Predictors of Nonadherence to Antiretroviral Therapy among HIV-Infected Adults in Dar es Salaam,Tanzania

Abstract

Background:

Adherence rates of ≥95% to antiretroviral therapy (ART) are necessary to maintain viral suppression in HIV-infected individuals. We identified predictors of nonadherence to scheduled antiretroviral drug pickup appointments in a large HIV care and treatment program in Tanzania.

Methods:

We performed a prospective cohort study of 44, 204 HIV-infected adults on ART between November 2004 and September 2012. Multivariate generalized estimating equation for repeated binary data was used to estimate the relative risk and 95% confidence intervals of nonadherence.

Results:

Nonadherence was significantly greater among patients with high CD4 counts, high body mass indices, males, younger patients, patients with longer durations on ART, and those with perceived low social support.

Conclusions:

Targeted interventions should be developed to improve ART adherence among healthier, younger, and more experienced patients who are on ART for longer durations within HIV care and treatment programs. Social support for patients on ART should be emphasized.

Keywords

HIV/AIDS antiretroviral treatment nonadherence Tanzania

Introduction

Access to antiretroviral therapy (ART) was limited in sub-Saharan Africa until a decade ago when scale-up was possible with financial and technical assistance from various governments and international donor agencies.¹ As in other parts of the world, the wider use of combination ART has transformed AIDS into a chronic treatable condition for a larger proportion of people living with HIV (PLHIV) in sub-Saharan Africa (SSA).² The increased longevity for PLHIV, however, is accompanied by concerns about ART nonadherence that increases with longer duration of therapy. Policy makers and public health specialists have warned of the necessity to monitor adherence to ART to avoid widespread drug resistance and a potential dramatic increase in per-capita cost of maintaining patients on ART who would be using second-line treatment.³

An intake of ≥95% of prescribed antiretroviral (ARV) drug is defined as optimal adherence to ART. Optimal adherence is needed to achieve sustained viral suppression and avoid drug resistance and keeping as many people as possible on first-line regimens.^4

–7 The methods of assessing nonadherence are diverse,^8,9 sometimes cumbersome to implement and may not actually correlate with immunologic or virologic failure.¹⁰ In addition, factors associated with nonadherence to ART vary by social, cultural, and local contexts.¹¹ Studies of the predictors of nonadherence to ART among HIV-positive patients in areas of high HIV/AIDS burden are crucial to determine local factors associated with adherence to ART and to assist with the development and implementation of programs that maximize adherence. Measuring adherence to ART using methods such as pill counts and self-reports is difficult in resource-limited settings.¹² In addition, these methods are limited to the assessment of adherence among those who are adherent to visit schedules. Self-reported measures of adherence and pill count have been found to be highly correlated with regular clinic attendance within 3 days of schedule.⁸ In turn, adherence to medication and consistency with clinic visits has been shown to be significantly associated with improved clinical outcomes in patients living with HIV/AIDS in these settings.^8,13

–16 Consistent clinic attendance to a scheduled ART refill appointment has been suggested as a relatively easy way to identify nonadherent patients on ART over time in resource-poor settings.^13,16,17 In this study, we assessed the incidence of adherence to ART in a cohort of HIV-infected patients in a large urban Tanzanian HIV Care and Treatment program using adherence to clinic ART refill visits as a surrogate measure. Therefore, using on-time visit for ART drug pickup as a metric of adherence, we sought to identify modifiable risk factors associated with nonadherence.

Methods

Study Population

In this prospective cohort study, we included HIV-infected adult (>15 years old) patients on ART within the urban public HIV care and treatment clinics (CTC) supported by Management and Development for Health (MDH) with support from the US President’s Emergency Plan for AIDS Relief (PEPFAR) in the Dar es Salaam region of Tanzania. All HIV-infected adults who had been initiated on ART and attended at least 2 scheduled ART pickup/refill visits at 1 of 29 CTCs in Dar es Salaam city clinics between November 2004 and September 2012 were included in this study. All patients enrolled in MDH supported CTCs receive ART and cotrimoxazole which is provided free of charge. Patients were enrolled to participate in this study following a written informed consent which was subject to ethical review by the Muhimbili University of health and Allied Sciences Review Board and the Harvard School of Public Health Institutional review board.

Patients ART Follow-up Visits Plan

At enrollment, patients were evaluated for ART eligibility and initiated on ART according to Tanzanian National guidelines. Patients were assessed at baseline by a physician and, if eligible, were initiated ART 2 to 4 weeks later. Antiretroviral drugs are provided at no cost by the Tanzanian government. At the time of this study, standard first-line ART regimens included stavudine (d4T) or zidovudine (ZDV), plus lamivudine (3TC) and efavirenz (EFV) or nevirapine (NVP). Patients are followed every 28 days and at each visit they undergo a physical examination and assessment by a physician, pick up 30 days’ supply of ART, and meet with a counselor to discuss adherence, dosing, and ART side effects. CD4 counts, complete blood count, and alanine transferase (ALT) levels are performed every 6 months. Viral loads are not routinely available at MDH-supported facilities. During the study enrollment period, no patients experienced treatment interruptions owing to drug shortages.

