Retention and predictors of attrition among patients who started antiretroviral therapy in Zimbabwe's national antiretroviral therapy programme between 2012 and 2015.

BACKGROUND: The last evaluation to assess outcomes for patients receiving antiretroviral therapy (ART) through the Zimbabwe public sector was conducted in 2011, covering the 2007-2010 cohorts. The reported retention at 6, 12, 24 and 36 months were 90.7%, 78.1%, 68.8% and 64.4%, respectively. We report findings of a follow-up evaluation for the 2012-2015 cohorts to assess the implementation and impact of recommendations from this prior evaluation.
METHODS: A nationwide retrospective study was conducted in 2016. Multi-stage proportional sampling was used to select health facilities and study participants records. The data extracted from patient manual records included demographic, baseline clinical characteristics and patient outcomes (active on treatment, died, transferred out, stopped ART and lost to follow-up (LTFU)) at 6, 12, 24 and 36 months. The data were analysed using Stata/IC 14.2. Retention was estimated using survival analysis. The predictors associated with attrition were determined using a multivariate Cox regression model.
RESULTS: A total of 3,810 participants were recruited in the study. The median age in years was 35 (IQR: 28-42). Overall, retention increased to 92.4% (p-value = 0.060), 86.5% (p-value<0.001), 79.2% (p-value<0.001) and 74.4% (p-value<0.001) at 6, 12, 24 and 36 months respectively. LTFU accounted for 98% of attrition. Being an adolescent or a young adult (15-24 years) (vs adult;1.41; 95% CI:1.14-1.74), children (<15years) (vs adults; aHR 0.64; 95% CI:0.46-0.91), receiving care at primary health care facility (vs central and provincial facility; aHR 1.23; 95% CI:1.01-1.49), having initiated ART between 2014-2015 (vs 2012-2013; aHR1.45; 95%CI:1.24-1.69), having WHO Stage IV (vs Stage I-III; aHR2.06; 95%CI:1.51-2.81) and impaired functional status (vs normal status; aHR1.25; 95%CI:1.04-1.49) predicted attrition.
CONCLUSION: The overall retention was higher in comparison to the previous 2007-2010 evaluation. Further studies to understand why attrition was found to be higher at primary health care facilities are warranted. Implementation of strategies for managing patients with advanced HIV disease, differentiated care for adolescents and young adults and tracking of LTFU clients should be prioritised to further improve retention.

Introduction

Globally by the end of 2018, there were 37.9 million [32.7 million–44.0 million] people living with HIV (PLHIV) with 61% of these residing in Eastern and Southern Africa (ESA) [1]. Over the past two to three decades, investments in the global HIV response have achieved unprecedented results with the number of new HIV infections significantly reduced from2.9 million [2.3 million–3.8 million] in 1997 to 1.7 million [1.6 million–2.3 million] new infections by 2018and 23.3million PLHIV put on treatment globally [1]. Between 2010 and 2018 new HIV infections and deaths decreased by 16% and 33%, respectively [2]. Despite these remarkable achievements, patient attrition and losses to follow-up (LTFU) still remain legitimate threats to the long-term success of antiretroviral therapy (ART) scale up [3].

PLHIV on ART who are not retained in care are at increased risk of developing drug resistance and dying [4]. Based on a systematic review of several studies that have been conducted in low resource settings, key predictors of high attrition include patients with advanced HIV-disease progression [marked by body mass index (BMI) <18 kg/m2, baseline CD4 counts <200 cells/mL, World Health Organisation (WHO) Stage—III and IV, poorer level of functionality], male sex, younger age and having lower levels of education [5]. Early ART initiation as measured by shorter time duration between HIV testing and ART initiation has been shown to reduce risk of attrition [6,7]. Other studies have however shown that those who initiate ART at a higher baseline CD4 may also be prone to attrition [8,9].

Numerous studies assessed attrition and the effect of different initiatives on treatment outcomes at selected health facilities in resource-limited settings [7,8,10-12]. They mainly reflect the experience of academic, standalone, donor supported or private institutions that generally have better data collection systems, are well financed and have better human resources for health, which may not be generalizable. In scenarios where outcomes for routine programmes have been reported, evaluations have often been regional, limited to a few facilities or targeting only a specific subpopulation of the HIV cohorts (children, adolescents or adults) [6,13-17]. There is paucity of data demonstrating treatment outcomes and impact of ART at a national level in many resource-limited settings. Periodic national treatment outcome evaluations provide information needed for targeted interventions that will allow national HIV programmes to achieve the ambitious UNAIDS 3rd 90 by 2020.

Zimbabwe started rolling out ART in the public health system in 2004 in 5 pilot facilities. Since then, there has been significant scale-up with more than 1500 health institutions offering ART by the end of 2017 [18]. This rapid scale-up has been mostly attributed to the rapid adaptation of the WHO HIV guidelines and the widespread decentralization of comprehensive HIV services which was supported by health care worker task-shifting policies and significant investments in training, supportive supervision and clinical mentoring [19]. Funding from the Government of Zimbabwe, Global Fund for AIDS, Tuberculosis, and Malaria (GFATM), the United States of America President’s Emergency Plan for AIDS Relief (PEPFAR) and other donors supported the scale-up the program.

