.1) |
| ZDV, 3TC, NVP | 262 (17.5) | 909 (60.8) | 256 (17.1) | 1427 (31.8) | 418 (27.9) |
| Other regimens | 70 (4.7) | 175 (11.7) | 437 (29.2) | 682 (15.2) | 198 (13.2) |
| Missing | 93 (6.2) | 14 (0.9) | 68 (4.5) | 175 (3.9) | 26 (1.7) |
| CD4+ cell count prior to ART: n (%) | |||||
| >250 cells/μl | 213 (14.2) | 111 (7.4) | 243 (16.2) | 567 (12.6) | 198 (13.2) |
| ≤250 cells/μl | 995 (66.4) | 1087 (72.7) | 938 (62.7) | 3020 (67.3) | 989 (66.1) |
| Missing | 290 (19.4) | 297 (19.9) | 315 (21.1) | 902 (20.1) | 310 (20.7) |
3TC, lamivudine; D4T, stavudine; EFV, efavirenz; FTC, emtricitabine; IQR, interquartile range; NVP, nevirapine; TDF, tenofovir; VL, viral load; ZDV, zidovudine.
P<0.05 comparing the subset of participants from VL testing sites to the total study sample.
Proportion with incomplete adherence
Results from the self-reported measures of adherence found that 3.2% (141/4425) of participants missed at least 48 consecutive hours of ART in the past three months and 58% (2599/4450) reported taking less than 100% of their ART in the past 30 days using a VAS (Table 2). Providers estimated that 13.5% of participants were less than 80% adherent and 71.3% were less than 100% adherent in the past 30 days. The pharmacy MPR data found that 40% (1634/4070) of participants had less than 90% of their ART during the previous 6 months.
Table 2.
Number of individuals with incomplete adherence by the different adherence measures and standardized cut-off points among adults in selected antiretroviral treatment programs in Tanzania, Uganda and Zambia, 2011.
| Complete population |
Sample from sites with VL measurements |
||||
|---|---|---|---|---|---|
| Cut-off | n/N a | % less than the cut-off | n/N b | % | |
| Self-reported adherence measures | |||||
| % ARV drugs missed – last 3 days | 100% | 249/4299 | 5.8 | 109/1470 | 7.4 |
| 95% | 249/4299 | 5.8 | 109/1470 | 7.4 | |
| 90% | 235/4299 | 5.5 | 106/1470 | 7.2 | |
| 80% | 130/4299 | 3.0 | 55/1470 | 3.7 | |
| % ARV drugs missed – last 30 days | 100% | 517/4289 | 12.1 | 239/1460 | 16.4 |
| 95% | 113/4289 | 2.6 | 43/1460 | 2.9 | |
| 90% | 46/4289 | 1.1 | 14/1460 | 1.0 | |
| 80% | 22/4289 | 0.5 | 8/1460 | 0.5 | |
| % ARV drugs missed – visual analogue scale, last 30 days |
100% | 2599/4450 | 58.4 | 682/1484 | 46.0 |
| 95% | 1730/4450 | 38.9 | 427/1484 | 28.8 | |
| 90% | 1119/4450 | 25.1 | 253/1484 | 17.0 | |
| 80% | 511/4450 | 11.5 | 85/1484 | 5.7 | |
| Case Adherence Index | >10 | 751/4473 | 16.8 | 304/1493 | 20.4 |
| Missed ≥48 consecutive hours of ARV drugs in last 3 months |
Yes | 141/4425 | 3.2 | 64/1471 | 4.4 |
| Other adherence measures | |||||
| Healthcare provider report | 100% | 3150/4415 | 71.3 | 792/1463 | 54.1 |
| 95% | 1835/4415 | 41.6 | 476/1463 | 32.5 | |
| 90% | 1236/4415 | 28.0 | 302/1463 | 20.6 | |
| 80% | 594/4415 | 13.5 | 157/1463 | 10.7 | |
| Pharmacy medication possession ratio (MPR) |
100% | 3261/4070 | 80.1 | 1275/1450 | 87.9 |
| 95% | 2131/4070 | 52.4 | 837/1450 | 57.7 | |
| 90% | 1634/4070 | 40.1 | 659/1450 | 45.4 | |
| 80% | 979/4070 | 24.1 | 429/1450 | 29.6 | |
Number nonadherent/total number of individuals with data in the population. ARV, antiretroviral.
The denominators for these adherence measures vary due to missing data mainly on the current ART regimen and associated daily pill burden needed to calculate adherence.
Selection of optimal adherence measures based on association with virologic failure
One thousand, four hundred and ninety-eight participants received HIV RNA testing and were younger, on ART for less time and had different ART regimens, compared with those who had not received testing. The proportion of participants with at least 1000 copies/ml ranged across the testing sites from 7.2 to 17.2% (mean 9.9%). Using the ROC curve analysis cutoff points, only three adherence measures were significantly related to virologic failure (Table 3), with at least 48 consecutive hours of missed ART having the strongest association with an odds ratio (OR) of 2.86 (95% CI 1.56–5.26) followed by the pharmacy MPR less than 90% (OR 1.47, 95% CI 1.02–2.15) and the healthcare provider report less than 98% (OR 1.57, 95% CI 1.02–2.41). On the basis of these results and previous evidence of the importance of treatment interruptions as a predictor of virologic failure and resistance [37,38], combined with provider estimates most likely reflecting a clinical assessment rather than pill-taking behaviours [39–41], the team selected the variable at least 48 consecutive hours of missed ART as the primary outcome variable to examine the factors associated with incomplete adherence.
Table 3.