Definition of Outcomes

Since ≥95% adherence to ART intake is required for optimal viral suppression, and because of the high reported correlation with scheduled ART pickup visits,^13,16 we defined nonadherence as ≥5% noncompliance with scheduled ART pickup visits. In our study clinics, patients were given scheduled appointments for ART drug refill every 28 days and each ART refill included a 30-day supply of drugs. Because patients received 2 extra pills for each on-time refill, patients were classified as nonadherent for a visit if they exceeded more than 3 days beyond their recorded appointment date or if the intervisit interval was ≥35 days in cases where there was no record of a future appointment date. The ≥35 days intervisit interval definition of nonadherence was chosen to accommodate the fact that a given patient can have adequate drug coverage for a maximum of 34 days if that patient was exactly on time for their most recently past drug pickup appointment (allowing for the 2 extra pills) but presented up to 2 days late for their next appointment.

Data Collection

Health care providers from the study clinics completed standard case report forms to capture sociodemographic, clinical, laboratory, and therapeutic information at all baseline and follow-up visits for each patient who was identified by a unique study identification number. Dedicated data reviewers stationed at respective clinics ensured adequacy and completeness of the recorded data by the health care workers. Data were collected daily and stored in a secured computerized database that was updated daily by dedicated data entry clerks. Weekly quality assurance and quality control checks of the database were performed by the data management team to ensure accuracy. Data collected included sociodemographic characteristics, body mass index (BMI kg/m²), date of first confirmed HIV positive test, history of TB, current TB treatment, use of cotrimoxazole, pregnancy, ART regimen at initiation and follow-up, ART regimen type, and next appointment date. Laboratory data collected at baseline and at 6-month intervals included hemoglobin, CD4 counts, and ALT. Additional information on patients’ socioeconomic and psychosocial characteristics including depression, stigma, social support, employment, per-capital food expenditure, and household belongings and others were available in a subset of 3620 patients who were involved in a Trial of Vitamin 3 study conducted to examine the effects of multivitamins (including B, C, and E) on HIV disease progression among HIV-positive Tanzanian adults on ART. The socioeconomic and psychosocial variables were categorized in as follows: participants were assessed for depressive symptoms using the Hopkins Symptom Checklist and a cutoff score of >1.06 was used to define depression.¹⁸ The social support score and stigma scores were grouped by tertiles, and household belongings score and per-capita daily food expenditure were grouped by quartiles with the following cut points: number of household belongings (≤2/>2–3/>3–5/>5), social support score (≤2.5/>2.5–3.625/>3.625), stigma score (1/>1–1.167/>1.167), and per-capita household expenditure (Tanzanian Shillings [Tsh] ≤625 /≤Tsh857/≤Tsh1250/>Tsh1250). High meat intake was defined as eating meat at least 3 to 4 times a week.

Statistical Analysis

We used mean and standard deviation (SD) for continuous variables, and proportions were used to describe basic characteristics of the study population at the time of enrollment. Generalized estimating equation (GEE) models assuming an exchangeable working correlation matrix and the Poisson distribution with the log link were used to estimate relative risks (RRs) and 95% confidence intervals (CIs) for ≥5% patient nonadherence with scheduled ART refills.¹⁹ Multivariate analyses identified the independent effect of each risk factor, adjusting liberally for all other measured potential risk factors including sociodemographics, clinical and immunologic characteristics, district of residence, year of ART initiation, duration on ART initiation, type of ART regimen, and season of consultation. Variables that were associated with nonadherence at a P value ≤.20 in univariate models were further considered in multivariate models.²⁰ Statistical analyses were conducted using SAS version 9.1 statistical software (SAS Institute, Cary, North Carolina). Ethical clearance for this study was provided by the institutional review boards of the Harvard School of Public Health and Muhimbili University of Health and Allied Sciences.

Results

Of 47, 010 HIV-infected patients who were enrolled at and initiated on ART enrolled at MDH-supported CTCs, a total of 44, 204 patients had 2 or more visits and were included in this study. Patients were followed for an average of 1.8 (SD = 1.5) years from ART initiation and 2.2 (SD = 1.8) years from enrollment. At ART initiation, the mean age was 37.5 (SD = 9.59) years, 64% were <40 years old and 69% were females. The mean CD4 counts and BMI at enrollment were 141 (SD = 128) cells/mm³ and 21 (SD = 4.5) kg/m², respectively. Most patients were ARV naive prior to enrollment at MDH (93%), were anemic or severely anemic (64%), and had World Health Organization clinical stage 3 or 4 disease (76.7%) at ART initiation; 53% were initiated on ART regimen that contained EFV (50%) and 46% received a regimen that contained ZDV or d4T (Table 1).