An evaluation of the Zimbabwe national ART programme, which only included adults (> 15 years) living with HIV started on ART between 2007–2010, reported 90.7%, 78.1%, 68.8% and 64.4% retention at 6, 12, 24 and 36 months, respectively [17]. These findings mirrored treatment outcomes in other parts of sub-Saharan Africa and recommendations given included strengthening earlier diagnosis and linkages to treatment; further decentralization of comprehensive HIV services to improve ART coverage and adaptation of innovative strategies aiming at improving patient retention (adherence clubs, food supplementation and mobile short messages service (SMS) reminders) [17].

In our study, we report results of a 2016 follow-up evaluation of treatment outcomes of the Zimbabwe National ART Programme, conducted among individuals who started ART between 2012–2015, prior to the implementation of the HIV ‘Treat All’ policy. This nationwide follow-up evaluation therefore assessed the impact of the interventions implemented since the first evaluation. The assessment was expanded to include children and adolescents. We expected an increase in retention on ART following interventions recommended after the previous evaluation. The aim was to estimate retention, compare it with the previous evaluation, and explore predictors of attrition (either lost to follow-up or died).

Material & methods

Study design

A retrospective cohort analysis was undertaken among children, adolescents and adults living with HIV who started ART in Zimbabwe between October 2012 and January 2015.

Study setting

The study was conducted at selected public health institutions across all the country’s 10 provinces. Zimbabwe has a population of around 13 million [20]. The country has a generalised HIV epidemic and prevalence has continued to hover between 13% and 16% for the past decade. Currently there are about 1.4 million PLHIV of which 1.0 million (71%) were on (ART) by December 2018 [18]. Since 2004 the country has adopted all the successive changes in the WHO recommendations to start ART. Patients were eligible to start ART when they had a CD4 count <350 cells/mL in 2010, CD4 <500 cells/mL in 2013 and since 2016, patients are eligible regardless of their CD4 count (Treat All). The country has made significant progress towards achieving the 90-90-90 targets. In a recent survey, 74.2% reported knowing their HIV status; 86.8% self-reported current use of ART and among those who self-reported current use of ART, 86.5% were virally suppressed [21].

During the period between 2011 and 2015, the Zimbabwe National AIDS Strategic Plan II (ZNASP-II) 2011–15 was developed to guide the scale-up of HIV care and treatment services towards universal access [22]. In 2013, the country adopted the 2013 WHO guidelines which recommended the CD4 < 500 cells/mL threshold for ART initiation. A ‘test and treat’ approach was adopted for all HIV-positive children under 5 years, TB/HIV co-infected, HBV/HIV co-infected, the HIV-positive partner in HIV sero-discordant relationship and pregnant and breastfeeding mothers (Option B+). The preferred first-line regimen for adults, adolescents, and older children was changed from stavudine, lamivudine and nevirapine (d4T/3TC/NVP) to a once-daily pill of tenofovir, lamivudine and efavirenz (TDF/3TC/EFV) with zidovudine and nevirapine as alternatives for TDF and EFV respectively. After ART initiation stable patients were reviewed every 3 months and drugs were dispensed directly from the health facilities. Monitoring was mainly clinical (weight, WHO clinical stage and assessment of opportunistic infections) at every visit, complemented with laboratory tests (CD4 testing) every 6 months. Compared to the previous period access to CD4 testing improved significantly. By the end of 2015, laboratory based CD4 testing was available in each of the 63 districts as compared to 47 in 2009 [23]. On top of laboratory based CD4 testing, more than 265 point of care (POC) CD4 testing devices were procured and distributed throughout the country mainly to support the roll out of Option B+ [24]. Blood collection and transportation systems were put in place to support facilities without access to CD4 testing. Routine viral load testing was not available in the public sector and there was limited access to targeted viral load testing for patients with suspected treatment failure. The monitoring and evaluation system was mainly paper based. A few high-volume facilities started to implement an electronic patient monitoring system (ePMS) which operated concurrently with the usual paper-based system.

Sample size and sampling criteria

A minimum sample of 4000 patient charts was required to estimate 12-month ART retention (the outcome of interest) after assuming: 50% ART retention at 12 months after initiation, 20% of charts would be missing, a margin of error of 5% and a design effect of 2.0. The design effect of 2 was used to correct the loss of effective sample which was anticipated from intra-stratum correlation. For logistical and financial feasibility purposes, sampling was restricted to 1,389 ART sites across all 10 provinces that were supporting ≥50 HIV-positive clients on ART for at least 12 months by January 2016. The sites were stratified into seven strata according to their patient volumes receiving ART care and ordered by province and district within each stratum. From these, 70 (5%) ART sites were sampled using a probability proportional to size sampling criteria based on client volumes receiving ART at each site. At the sampled sites, a line-listing was generated in Excel for all HIV positive children, adolescents, adults and pregnant women initiated on ART at the sampled sites between 1 October 2012 and 31 January 2015 using their unique ART numbers obtained from the facility ART register. Following this, the required number of patient ART care booklets were randomly selected without replacement according to their stratum based on random schedule generator in Stata. If a selected ART care booklet was not traceable, the next record was traced until the required sample size by site was reached.