Association between adherence measures and virologic failure at least 1000 copies/ml among adults in selected antiretroviral treatment programmes in Tanzania, Uganda and Zambia, 2011.
| Patients with virologic failure ≥1000 copies/ml |
|||
|---|---|---|---|
| n/N a | OR (95% CI)b | P * | |
| Self-reported adherence measuresc | |||
| % ARV drugs missed – last 3 days | |||
| <93% | 12/107 (11.2%) | 1.28 (0.68–2.42) | 0.450 |
| ≥93% | 132/1363 (9.7%) | ||
| % ARV drugs missed – last 30 days | |||
| <99% | 22/237 (9.3%) | 1.04 (0.63–1.69) | 0.888 |
| ≥99% | 122/1223 (10.0%) | ||
| % ARV drugs missed – visual analogue scale – last 30 days | |||
| <99% | 68/633 (10.7%) | 0.86 (0.57–1.30) | 0.479 |
| ≥99% | 80/851 (9.4%) | ||
| CASE Adherence Index | |||
| <11 | 33/304 (10.9%) | 1.30 (0.84–2.01) | 0.233 |
| ≥11 | 115/1189 (9.7%) | ||
| Missed ≥48 consecutive hours of ARV drugs in past 3 months | |||
| ≥48 h | 16/64 (25.0%) | 2.86 (1.56–5.26) | 0.001 |
| <48 h | 128/1407 (9.1%) | ||
| Other adherence measures | |||
| Healthcare provider report | |||
| <98% | 82/628 (13.1%) | 1.57 (1.02–2.41) | 0.042 |
| ≥98% | 60/835 (7.2%) | ||
| Pharmacy medication possession ratio (MPR) | |||
| <90% | 72/659 (10.9%) | 1.48 (1.02–2.15) | 0.037 |
| ≥90% | 67/791 (8.5%) | ||
ARV, antiretroviral; CI, confidence interval; OR, odds ratio.
N, number of patients in subgroup; n, number of failures in subgroup.
OR of failure among incomplete-adherent versus adherent patients from logistic regression.
Cut-off points determined by receiver operator curve (ROC) analysis.
P for association between adherence measure and failure, adjusted for ART study site.
Factors associated with incomplete adherence
During univariate analysis, significant associations at the 0.10 level between individual demographic and psychosocial characteristics and incomplete adherence were male sex, high levels of internalized stigma, screening positive for depression, low levels of social support care, screening positive for alcohol abuse, having ever consulted a traditional healer/herbalist because of HIV and having a current ART regimen that does not contain NVP (Table 4). In terms of programme characteristics, significant associations at the 0.20 level were having routine ART refills every 2–3 months compared with once a month, the level and type of health facility, urban location and not offering community-based distribution of ART.
Table 4.
Bivariate analysis of missed at least 48 consecutive hours of antiretroviral therapy in the past 3 months among adults in selected antiretroviral treatment programmes in Tanzania, Uganda and Zambia, 2011.
| N=4425 | Total % | <48 h (n=4284) % |
≥48 h (n=141) % |
Odds ratio (95% CI) |
P | |
|---|---|---|---|---|---|---|
| Demographics | ||||||
| Age (in years) | ||||||
| <35 | 1128 | 26.0 | 26.0 | 26.2 | 0.97 (0.66–1.43) | 0.867 |
| ≥35 years | 3215 | 74.0 | 74.0 | 73.8 | 1 | |
| Missing | 82 | 82 | 0 | |||
| Female sex | 3006 | 67.9 | 68.2 | 61.0 | 0.72 (0.51–1.02) | 0.065 |
| Marital status | ||||||
| Single | 434 | 9.8 | 9.7 | 12.8 | 1.54 (0.90–2.63) | 0.227 |
| Separated/divorced/widowed | 1523 | 34.5 | 34.4 | 37.6 | 1.24 (0.86–1.80) | |
| Married or cohabiting | 2456 | 55.7 | 55.9 | 49.6 | 1 | |
| Missing | 12 | 12 | 0 | |||
| DHS Wealth Index | ||||||
| Low | 1473 | 33.3 | 33.4 | 31.2 | 0.79 (0.50–1.25) | 0.579 |
| Middle | 1476 | 33.4 | 33.4 | 31.9 | 0.86 (0.56–1.31) | |
| High | 1476 | 33.4 | 33.2 | 36.9 | 1 | |
| Psychosocial factors | ||||||
| Stigma internalized | ||||||
| High (>median) | 1540 | 34.8 | 34.3 | 49.6 | 1.63 (1.15–2.32) | <0.006 |
| Low | 2883 | 65.2 | 65.7 | 50.4 | 1 | |
| Missing | 2 | 2 | 0 | |||
| Stigma disclosure | ||||||
| High (>median) | 1212 | 27.4 | 27.5 | 25.5 | 0.92 (0.59–1.44) | <0.728 |
| Low | 3211 | 72.6 | 72.5 | 74.5 | 1 | |
| Missing | 2 | 2 | 0 | |||
| Potential depression | ||||||
| Positive screen | 574 | 13.0 | 12.6 | 23.4 | 1.89 (1.25–2.87) | 0.003 |
| Negative screen | 3847 | 87.0 | 87.4 | 76.6 | 1 | |
| Missing | 4 | 4 | 0 | |||
| Ever-disclosed HIV status | ||||||
| Yes | 4154 | 94.0 | 94.0 | 92.9 | 0.90 (0.45–1.79) | 0.771 |
| No | 267 | 6.00 | 6.00 | 7.10 | 1 | |
| Missing | 4 | 4 | 0 | |||
| Social support care | ||||||
| Lower 10th percentile | 513 | 11.7 | 11.4 | 20.0 | 1.71 (1.10–2.67) | 0.017 |
| Higher | 3865 | 88.3 | 88.6 | 80.0 | 1 | |