Table 1.

Characteristics of Patient Population at ART Initiation.^a

Variable	Mean (SD)/%
Demographic variables
Age, years	37.7 (9.6)
<30	19
30–<40	44
40–<50	25
≥50 years	11
Sex
Female	69
Male	31
District of residence
Ilala	42
Kinondoni	32
Temeke	26
Clinical variables
CD4 counts, cells/mm³	141 (128)
<100	43
100–199	34
200–349	20
≥350	4
WHO clinical stage
I	8
II	16
III	56
IV	21
AIDS-defining symptom	21
Year of ART initiation
2004	0.1
2005	7
2006	11
2007	16
2008	19
2009	19
2010	17
2011	10
Previous antiretroviral drug use
No	93
Yes	7
ARV drugs for PMTCT only	0.04
NRTI component
ZDV (+3TC + [EFV or NVP])	56
d4T (+3TC + [EFV or NVP])	40
TDF (+ [3TC or FTC] + [EFV or NVP])	3
ABC (+3TC + [EFV or NVP])	0.3
NNRTI component
EFV (+ [ZDV or d4T or TDF or ABC] + [3TC or FTC])	53
NVP (+ [ZDV or d4T or TDF or ABC] + [3TC or FTC])	47
History of TB	27
On treatment for TB	12
Hypertension	6
Hemoglobin, mg/dL	10.2 (2.2)
<8.5	21
8.5–10.9	43
≥ 11.0	36
Alanine transaminase, IU/L	26 (21)
Body mass index, kg/m²	21 (4.5)
<18.5	30
18.5–<25	54
25–<30	12
≥30	4
Socioeconomic variables (n = 3 620)
Married	43
Employment
Unemployed	28
Not unemployed	72
Per-capita daily food expenditure (TSh)
≤625	27
626–857	24
858–1250	25
>1250	24
High meat intake
Household belongings score	26
Very low	43
Low	11
High	20
Very high	25
Psychosocial variables (n = 3620)
Social support score
Low	27
Medium	27
High	47
Stigma score
Low	65
Medium	25
High	9
Depressed	57
Shared HIV-test result with anyone outside the study	93
At least one reason given for nondisclosure of HIV status	35
At least one social problem reported from being given the HIV-test result	9
At least one social problem reported from being HIV positive	7

Abbreviations: TSh, Tanzanian Shilling; WHO, World Health Organization; ART, antiretroviral treatment; PMTCT, prevention of mother-to-child transmission of HIV; ZDV, zidovudine; EFV, efavirenz; NVP, nevirapine; d4T, stavudine; TDF, tenofovir; 3TC, lamivudine; FTC, emtricitabine; TB, tuberculosis; ABC, abacavir; SD, standard deviation.

^a N = 44 204.

Eighty-one percent of patients were compliant with ≥95% of follow-up visits and 96% of patients were compliant with ≥70% of visits. Several factors at ART initiation were independently associated with ≥5% nonadherence to visits following ART initiation, including younger age (RR among patients younger than 30 years compared with those ≥50 years = 1.07, 95% CI 1.05–1.09), advanced disease stage (RR for stage IV compared with stage I = 1.04, 95% CI 1.01–1.07), and high BMI (RR for ≥ 30 kg/m² compared with <18.5 = 1.09, 95% CI 1.07–1.12). Attendance at clinics in the Temeke district (RR = 1.11, 95% CI 1.09–1.12) was associated with higher nonadherence compared to the Ilala district. Men were at higher risk of nonadherence (RR = 1.10, 95% CI 1.09–1.12) as compared to nonpregnant women, whereas pregnant women were at significantly lower risk for nonadherence (RR = 0.53; 95% CI 0.43–0.64).

The risk of nonadherence was independently associated with later calendar years of the program, even after adjusting for duration on ART (P < .0001). Furthermore, nonadherence to clinic visits increased with longer duration on ART, despite adjustment for calendar year of ART initiation. Specifically, the RR of nonadherence rose from 1.33 (95% CI: 1.31–1.35) for patients on ART for 6 to 12 months to 2.62 (95% CI: 2.55–2.69) for patients on ART for ≥48 months (Table 2).

Table 2.

Sociodemographic and Clinical Variables in Relation to the Risk of ART Nonadherence.

Sociodemographic variables	Proportion of Nonadherent Interval Between Visits	Univariate RR (95% CI)	P Value	Multivariate^a RR (95% CI)	P Value
Current age, yrs			<.0001		<.0001
<30	15.77	1.07 (1.05–1.09)		1.07 (1.05–1.1)
30–39	14.64	1.00		1.00
40–<50	13.88	0.94 (0.93–0.96)		0.93 (0.92–0.95)
≥50	13.41	0.91 (0.89–0.93)		0.89 (0.87–0.91)
Sex/current pregnancy status
Nonpregnant female	14.38	1.00		1.00
Pregnant female	8.02	0.45 (0.37–0.55)	<.0001	0.53 (0.43–0.64)	<.0001
Male sex	14.90	1.03 (1.02–1.05)	<.0001	1.10 (1.09–1.12)	<.0001
Site/district at ART initiation			<.0001		<.0001
Ilala	14.37	1.00		1.00
Kinondoni	14.12	0.98 (0.96–0.99)		1.01 (0.99–1.02)
Temeke	15.21	1.07 (1.05–1.09)		1.11 (1.09–1.12)
Married at ART initiation			.10		.36
No	14.46	1.00		1.00
Yes	14.72	1.01 (1.00–1.03)		0.99 (0.98–1.01)
Clinical variables
Current CD4 counts, cells/mm³			<.0001		<.0001
<100	12.69	0.66 (0.65–0.68)		0.97 (0.95–0.99)
100–199	14.58	0.81 (0.80–0.82)		1.05 (1.03–1.07)
200–349	14.96	0.91 (0.90–0.92)		1.04 (1.03–1.06)
350+	14.86	1.00		1.00
WHO clinical stage at ART initiation (−90-14 days)			<.0001		.04
I	15.23	1.00		1.00
II	14.86	0.97 (0.94–1.00)		0.98 (0.95–1.01)
III	14.49	0.95 (0.92–0.97)		1.01 (0.98–1.04)
IV	14.16	0.94 (0.92–0.97)		1.04 (1.01–1.07)
AIDS-defining symptom at ART initiation			.29
No	14.41	1.00
Yes	14.98	1.01 (0.99–1.03)
Year of ART initiation			<.0001		<.0001
2004-2005	12.34	1.00		1.00
2006	12.37	0.99 (0.96–1.02)		1.03 (0.99–1.08)
2007	13.56	1.03 (1.00–1.07)		1.24 (1.19–1.28)
2008	15.60	1.17 (1.14–1.21)		1.49 (1.43–1.55)
2009	17.61	1.32 (1.28–1.36)		1.80 (1.73–1.88)
2010	18.13	1.34 (1.30–1.38)		2.01 (1.93–2.10)
2011	13.08	0.95 (0.91–1.00)		1.59 (1.50–1.67)
ART duration, months			<.0001		<.0001
<6 months on ART	11.49	1.00		1.00
6–12 months on ART	14.57	1.32 (1.30–1.34)		1.33 (1.31–1.35)
>12–24 months on ART	15.25	1.46 (1.44–1.48)		1.51 (1.49–1.54)
>24–36 months on ART	16.14	1.66 (1.64–1.68)		1.83 (1.79–1.86)
>36–48 months on ART	17.02	1.87(1.84–1.90)		2.17 (2.12–2.22)
>48 months on ART	18.44	2.13 (2.10–2.17)		2.62 (2.55–2.69)
History of ART use at initiation			<.0001		.03
No	14.69	1.00		1.00
Yes	13.21	0.95 (0.92–0.97)		0.97 (0.94–1.00)
Current NRTI component
ZDV (+ 3TC + [EFV or NVP])	15.93	1.00		1.00
d4T (+ 3TC + [EFV or NVP])	13.62	0.84 (0.82–0.85)	<.0001	0.99 (0.98–1.01)	.35
TDF (+ [3TC or FTC] + [EFV or NVP])	18.66	1.14 (1.11–1.17)	<.0001	1.05 (1.02–1.07)	.002
ABC (+ 3TC + [EFV or NVP])	12.87	0.84 (0.74–0.96)	.009	0.79 (0.70–0.90)	.0004
Current NNRTI component			.03		<.0001
EFV (+ [ZDV or d4T or TDF or ABC] + [3TC or FTC])	14.60	1.00		1.00
NVP (+ [ZDV or d4T or TDF or ABC] + [3TC or FTC])	15.48	1.02 (1.00–1.03)		0.94 (0.93–0.96)
Current body mass index, kg/m²			<.0001		<.0001
<18.5	13.23	0.89 (0.87–0.90)		0.96 (0.94–0.97)
18.5–<25	14.33	1.00		1.00
25–<30	15.07	1.11 (1.09–1.12)		1.05 (1.03–1.06)
≥ 30	16.20	1.21 (1.18–1.23)		1.09 (1.07–1.12)
TB history at ART initiation (−30–0 days)			<.0001		.002
No	14.76	1.00		1.00
Yes	13.93	0.96 (0.94–0.97)		0.97 (0.96–0.99)
Current TB treatment at ART initiation (0–30 days)			.27
No	14.53	1.00
Yes	14.29	0.99 (0.97–1.01)
Current hypertension (systolic BP > 90 mm Hg)			<.0001		<.0001
No	14.20	1.00		1.00
Yes	11.60	0.88 (0.86–0.90)		0.88 (0.86–0.90)
Current hemoglobin, mg/dL			<.0001		.04
<8.5	13.97	0.84 (0.82–0.86)		1.00 (0.98–1.02)
8.5–10.99	14.24	0.92 (0.91–0.93)		0.99 (0.98–1.01)
11+	14.65	1.00		1.00
Current alanine transferase, IU/L			.36
>40	14.27	1.00
≤40	14.26	0.99 (0.97–1.01)