| Missing | 48 | 47 | 1 | |||
| Social support help | ||||||
| Lower 10th percentile | 595 | 13.6 | 13.5 | 17.1 | 1.23 (0.77–1.95) | 0.387 |
| Higher | 3782 | 86.4 | 86.5 | 82.9 | 1 | |
| Missing | 48 | 47 | 1 | |||
| CAGE alcohol abuse | ||||||
| Positive ≥2 | 930 | 21.2 | 20.8 | 34.3 | 1.87 (1.29–2.70) | <0.001 |
| Negative<2 | 3448 | 78.8 | 79.2 | 65.7 | 1 | |
| Missing | 47 | 46 | 1 | |||
| Traditional healer | ||||||
| Ever consulted | 253 | 5.8 | 5.4 | 15.6 | 2.67 (1.55–4.60) | <0.001 |
| Never consulted | 4136 | 94.2 | 94.6 | 84.4 | 1 | |
| Missing | 36 | 36 | 0 | |||
| Cost to clinic | ||||||
| ≥1 USD | 1924 | 45.9 | 46.0 | 41.5 | 0.82 (0.56–1.20) | 0.310 |
| <1 USD | 2269 | 54.1 | 54.0 | 58.5 | 1 | |
| Missing | 232 | 226 | 6 | |||
| Time to clinic | ||||||
| ≥30 min | 2298 | 52.0 | 51.9 | 52.5 | 0.93 (0.65,1.34) | 0.694 |
| <30 min | 2215 | 48.0 | 48.1 | 47.5 | 1 | |
| Missing | 2 | 2 | 0 | |||
| ART and clinical factors | ||||||
| ART regimens containing | ||||||
| EVF | 1540 | 34.8 | 34.7 | 38.3 | 1.35 (0.94–1.95) | 0.103 |
| NVP | 2567 | 58.0 | 58.3 | 48.9 | 0.57 (0.40–0.81) | 0.002 |
| ZDV | 2224 | 50.3 | 50.2 | 51.8 | 0.85 (0.58–1.25) | 0.410 |
| TDF | 1131 | 25.6 | 25.7 | 20.6 | 0.95 (0.60–1.50) | 0.823 |
| D4T | 805 | 18.2 | 18.2 | 17.0 | 0.84 (0.50–1.39) | 0.490 |
| FTC | 1020 | 23.1 | 23.2 | 17.7 | 0.92 (0.56–1.51) | 0.752 |
| 3TC | 3191 | 72.1 | 72.1 | 72.3 | 0.72 (0.47–1.11) | 0.141 |
| PI | 136 | 3.10 | 3.10 | 3.50 | 1.13 (0.45–2.86) | 0.798 |
| Time on ART | ||||||
| <2.2 | 1089 | 24.6 | 24.5 | 27.7 | 1.18 (0.73–1.90) | 0.597 |
| 2.2–5.3 | 2213 | 50.0 | 50.1 | 46.8 | 0.96 (0.62–1.46) | |
| >5.3 | 1123 | 25.4 | 25.4 | 25.5 | ||
| Pill burden (self-report) | ||||||
| <4 | 3510 | 81.3 | 81.2 | 84.7 | 1.15 (0.70–1.90) | 0.578 |
| ≥4 | 809 | 18.7 | 18.8 | 15.3 | 1 | |
| Missing | 106 | 96 | 10 | |||
| HIV symptom index (≥ median) | 2458 | 55.5 | 55.0 | 73.0 | 1.98 (1.34–2.94) | <0.001 |
| Pre-ART WHO stage | ||||||
| Missing | 381 | 8.6 | 8.7 | 7.1 | 0.95 (0.47–1.95) | 0.144 |
| Stage IV | 517 | 11.7 | 11.4 | 19.1 | 1.74 (1.02–2.96) | |
| Stage III | 1842 | 41.6 | 41.6 | 42.6 | 1.38 (0.92–2.09) | |
| Stage I and II | 1685 | 38.1 | 38.3 | 31.2 | ||
| Pre-ART CD4+ cell count (cells/μl) | ||||||
| Missing | 889 | 20.1 | 20.0 | 22.0 | 1.11 (0.73–1.69) | .038 |
| >250 | 558 | 12.6 | 12.7 | 9.2 | 0.69 (0.38–1.26) | |
| ≤250 | 2978 | 67.3 | 67.3 | 68.8 | 1 | |
| Site-level factors | ||||||
| Buddy required to start ART (yes) | 3440 | 77.7 | 77.5 | 85.1 | 1.51 (0.54–4.21) | 0.4051 |
| Community ART dispensing (yes) | 1228 | 27.8 | 28.0 | 19.9 | 0.53 (0.20–1.39) | 0.182 |
| ART dispensing in clinic | 2823 | 63.8 | 63.4 | 74.5 | 1.54 (0.66–3.60) | 0.301 |
| ART refill frequency | ||||||
| Every 3 months | 1224 | 27.7 | 27.3 | 39.0 | 3.15 (0.41–4.11) | 0.152 |
| Every 2 months | 2217 | 50.1 | 50.1 | 49.6 | 2.35 (0.77–7.19) | |
| Every month | 984 | 22.2 | 22.6 | 11.3 | 1 | |
| Level of health facility | ||||||
| National | 732 | 16.5 | 16.4 | 21.3 | 1.30 (0.41–4.11) | 0.174 |
| Provincial/regional | 983 | 22.2 | 21.9 | 30.5 | 1.22 (0.39–3.79) | |
| District | 1480 | 33.4 | 34.0 | 17.7 | 0.42 (0.14–1.26) | |
| Primary/community | 1230 | 27.8 | 27.7 | 30.5 | 1 | |
| Type of health facility | ||||||
| Government | 2467 | 55.8 | 56.4 | 36.2 | 0.40 (0.16–1.00) | 0.086 |
| Mission | 976 | 22.1 | 21.8 | 31.2 | 0.84 (0.30–2.39) | |
| NGO | 982 | 22.2 | 21.8 | 32.6 | 1 | |
| Health facility size | ||||||
| <1000 on ART | 1344 | 30.4 | 30.3 | 31.9 | 1.05 (0.28–3.95) | 0.955 |
| 1000–2000 | 370 | 8.4 | 8.4 | 7.1 | 1.20 (0.20–7.07) | |
| 2000–4000 | 1723 | 38.9 | 38.8 | 41.8 | 1.34 (0.38–4.72) | |
| >4000 | 988 | 22.3 | 22.4 | 19.1 | 1 | |
| Locale | ||||||
| Nonurban | 1470 | 33.2 | 33.5 | 26.2 | 0.49 (0.19–1.22) | 0.118 |
| Urban | 2955 | 66.8 | 66.5 | 73.8 | 1 | |
| Lay person provides adherence counselling | 3566 | 80.6 | 80.6 | 80.1 | 0.86 (0.30–2.42) | 0.758 |
| ART stock out past 6 months | 497 | 11.2 | 11.4 | 5.0 | 0.41 (0.10–1.72) | 0.209 |
3TC, lamivudine; ART, antiretroviral therapy; CI, confidence interval; D4T, stavudine; DHS, demographic and health survey; EFV, efavirenz; FTC, emtricitabine; NGO, nongovernmental organization; NVP, nevirapine; PI, protease inhibitors; TDF, tenofovir; USD, United States dollar; ZDV, zidovudine.