Abbreviations: ART, antiretroviral therapy, ZDV, zidovudine; EFV, efavirenz; NVP, nevirapine; d4T, stavudine; TDF, tenofovir; TB, tuberculosis; RR, relative risk; CI, confidence interval; yrs, years; FTC, emtricitabine; 3TC, lamivudine; WHO, World Health Organization; NRTI, nucleoside reverse transcriptase inhibitor; ABC, abacavir; BP, blood pressure.

^a The multivariate model also adjusted for employment status, reason for nondisclosure of HIV status, social problems from HIV status, high meat intake, depressed, social support score, household score, and per-capita daily food expenditure.

Antiretroviral regimens containing tenofovir (TDF) as the nucleoside reverse transcriptase inhibitor base (RR 1.05, 95% CI: 1.02–1.07, P = .002) were associated with higher risk of nonadherence. However, those on abacavir (ABC) containing regimens had a significant lower risk of nonadherence relative to regimen containing d4T, ZDV, or TDF (RR 0.79, 95% CI: 0.70–0.90, P = .0004). The risk of nonadherence to clinic visits was significantly lower for patients initiated on the nonnucleoside reverse transcriptase inhibitors (NNRTIs) containing NVP (RR = 0.94, 95% CI: 0.93–0.96, P value < .0001) than those on an EFV containing regimen. Anemia status, tuberculosis disease history, and history of ART at time of initiation were not independently associated with >5% nonadherence to clinic visits (Table 2).

Nonadherence to clinic visits was highest among patients with perceived low social support (RR: 0.91, 95% CI: 0.86–0.95). The majority of patients were not employed (72%) and their average per-capita daily food expenditure was Tsh1093, equivalent to less than 1 US dollar per day. Although employment was not associated with the risk of being nonadherent, high meat intake (P = .05) was found to be independently associated with the risk of being nonadherent to clinic visits. Disclosure of HIV status, depression, and stigma were not significantly independently associated with nonadherence (Table 3).

Table 3.

Psychosocial and Socioeconomic Status at Time of ART Initiation, in Relation to the Risk of ART Nonadherence.

	Proportion of Nonadherent Intervals Between Visits	Univariate RR (95% CI)	P Value	Multivariate ^a RR (95% CI)	P Value
Depressed			.11		.42
No	14.13	1.00		1.00
Yes	13.54	0.96 (0.92–1.01)		0.98 (0.93–1.03)
High meat intake			.003		.05
No	13.58	1.00		1.00
Yes	14.44	1.08 (1.03–1.13)		1.05 (1.00–1.11)
Employment			.17		.46
Unemployed	13.67	1.00		1.00
Employed	13.94	1.04 (0.98–1.09)		0.98 (0.93–1.03)
Per-capita daily food expenditure^b			<.0001		.1
Very low	12.77	1.00		1.00
Low	13.80	1.08 (1.01–1.16)		1.05 (0.98–1.12)
High	14.24	1.11 (1.03–1.18)		1.06 (0.99–1.13)
Very high	14.66	1.14 (1.07–1.22)		1.07 (1.00–1.15)
Household belongings score^b			.14		.18
Very low	13.65	1.00		1.00
Low	13.52	0.99 (0.92–1.06)		0.97 (0.90–1.04)
High	13.75	1.01 (0.95–1.07)		1.00 (0.94–1.06)
Very high	14.11	1.05 (0.99–1.11)		1.02 (0.97–1.08)
Social support score^c			<.0001		<.0001
Low	15.08	1.00		1.00
Medium	13.05	0.87 (0.82–0.92)		0.89 (0.83–0.94)
High	13.52	0.89 (0.85–0.94)		0.91 (0.86–0.95)
Stigma score^c			.81
Low	13.85	1.00
Medium	13.50	0.97 (0.93–1.03)
High	14.33	1.04 (0.96–1.13)
Shared HIV status with others			.33
No	14.12	1.00
Yes	13.77	0.96 (0.88–1.04)
Reason for nondisclosure of HIV status			.05		.07
None	13.26	1.00		1.00
≥1 reason given	13.93	1.06 (1.00–1.13)		1.06 (1.00–1.13)
Social problem from being given HIV test result			.36
No	13.76	1.00
Yes	14.07	1.04 (0.96–1.12)
Social problem from HIV status			.14		.06
No	13.73	1.00		1.00
Yes	14.59	1.07 (0.98–1.17)		1.09 (1.00–1.19)

Abbreviations: RR, relative risk; CI, confidence interval; WHO, World Health Organization; NRTI, nucleoside reverse transcriptase inhibitor; ART, antiretroviral therapy; TB, tuberculosis; NNRTI, nonnucleoside reverse transcriptase inhibitor.

^a The multivariate model also adjusted for age, body mass index, current CD4 counts, current NRTI component, current NNRTI component, history of ART use at initiation, current hemoglobin, sex/current pregnancy status, year of ART initiation, ART duration (months), year of ART initiation, site/district at ART initiation, married at ART initiation, WHO stage at ART initiation, TB history at ART initiation, and current hypertension.

^b Quartiles.

^c Tertiles.

Discussion

This 6-year follow-upstudy is based upon one of the largest cohorts of patients on ART in Africa including 44 204 HIV-infected adults on treatment. Most patients (78%) had advanced AIDS at ART initiation as defined by CD4 counts <200 cells/mm³.²¹ We noted that optimal ART adherence to on-time drug pickup occurred among 81% of the patients consistent with adherence levels reported in other studies using on-time drug pickup visits^13,16 as well as in studies using other methods to assess adherence including pill count^3,22 and self-report.⁷ We observed that nonadherence is greatest among healthier patients as indicated by CD4 counts, disease stage, and BMI, suggesting that these patients perceive less concern about their disease as they get healthier^3,12,23 and pointing to a possible serious gap in patients’ awareness of the health implications of nonadherence to ART in absence of clinical symptoms.

Interestingly, young patients and male patients had a significantly higher risk of nonadherence, consistent with the findings from another study by Poles et al of a smaller cohort of 809 HIV-infected patients within a similar setting.¹⁵ However, other smaller studies with less than 500 participants did not find significant associations of nonadherence with age and sex.^22,24–25 Nonadherence was significantly higher among patients from the district of Temeke, which is the most rural district in Dar es Salaam, whose population is more dispersed with lower socioeconomic status, hence hindering access to quality health services. Long distances with unreliable transport to clinic, as well as rural locations, have been found to be associated with noncompliance to ART in Uganda and Cote d’Ivoire.^26,27

Patients with low socioeconomic status have difficulty in accessing services, especially when they need to pay for regular transport. Studies in Tanzania and Uganda observed that poor access to ART services is common when people need to choose between buying food and paying for transport, coupled with the belief that taking ARV drugs with inadequate food intake could result in severe adverse effects.^28,29 Although patients in this study were accessing free services from public facilities, most of them would have to pay for transport to reach their respective clinics. Although provision of free services has been perceived to increase access and promote adherence, it does not guarantee consistent access to ART services among individuals with limited financial resources. To improve adherence, services should be provided closer to the community and possibly integrated in primary health services.

We noted that nonadherence was significantly higher among patients who were prescribed d4T- and TDF-containing regimens compared to ZDV-containing regimens. Stavudine is known to produce undesirable side effects, including lipodystrophy which likely leads to patient noncompliance. Several studies provided consistent findings in this regard, leading to recommendations to remove d4T from the ARV regimen.^30
–32 We also observed that TDF was independently associated with nonadherence, despite the fact that TDF produces minimal side effects and hence is less likely to be associated with nonadherence.³³ Tenofovir was introduced in Tanzania in 2009 as an alternative to d4T and ZDV and therefore may have been more likely to be used in patients who were switching from d4T because of undesirable side effects and/or nonadherence. We also observed patients on EFV to be more nonadherent than those on NVP-based regimens. These findings are contrary to those from other studies which report no significant difference for adherence and viral suppression between participants on different NNRTIs.³⁴

The risk of nonadherence with on-time ART refill increased significantly with increased time since ART initiation and was 75% greater after 4 or more years on ART. This suggests that as patients became healthier, they may become more risk taking and nonadherent. Furthermore, we noted that patients who were initiated on ART in recent years were more likely to be nonadherent. This finding may be attributed to the fact that the existing clinics are getting more congested in recent years, possibly resulting in a poorer quality of ART services within our context. Future studies should examine the relationship of specific domains of quality of care and relationships to adherence.