The final multiple regression model found five independent variables associated with incomplete adherence (Table 5): having high levels of internalized stigma, screening positive for alcohol abuse, ever consulting a traditional healer/herbalist because of HIV and having higher numbers of HIV-related symptoms. The analysis also found that having an ART regimen containing NVP was protective against incomplete adherence and a lack of social support care was marginally associated with incomplete adherence. The final model did not include any of the programme characteristics.
Table 5.
Final multivariable model for factors associated with incomplete adherence: missed at least 48 consecutive hours of antiretroviral drugs among adults in selected antiretroviral treatment programmes in Tanzania, Uganda and Zambia, 2011.
| Adjusted odds ratio (95% CI) |
P | |
|---|---|---|
| Internalized stigma: high | 1.50 (1.05–2.13) | 0.025 |
| Social support care: low | 1.55 (0.99–2.43) | 0.054 |
| Alcohol abuse: positive | 1.68 (1.16–2.44) | 0.006 |
| HIV symptoms: high | 1.79 (1.20–2.67) | 0.004 |
| Traditional medicine: ever visited | 2.41 (1.39–4.18) | 0.002 |
| ART regimen includes NVP: Yes | 0.60 (0.42–0.85) | 0.005 |
ART, antiretroviral treatment; CI, confidence interval; NVP, nevirapine.
Discussion
This first multicountry study of individual and programme-level factors related to adherence among patients in sub-Saharan Africa on ART for at least 6 months found that social and behavioural factors, including HIV stigma and alcohol abuse, are associated with lower ART adherence. These data support the development of interventions to address high-risk alcohol use and internalized stigma as potential ways to enhance ART adherence and HIV-related health outcomes. The findings also underscore the need to examine how traditional healers may support or hinder ART adherence. Furthermore, these data highlight the variability of existing adherence measures and the need for both accurate research and programme-level methods for assessing pill-taking behaviours in order to inform programme strategies and assess intervention impact.
This research is also the first multicountry study to examine self-reported prevalence of missed antiretroviral drugs for at least 48 consecutive hours, in addition to the more common adherence measures of 3 and 30-day self-reports, pharmacy refill and provider reports. Similar to previous studies, the adherence estimates varied greatly with the 3 and 30-day aggregate questions generating the lowest estimates of missed ART [42] and corresponding poorly with virologic failure. The study results also reaffirm findings that pharmacy MPRs are associated with virologic failure [43], and contrary to other studies [39,41], found that provider estimates were also associated with virologic failure. Providers, including clinicians, pharmacists and adherence counsellors, gave their adherence estimates after a patient visit based on their knowledge of the patient and the patient’s medical record. As such, their assessment presumably reflects a combination of factors including knowledge of patient behaviours as well as clinical and laboratory treatment response prior to the assessment. Similarly, the pharmacy MPR takes into account patient’s medication refills over a 6-month time period capturing the maximum adherence a patient may achieve rather than actual pill-taking behaviours. Although only able to identify 11% of virological failures in our cross-sectional assessment, the strength of the relationship between the at least 48 consecutive hours measure with virologic failure is consistent with several observations that patterns of adherence, namely interruptions during otherwise good adherence, is an important cause of virologic failure [37,38].
This is also the first study to examine the factors associated with incomplete adherence using at least 48 consecutive hours of missed ART as the main outcome. The factor most strongly related to incomplete adherence was having ever consulted a traditional healer or herbalist because of HIV. About three-quarters of the 253 participants who reported having ever visited a traditional healer or herbalist were from Tanzania. During data collection, a famous healer in the Manyara District of Tanzania was offering a liquid cure for chronic illnesses including HIV/AIDS [44,45]. Although our questionnaire did not investigate details regarding the healer, this finding does contribute to a growing body of research on the role that traditional healers and alternative medicines may have in delaying HIV testing and care-seeking behaviours [46–48] and influencing sustained HIV ART adherence [49–52].
This research also reinforces findings that alcohol abuse [53,54] and HIV stigma [55] are broad and consistent correlates of ART adherence. The CAGE questionnaire is a concise four-question tool that may be incorporated into clinic assessments to identify ART recipients who may benefit from alcohol counselling interventions. More than one-third of participants reported high levels of internalized stigma defined as feelings of worthlessness, shame and guilt because of one’s HIV-positive status, despite the fact that most of the participants in this research had initiated ART more than 2 years earlier. Stigma is associated with lower rates of HIV disclosure, which can compromise adherence due to both concealing the medication and preventing access to social support [56–58]. Several qualitative studies have found social support to be critical for people to overcome barriers, particularly structural and economic barriers, to adherence [16,59–61]. Although this study did not find a significant relationship between structural barriers (e.g. transport costs to the clinic) and adherence, we did find that people with lower levels of social support care were more likely to have missed at least 48 consecutive hours of ART in the past 3 months (P < 0.054).
Other factors related to incomplete adherence include having a greater burden of HIV-related symptoms as well as an ART regimen that did not include NVP. Women of childbearing age and/or actively considering pregnancy were given NVP rather than EFV during the study period out of concern for EFV-related teratogenicity. Pregnancy has been shown to positively impact adherence in some but not all studies [62,63]. Although our analysis controlled for age and sex, we did not control for fertility desire or pregnancy intention. Therefore, the association between NVP and adherence may be confounded by preferential prescription of NVP to patients pregnant or considering pregnancy. Alternatively, higher adherence to NVP may be related to differing side effect profiles for NVP versus EFV, which was not fully controlled for in our analysis.