In some programs in which ART has to be purchased by the patient, the inability to pay for ART was the most important barrier to consistent ART intake and nonadherence.²⁸ However, since ART was provided for free in our program, we observed that unemployment was not a significant predictor of ART nonadherence, consistent with the findings from studies in a similar setting in Uganda in which ART services were also provided for free.²⁴ Depression and stigma were also not found to be associated with nonadherence in this study, which was similar to our earlier findings but differed from an older study of nonadherence conducted in South Africa.²¹ As HIV Care and Treatment programs in SSA have grown, societies have become more accepting of providing care and support to relatives and friends with HIV/AIDS, which may have resulted in less stigma-related nonadherence. On the other hand, depression has been associated with decline of CD4 counts and higher mortality which may be due to noncompliance.^11,25,32 In this study, in contrast, depression was not associated with nonadherence.

There were several limitations that may have influenced our findings. We have used a measure of nonadherence to drug pickup visits which could have missed patients who shifted to other clinics and were enrolled as new patients. We did not measure viral load, therefore we were unable to correlate adherence with virologic outcomes.

In conclusion, we noted that nonadherence is low in this urban Tanzania HIV-infected population. However, a quarter of the population still had suboptimal adherence, underscoring the urgency of strengthening adherence initiatives in clinics. Such initiatives could target patients with higher risk, especially the young, relatively healthier individuals, patients with higher socioeconomic status, and males. As treatment programs are being rapidly scaled-up into more rural areas, it is crucial to take into account the socioeconomic status of the general population and tailor initiatives to intervene on predictors that hinder optimal adherence. Further research is needed on the relationship between quality of care and treatment adherence, general health, and survival.

Footnotes

Authors’ Note

This study is done under the support of the President’s Emergency Plan for AIDS Relief (PEPFAR) through the US Centers for Disease Control and Prevention (CDC). The findings and conclusions in this article are those of the authors and do not necessarily represent the views of the US Centers for Disease Control and Prevention. Use of trade names is for identification only and does not imply endorsement by the US Centers for Disease Control and Prevention or the US Department of Health and Human Services.

Acknowledgements

The authors are grateful for the great collaboration of the United States government offered through the US President’s Emergency Plan for AIDS Relief (PEPFAR) and to the Government of Tanzania through its Ministry of Health and Social Welfare (MoHSW) in supporting the implementation of the HIV care and treatment program in Tanzania. We are also thankful to all patients who participated in this study, and to health care providers from the participating clinics, and the 3 municipals health management teams.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

UNAIDS. Global Report: UNAIDS Report on the Global AIDS Epidemic 2010. Geneva, Switzerland: Joint United Nations Programme on HIV/AIDS; 2010:20.

Effros

Fletcher

Gebo

. Aging and infectious diseases: workshop on HIV infection and aging: what is known and future research directions. Clin Infect Dis. 2008;47(4):542–553.

Altice

Mostashari

Friedland

. Trust and the acceptance of and adherence to antiretroviral therapy. J Acquir Immune Defic Syndr. 2001;28(1):47–58.

Amberbir

Woldemichael

Getachew

Girma

Deribe

. Predictors of adherence to antiretroviral therapy among HIV-infected persons: a prospective study in Southwest Ethiopia. BMC Public Health. 2008;8:265.

Berg

Litwin

Heo

Arnsten

. Directly observed antiretroviral therapy improves adherence and viral load in drug users attending methadone maintenance clinics: a randomized controlled trial. Drug Alcohol Depend. 2011;113(2–3):192–199.

Carrieri

Cailleton

Le Moing

. The dynamic of adherence to highly active antiretroviral therapy: results from the French National APROCO cohort. J Acquir Immune Defic Syndr. 2001;28(3):232–239.

Crespo-Fierro

. Compliance/adherence and care management in HIV disease. J Assoc Nurses AIDS Care. 1997;8(4):43–54.

Ross-Degnan

Pierre-Jacques

Zhang

. Measuring adherence to antiretroviral treatment in resource-poor settings: the clinical validity of key indicators. BMC Health Serv Res. 2010;10:42.

Diabate

Alary

Koffi

. Short-term increase in unsafe sexual behaviour after initiation of HAART in Cote d'Ivoire. AIDS. 2008;22(1):154–156.

10.

Chaiyachati

Hirschhorn

Tanser

Newell

Barnighausen

. Validating 5 questions of antiretroviral nonadherence in a public-sector treatment program in rural South Africa. AIDS Patient Care STDS. 2011;25(3):163–170.

11.

Reynolds

Testa

Marc

. Factors influencing medication adherence beliefs and self-efficacy in persons naive to antiretroviral therapy: a multicenter, cross-sectional study. AIDS Behav. 2004;8(2):141–150.

12.

Oyugi

Byakika-Tusiime

Charlebois

. Multiple validated measures of adherence indicate high levels of adherence to generic HIV antiretroviral therapy in a resource-limited setting. J Acquir Immune Defic Syndr. 2004;36(5):1100–1102.

13.