These factors –— internalized stigma, alcohol abuse, low levels of social support, visiting traditional healers and burden of HIV symptoms –— all give insights into the challenges treatment-experienced people living with HIV are dealing with on a daily basis even after several years of taking ART. The current biomedical approach to HIV treatment is that you take your ART regularly, you feel better and your life returns to ‘normal’. This script, however, does not always play out in people’s lives [64]. Tsai et al. [65] suggest that stigma in low-resource settings is tied with concepts of disability, economic incapacity and death, and that poverty alleviation strategies may be effective in reducing HIV-related stigma. More innovative research around self-identity and living with HIV as a chronic illness [64], combined with strategies to overcome and reduce stigma through programmes, such as poverty alleviation that can restore one’s identity, are needed, as HIV services seek to support life-long ART adherence and healthy outcomes among people living with the virus.
Strengths and limitations
The strengths of this study include representation from 19 ART clinics in 18 different health facilities in three countries using standardized data collection tools, an HIV RNA test among a subsample and a systematic sampling strategy among ART patients who initiated therapy at least 6 months prior to data collection. Although the number of participating sites is one of the study’s strengths, 19 ART clinics were too few to fully examine the association of programme characteristics with ART adherence. The cross-sectional design of the study restricts interpretations to associations rather than temporal relationships and causation. For example, virologic failure could have happened prior to the time frame of 3 and 30-days captured by the self-reported adherence measures, potentially explaining the low overall correlations we found between the different measures of adherence and virologic failure. Study sites were also not randomly selected, and this could have introduced selection bias. However, the selection was conducted in consultation with country-specific stake-holders and aimed to balance site characteristics that might influence retention and adherence. Within study sites, we were also unable to assess whether ART clients who declined to participate differed from those who chose to participate.
Conclusion
These data support the importance of social and behavioural factors on impacting adherence in resource-limited settings. Interventions to reduce alcohol abuse and stigma should be paramount and efforts to enhance social support may improve adherence. The findings also support a more in-depth investigation of the role that traditional healers and alternative medicines play in how adults living with HIV in sub-Saharan Africa manage their infection. Although the study did not find evidence that programme characteristics relate to incomplete adherence, there is a need to critically assess programme approaches, as well as develop interventions to address alcohol abuse and internalized stigma, among treatment-experienced adult ART clients.
Acknowledgements
Y.D.M., J.A.D, K.T, E.V.P, G.P. K., F.W.M, M.M., D.R.B, R.C., J.M and O.K. designed the study. J.A.D., O.K., J.M., R.C. and D.R.B. wrote the first draft of the article. J.M and S.T. performed the statistical analysis. A.A., S.A. and J.K. contributed critical revisions to the analysis and interpretation. All authors contributed to the writing of the final draft.
This research has been supported by the President’s Emergency Plan for AIDS Relief (PEPFAR) through the Centers for Disease Control and Prevention (CDC) and the Health Resources & Services Administration (HRSA) under the terms of the contract no. 2006-N-08428 with FHI 360. The findings and conclusions in this report are those of the author(s) and do not necessarily represent the official position of CDC, HRSA or any other federal agency or office.
The authors gratefully acknowledge the invaluable contributions of the people living with HIV who participated in this study and to the staff of the health facilities where the research was conducted. We also gratefully acknowledge Dr Seymour Williams of CDC who was instrumental in designing the study and to the study coordinators who oversaw study implementation in each of the countries, Drs Marysia Antony Tukai, Esther Birabwa and Sebastian Hachizovu.
Footnotes
Conflicts of interest
The authors have no conflicts of interest to report.
References
- 1.WHO . Global update on HIV treatment 2013: results, impact and opportunities: WHO report in partnership with UNICEF and UNAIDS. WHO; Geneva, Switzerland: 2013. [Google Scholar]
- 2.UNAIDS . report: UNAIDS report on the global AIDS epidemic 2013. Joint United Nations Programme on HIV/AIDS (UNAIDS); Geneva, Switzerland: 2013. [Google Scholar]
- 3.Bangsberg DR, Perry S, Charlebois ED, Clark RA, Roberston M, Zolopa AR, et al. Nonadherence to highly active antiretroviral therapy predicts progression to AIDS. AIDS. 2001;15:1181–1183. doi: 10.1097/00002030-200106150-00015. [DOI] [PubMed] [Google Scholar]
- 4.Gardner EM, Burman WJ, Steiner JF, Anderson PL, Bangsberg DR. Antiretroviral medication adherence and the development of class-specific antiretroviral resistance. AIDS (London, England) 2009;23:1035. doi: 10.1097/QAD.0b013e32832ba8ec. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Nachega JB, Hislop M, Dowdy DW, Lo M, Omer SB, Regensberg L, et al. Adherence to highly active antiretroviral therapy assessed by pharmacy claims predicts survival in HIV-infected South African adults. J Acquir Immune Defic Syndr. 2006;43:78–84. doi: 10.1097/01.qai.0000225015.43266.46. [DOI] [PubMed] [Google Scholar]