Bastard

Pinoges

Balkan

Szumilin

Ferreyra

Pujades-Rodriguez

. Timeliness of clinic attendance is a good predictor of virological response and resistance to antiretroviral drugs in HIV-infected patients. PLoS One. 2012;7(11):e49091.

14.

Nyandiko

Vreeman

Liu

. Nonadherence to clinic appointments among HIV-infected children in an ambulatory care program in Western Kenya. J Acquir Immune Defic Syndr. 2013;63:e49–e55.

15.

Poles

Siril

. Factors associated with different patterns of nonadherence to HIV care in Dar es Salaam, Tanzania. J Int Assoc Provid AIDS Care (Chic). 2014;13(1):78–84.

16.

El-Khatib

Katzenstein

Marrone

. Adherence to drug-refill is a useful early warning indicator of virologic and immunologic failure among HIV patients on first-line ART in South Africa. PLoS One. 2011;6(3):e17518.

17.

Samet

Libman

Steger

. Compliance with zidovudine therapy in patients infected with human immunodeficiency virus, type 1: a cross-sectional study in a municipal hospital clinic. Am J Med. 1992;92(5):495–502.

18.

Kaaya

Fawzi

Mbwambo

Lee

Msamanga

Fawzi

. Validity of the Hopkins Symptom Checklist-25 amongst HIV-positive pregnant women in Tanzania. Acta Psychiatr Scand. 2002;106(1):9–19.

19.

Fitzmaurice

GM LN

Ware

. Applied Longitudinal Analysis. John Wiley & Sons; 2004.

20.

Maldonado

Greenland

. Simulation study of confounder-selection strategies. Am J Epidemiol. 1993;138(11):923–936.

21.

Chalamilla

Hawkins

Okuma

. Mortality and treatment failure among HIV-infected adults in Dar Es Salaam, Tanzania. J Int Assoc Physicians AIDS Care (Chic). 2011.

22.

Nachega

Stein

Lehman

. Adherence to antiretroviral therapy in HIV-infected adults in Soweto, South Africa. AIDS Res Hum Retroviruses. 2004;20(10):1053–1056.

23.

Bangsberg

Acosta

Gupta

. Adherence-resistance relationships for protease and non-nucleoside reverse transcriptase inhibitors explained by virological fitness. AIDS. 2006;20(2):223–231.

24.

Byakika-Tusiime

Oyugi

Tumwikirize

Katabira

Mugyenyi

Bangsberg

. Adherence to HIV antiretroviral therapy in HIV+ Ugandan patients purchasing therapy. Int J STD AIDS. 2005;16(1):38–41.

25.

Singh

Squier

Sivek

Wagener

Nguyen

. Determinants of compliance with antiretroviral therapy in patients with human immunodeficiency virus: prospective assessment with implications for enhancing compliance. AIDS Care. 1996;8(3):261–269.

26.

Lubega

Musenze

Joshua

. Sex inequality, high transport costs, and exposed clinic location: reasons for loss to follow-up of clients under prevention of mother-to-child HIV transmission in eastern Uganda - a qualitative study. Patient Prefer Adherence. 2013;7:447–454.

27.

Diabate

Alary

Koffi

. Determinants of adherence to highly active antiretroviral therapy among HIV-1-infected patients in Cote d'Ivoire. AIDS. 2007;21(13):1799–1803.

28.

Kumarasamy

Safren

Raminani

. Barriers and facilitators to antiretroviral medication adherence among patients with HIV in Chennai, India: a qualitative study. AIDS Patient Care STDS. 2005;19(8):526–537.

29.

Mshana

Wamoyi

Busza

. Barriers to accessing antiretroviral therapy in Kisesa, Tanzania: a qualitative study of early rural referrals to the national program. AIDS Patient Care STDS. 2006;20(9):649–657.

30.

Musheke

Bond

Merten

. Individual and contextual factors influencing patient attrition from antiretroviral therapy care in an urban community of Lusaka, Zambia. J Int AIDS Soc. 2012,15(suppl 1):1–9.

31.

Menezes

Maskew

Sanne

Crowther

Raal

. A longitudinal study of stavudine-associated toxicities in a large cohort of South African HIV infected subjects. BMC Infect Dis. 2011;11:244.

32.

Ickovics

Meade

. Adherence to HAART among patients with HIV: breakthroughs and barriers. AIDS Care. 2002;14(3):309–318.

33.

Moyle

Sabin

Cartledge

. A randomized comparative trial of tenofovir DF or abacavir as replacement for a thymidine analogue in persons with lipoatrophy. AIDS. 2006;20(16):2043–2050.

34.

Bangsberg

Hecht

Charlebois

. Adherence to protease inhibitors, HIV-1 viral load, and development of drug resistance in an indigent population. AIDS. 2000;14(4):357–366.