- 6.Babiker AG, Emery S, Fätkenheuer G, Gordin FM, Grund B, Lundgren JD, et al. Considerations in the rationale, design and methods of the Strategic Timing of AntiRetroviral Treatment (START) study. Clin Trials. 2013;10:S5–S36. doi: 10.1177/1740774512440342. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Braitstein P, Brinkhof M, Dabis F, Schechter M, Boulle A, Miotti P, et al. Mortality of HIV-1-infected patients in the first year of antiretroviral therapy: comparison between low-income and high-income countries. Lancet. 2006;367:817–824. doi: 10.1016/S0140-6736(06)68337-2. [DOI] [PubMed] [Google Scholar]
- 8.Ferradini L, Jeannin A, Pinoges L, Izopet J, Odhiambo D, Mankhambo L, et al. Scaling up of highly active antiretroviral therapy in a rural district of Malawi: an effectiveness assessment. Lancet. 2006;367:1335–1342. doi: 10.1016/S0140-6736(06)68580-2. [DOI] [PubMed] [Google Scholar]
- 9.Mills EJ, Nachega JB, Buchan I, Orbinski J, Attaran A, Singh S, et al. Adherence to antiretroviral therapy in sub-Saharan Africa and North America: a meta-analysis. JAMA. 2006;296:679–690. doi: 10.1001/jama.296.6.679. [DOI] [PubMed] [Google Scholar]
- 10.Byakika-Tusiime J, Crane J, Oyugi JH, Ragland K, Kawuma A, Musoke P, et al. Longitudinal antiretroviral adherence in HIV+ Ugandan parents and their children initiating HAART in the MTCT-Plus family treatment model: role of depression in declining adherence over time. AIDS Behav. 2009;13:82–91. doi: 10.1007/s10461-009-9546-x. [DOI] [PubMed] [Google Scholar]
- 11.Parruti G, Manzoli L, Toro PM, D’amico G, Rotolo S, Graziani V, et al. Long-term adherence to first-line highly active anti-retroviral therapy in a hospital-based cohort: predictors and impact on virologic response and relapse. AIDS Pat Care STDs. 2006;20:48–56. doi: 10.1089/apc.2006.20.48. [DOI] [PubMed] [Google Scholar]
- 12.Mannheimer S, Friedland G, Matts J, Child C, Chesney M. The consistency of adherence to antiretroviral therapy predicts biologic outcomes for human immunodeficiency virus: infected persons in clinical trials. Clin Infect Dis. 2002;34:1115–1121. doi: 10.1086/339074. [DOI] [PubMed] [Google Scholar]
- 13.Liu H, Miller LG, Hays RD, Golin CE, Wu T, Wenger NS, et al. Repeated measures longitudinal analyses of HIV virologic response as a function of percentage adherence, dose timing, genotypic sensitivity, and other factors. J Acquir Immune Defic Syndr. 2006;41:315–322. doi: 10.1097/01.qai.0000197071.77482.6e. [DOI] [PubMed] [Google Scholar]
- 14.Gill CJ, Hamer DH, Simon JL, Thea DM, Sabin LL. No room for complacency about adherence to antiretroviral therapy in sub-Saharan Africa. AIDS. 2005;19:1243–1249. doi: 10.1097/01.aids.0000180094.04652.3b. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Mayston R, Kinyanda E, Chishinga N, Prince M, Patel V. Mental disorder and the outcome of HIV/AIDS in low-income and middle-income countries: a systematic review. AIDS. 2012;26:S117–S135. doi: 10.1097/QAD.0b013e32835bde0f. [DOI] [PubMed] [Google Scholar]
- 16.Watt MH, Maman S, Golin CE, Earp JA, Eng E, Bangdiwala SI, et al. Factors associated with self-reported adherence to anti-retroviral therapy in a Tanzanian setting. AIDS Care. 2010;22:381–389. doi: 10.1080/09540120903193708. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Siedner MJ, Haberer JE, Bwana MB, Ware NC, Bangsberg DR. High acceptability for cell phone text messages to improve communication of laboratory results with HIV-infected patients in rural Uganda: a cross-sectional survey study. BMC Med Inform Decis Mak. 2012;12:56. doi: 10.1186/1472-6947-12-56. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Lankowski AJ, Siedner MJ, Bangsberg DR, Tsai AC. Impact of geographic and transportation-related barriers on HIV outcomes in sub-Saharan Africa: a systematic review. AIDS Behav. 2014;18:1199–1223. doi: 10.1007/s10461-014-0729-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Kagee A, Remien R, Berkman A, Hoffman S, Campos L, Swartz L. Structural barriers to ART adherence in Southern Africa: challenges and potential ways forward. Glob Public Health. 2011;6:83–97. doi: 10.1080/17441691003796387. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Koole O, Tsui S, Wabwire-Mangen F, Kwesigabo G, Menten J, Mulenga M, et al. Retention and risk factors for attrition among adults in antiretroviral treatment programmes in Tanzania, Uganda and Zambia. Trop Med Int Health. 2014;19:1397–1410. doi: 10.1111/tmi.12386. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Chesney MA, Ickovics J, Chambers D, Gifford A, Neidig J, Zwickl B, et al. Self-reported adherence to antiretroviral medications among participants in HIV clinical trials: the AACTG adherence instruments. AIDS Care. 2000;12:255–266. doi: 10.1080/09540120050042891. [DOI] [PubMed] [Google Scholar]
- 22.Mannheimer S, Mukherjee R, Hirschhorn L, Dougherty J, Celano S, Ciccarone D, et al. The CASE adherence index: a novel method for measuring adherence to antiretroviral therapy. AIDS Care. 2006;18:853–861. doi: 10.1080/09540120500465160. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Giordano TP, Guzman D, Clark R, Charlebois E, Bangsberg D. Measuring adherence to antiretroviral therapy in a diverse population using a visual analogue scale. HIV Clin Trials. 2004;5:74–79. doi: 10.1310/JFXH-G3X2-EYM6-D6UG. [DOI] [PubMed] [Google Scholar]
- 24.Central Statistical Office MoH. Tropical Diseases Research Centre. University of Zambia, Inc MI . Zambia Demographic and Health Survey 2007. CSO and Macro International Inc Calverton; Maryland: 2009. [Google Scholar]
- 25.Statistics UBo, ICF International . Uganda Demographic and Health Survey 2011. UBOS and Calverton; Maryland: ICF International Inc; Kampala, Uganda: 2012. [Google Scholar]
- 26.National Bureau of Statistics (NBS) [Tanzania] and ICF Macro . Tanzania demographic and health survey 2010. NBS and ICF Macro; Dar es Salaam, Tanzania: 2011. [Google Scholar]
- 27.Rutstein Shea O and Kiersten Johnson . The DHS wealth index. ORC Macro; Calverton, Maryland: 2004. DHS Comparative Reports No. 6. [Google Scholar]
- 28.Simbayi LC, Kalichman S, Strebel A, Cloete A, Henda N, Mqeketo A. Internalized stigma, discrimination, and depression among men and women living with HIV/AIDS in Cape Town, South Africa. Soc Sci Med. 2007;64:1823–1831. doi: 10.1016/j.socscimed.2007.01.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Kalichman SC, Simbayi LC, Cloete A, Mthembu PP, Mkhonta RN, Ginindza T. Measuring AIDS stigmas in people living with HIV/AIDS: the Internalized AIDS-Related Stigma Scale. AIDS Care. 2009;21:87–93. doi: 10.1080/09540120802032627. [DOI] [PubMed] [Google Scholar]
- 30.Kaaya SF, Fawzi M, Mbwambo J, Lee B, Msamanga G, Fawzi W. Validity of the Hopkins Symptom Checklist-25 amongst HIV-positive pregnant women in Tanzania. Acta Psychiatr Scand. 2002;106:9–19. doi: 10.1034/j.1600-0447.2002.01205.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Broadhead W, Gehlbach SH, De Gruy FV, Kaplan BH. The Duke-UNC Functional Social Support Questionnaire: measurement of social support in family medicine patients. Med Care. 1988;26:709–723. doi: 10.1097/00005650-198807000-00006. [DOI] [PubMed] [Google Scholar]
- 32.Antelman G, Kaaya S, Wei R, Mbwambo J, Msamanga GI, Fawzi WW, et al. Depressive symptoms increase risk of HIV disease progression and mortality among women in Tanzania. J Acquir Immune Defic Syndr. 2007;44:470–477. doi: 10.1097/QAI.0b013e31802f1318. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Claassen J. The benefits of the CAGE as a screening tool for alcoholism in a closed rural South African community. South Afr Med J. 1999;89:976. [PubMed] [Google Scholar]
- 34.Samet JH, Phillips SJ, Horton NJ, Traphagen ET, Freedberg KA. Detecting alcohol problems in HIV-infected patients: use of the CAGE questionnaire. AIDS Res Hum Retroviruses. 2004;20:151–155. doi: 10.1089/088922204773004860. [DOI] [PubMed] [Google Scholar]
- 35.Justice A, Holmes W, Gifford A, Rabeneck L, Zackin R, Sinclair G, et al. Development and validation of a self-completed HIV symptom index. J Clin Epidemiol. 2001;54:S77–S90. doi: 10.1016/s0895-4356(01)00449-8. [DOI] [PubMed] [Google Scholar]
- 36.Royston P. Multiple imputation of missing values. Stata J. 2004;4:227–241. [Google Scholar]
- 37.Oyugi JH, Byakika-Tusiime J, Ragland K, Laeyendecker O, Mugerwa R, Kityo C, et al. Treatment interruptions predict resistance in HIV-positive individuals purchasing fixed-dose combination antiretroviral therapy in Kampala, Uganda. AIDS. 2007;21:965–971. doi: 10.1097/QAD.0b013e32802e6bfa. [DOI] [PubMed] [Google Scholar]
- 38.Genberg BL, Wilson IB, Bangsberg DR, Arnsten J, Goggin K, Remien RH, et al. Patterns of antiretroviral therapy adherence and impact on HIV RNA among patients in North America. AIDS. 2012;26:1415–1423. doi: 10.1097/QAD.0b013e328354bed6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Bangsberg DR, Hecht FM, Clague H, Charlebois ED, Ciccarone D, Chesney M, et al. Provider assessment of adherence to HIV antiretroviral therapy. J Acquir Immune Defic Syndr. 2001;26:435–442. doi: 10.1097/00126334-200104150-00005. [DOI] [PubMed] [Google Scholar]
- 40.Paterson DL, Swindells S, Mohr J, Brester M, Vergis EN, Squier C, et al. Adherence to protease inhibitor therapy and outcomes in patients with HIV infection. Ann Intern Med. 2000;133:21–30. doi: 10.7326/0003-4819-133-1-200007040-00004. [DOI] [PubMed] [Google Scholar]
- 41.Gross R, Bilker WB, Friedman HM, Coyne JC, Strom BL. Provider inaccuracy in assessing adherence and outcomes with newly initiated antiretroviral therapy. AIDS. 2002;16:1835–1837. doi: 10.1097/00002030-200209060-00021. [DOI] [PubMed] [Google Scholar]
- 42.Bärnighausen T, Chaiyachati K, Chimbindi N, Peoples A, Haberer J, Newell M-L. Interventions to increase antiretroviral adherence in sub-Saharan Africa: a systematic review of evaluation studies. Lancet Infect Dis. 2011;11:942–951. doi: 10.1016/S1473-3099(11)70181-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.McMahon JH, Jordan MR, Kelley K, Bertagnolio S, Hong SY, Wanke CA, et al. Pharmacy adherence measures to assess adherence to antiretroviral therapy: review of the literature and implications for treatment monitoring. Clin Infect Dis. 2011;52:493–506. doi: 10.1093/cid/ciq167. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Ashforth A. AIDS, religious enthusiasm and spiritual insecurity in Africa. Global Pub Health. 2011;6:S132–S147. doi: 10.1080/17441692.2011.602702. [DOI] [PubMed] [Google Scholar]
- 45.Thielman NM, Ostermann J, Whetten K, Whetten R, Itemba D, Maro V, et al. Reduced adherence to antiretroviral therapy among HIV-Infected Tanzanians seeking cure from the Loliondo Healer. J Acquir Immune Defic Syndr. 2014;65:e104–e109. doi: 10.1097/01.qai.0000437619.23031.83. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 46.Audet CM, Blevins M, Rosenberg C, Farnsworth S, Salato J, Fernandez J, et al. Symptomatic HIV-positive persons in rural Mozambique who first consult a traditional healer have delays in HIV testing: a cross-sectional study. J Acquir Immune Defic Syndr. 2014;66:e80–e86. doi: 10.1097/QAI.0000000000000194. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 47.Audet CM, Sidat M, Blevins M, Moon TD, Vergara A, Vermund SH. HIV knowledge and health-seeking behavior in Zambezia Province, Mozambique. J Soc Asp HIV/AIDS. 2012;9:41–46. doi: 10.1080/17290376.2012.665257. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Okome-Nkoumou M, Okome-Miame F, Kendjo E, Obiang G-P, Kouna P, Essola-Biba O, et al. Delay between first HIV-related symptoms and diagnosis of HIV infection in patients attending the internal medicine department of the Fondation Jeanne Ebori (FJE), Libreville, Gabon. HIV Clin Trials. 2005;6:38–42. doi: 10.1310/ULR3-VN8N-KKB5-05UV. [DOI] [PubMed] [Google Scholar]
- 49.Musheke M, Bond V, Merten S. Individual and contextual factors influencing patient attrition from antiretroviral therapy care in an urban community of Lusaka, Zambia. J Int AIDS Soc. 2012;15:1–9. doi: 10.7448/IAS.15.3.17366. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 50.Gourlay A, Birdthistle I, Mburu G, Iorpenda K, Wringe A. Barriers and facilitating factors to the uptake of antiretroviral drugs for prevention of mother-to-child transmission of HIV in sub-Saharan Africa: a systematic review. J Int AIDS Soc. 2013;16:18588. doi: 10.7448/IAS.16.1.18588. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51.Littlewood RA, Vanable PA. A global perspective on complementary and alternative medicine use among people living with HIV/AIDS in the era of antiretroviral treatment. Curr HIV/ AIDS Rep. 2011;8:257–268. doi: 10.1007/s11904-011-0090-8. [DOI] [PubMed] [Google Scholar]
- 52.George G, Chitindingu E, Gow J. Evaluating traditional healers knowledge and practices related to HIV testing and treatment in South Africa. BMC Int Health Hum Rights. 2013;13:45. doi: 10.1186/1472-698X-13-45. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 53.Nakimuli-Mpungu E, Bass JK, Alexandre P, Mills EJ, Musisi S, Ram M, et al. Depression, alcohol use and adherence to antiretroviral therapy in sub-Saharan Africa: a systematic review. AIDS Behav. 2012;16:2101–2118. doi: 10.1007/s10461-011-0087-8. [DOI] [PubMed] [Google Scholar]
- 54.Hendershot CS, Stoner SA, Pantalone DW, Simoni JM. Alcohol use and antiretroviral adherence: review and meta-analysis. J Acquir Immune Defic Syndr. 2009;52:180–202. doi: 10.1097/QAI.0b013e3181b18b6e. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 55.Katz IT, Ryu AE, Onuegbu AG, Psaros C, Weiser SD, Bangsberg DR, et al. Impact of HIV-related stigma on treatment adherence: systematic review and meta-synthesis. J Int AIDS Soc. 2013;16:18640. doi: 10.7448/IAS.16.3.18640. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 56.Patel R, Ratner J, Gore-Felton C, Kadzirange G, Woelk G, Katzenstein D. HIV disclosure patterns, predictors, and psychosocial correlates among HIV positive women in Zimbabwe. AIDS Care. 2012;24:358–368. doi: 10.1080/09540121.2011.608786. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 57.Alonzo AA, Reynolds NR. Stigma, HIV and AIDS: an exploration and elaboration of a stigma trajectory. Soc Sci Med. 1995;41:303–315. doi: 10.1016/0277-9536(94)00384-6. [DOI] [PubMed] [Google Scholar]
- 58.Takada S, Weiser SD, Kumbakumba E, Muzoora C, Martin JN, Hunt PW, et al. The dynamic relationship between social support and HIV-related stigma in rural Uganda. Ann Behav Med. 2014;48:26–37. doi: 10.1007/s12160-013-9576-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 59.Ware NC, Idoko J, Kaaya S, Biraro IA, Wyatt MA, Agbaji O, et al. Explaining adherence success in sub-Saharan Africa: an ethnographic study. PLoS Med. 2009;6:e1000011. doi: 10.1371/journal.pmed.1000011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 60.O’Laughlin KN, Wyatt MA, Kaaya S, Bangsberg DR, Ware NC. How treatment partners help: social analysis of an African adherence support intervention. AIDS Behav. 2012;16:1308–1315. doi: 10.1007/s10461-011-0038-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 61.Mills EJ, Lester R, Ford N. Promoting long term adherence to antiretroviral treatment. BMJ. 2012;345:10. doi: 10.1136/bmj.e4173. [DOI] [PubMed] [Google Scholar]
- 62.Nachega JB, Uthman OA, Anderson J, Peltzer K, Wampold S, Cotton MF, et al. Adherence to antiretroviral therapy during and after pregnancy in low-income, middle-income, and high-income countries: a systematic review and meta-analysis. AIDS. 2012;26:2039–2052. doi: 10.1097/QAD.0b013e328359590f. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 63.Matthews LT, Heffron R, Mugo NR, Cohen C, Hendrix C, Celum C, et al. High medication adherence during periconception periods among HIV-1-uninfected women participating in a clinical trial of antiretroviral preexposure prophylaxis. J Acquir Immune Defic Syndr. 2014;67:91–97. doi: 10.1097/QAI.0000000000000246. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 64.Merten S, Kenter E, McKenzie O, Musheke M, Ntalasha H, Martin-Hilber A. Patient-reported barriers and drivers of adherence to antiretrovirals in sub-Saharan Africa: a meta-ethnography. Trop Med Int Health. 2010;15:16–33. doi: 10.1111/j.1365-3156.2010.02510.x. [DOI] [PubMed] [Google Scholar]
- 65.Tsai AC, Bangsberg DR, Weiser SD. Harnessing poverty alleviation to reduce the stigma of HIV in sub-Saharan Africa. PLoS Med. 2013;10:e1001557. doi: 10.1371/journal.pmed.1001557. [DOI] [PMC free article] [PubMed] [Google Scholar]