Study variables and treatment outcomes

The data were collected manually from the patient files. The data collected included demographics (age, sex and marital status), clinical and programmatic variables. Baseline clinical parameters collected at point of ART initiation included WHO stage, functional status and pregnancy. Programmatic data collected included level of care, date of HIV testing, date of enrolment and date of ART initiation. Date of last clinic visit, date of next scheduled clinic visit and date of transfer-out/death/stopping ART were also collected and used to determine the patient status (i.e. active on treatment, LTFU, dead, transferred out and stopped ART). The patient status was determined on the date of data abstraction, regardless of previous treatment interruptions. Patients were considered LTFU if they were more than 180 days without visit at the clinic on the date of data abstraction. If a patient was reported as LTFU in the patient file, data collectors checked pharmacy refill records to confirm LTFU. Refill dates were considered as visit dates.

Data collection procedures

Ten teams of three data abstractors collected the data using a structured tool. The data abstractors underwent a 5-day training programme to acquaint them with study tools and study procedures. Data collection tools were piloted to identify weaknesses which then were rectified. The primary source document for the study was the patient manual medical records (Patient OI/ART Booklets) being kept at the health institution. Other sources to complement or validate the information included the different registers and interviews with at least one experienced health care worker at each sampled site. Data was collected using Open Data Kit (ODK) software on Android devices. Data was downloaded at regular intervals for cleaning, quality control checks, merging and backing up. Data from a sample of 10% of the selected charts was re-abstracted as a quality control check measure. Discrepancies were addressed by both the team leader and data abstractors.

Data analysis

The data were analysed using Stata/IC 14.2 [25]. Our sample size calculation was adjusted to cater for missing data. Patients without outcome data were excluded. Explanatory variables which had more than 30% missingness were also excluded from the analysis. We assumed missingness to be at random. Categorical variables which had less than 30% missingness were included in the analysis (Table 1). We used the missing indicator method, showing the proportion missingness for the different variables (if any). The primary outcome was attrition, and this was defined as either being dead, LTFU or stopped ART.

Table 1

Baseline characteristics of 3810 patients who started ART between 2012–2015, in Zimbabwe.

	N	(%)
Total	3810	100.0
Year of ART initiation
2012	289	(7.6)
2013	1476	(38.7)
2014	1850	(48.6)
2015	195	(5.1)
Sex
Female	2262	(59.4)
Male	1548	(40.6)
Median Age (N = 3700) (IQR)	35 (28–42)
Age groups
Children (0–15 years)	254	(6.7)
Adolescents & young adults (15-24years)	385	(10.1)
Adults (25–49 years)	2735	(71.8)
Elderly (>=50 years)	424	(11.1)
Missing	12	(0.3)
Pregnancy at ART initiation
Confirmed	208	(5.5)
Not confirmed	3602	(94.5)
Marital status of patient
Single & divorced	863	(22.7)
Married	2125	(55.8)
Widowed	434	(11.4)
N/A Child	318	(8.3)
Missing	70	(1.8)
Baseline Functional Status
Impaired	811	(21.3)
Normal	2796	(73.4)
Missing	203	(5.3)
Baseline WHO Stage
WHO Stage I	821	(21.5)
WHO Stage II	1245	(32.7)
WHO Stage III	1447	(38.0)
WHO Stage IV	120	(3.1)
Unknown	177	(4.6)
Level of Care
Central Hospital	888	(23.3)
Provincial Hospital	341	(9.0)
Mission/District Hospitals	1549	(40.7)
Primary Health Care facilities	1032	(27.1)

N—Number of observations, ART—Antiretroviral Therapy, IQR—Interquartile Range, N/A—Not applicable, WHO—World Health Organisation.

A multivariable Cox proportional hazard model was used to compute crude and adjusted hazard ratios and their 95% confidence intervals estimating the association between explanatory variables and attrition. The proportional hazards assumption was assessed using log-log plots and Kaplan-Meier versus Cox predicted plots and results of these analyses suggested that the proportional hazards assumption holds. Time to attrition was calculated as the time between the date of ART initiation and the date of outcome or last patient clinic visit. Those active on ART and transferred out were censored on the date of data abstraction and the date they were transferred out respectively. Kaplan-Meier techniques were used to estimate retention on ART at 6, 12, 24 and 36 months. To identify independent risk factors for attrition, first variables were selected into the multivariable model based using a stepwise forward approach (p-value < 0.1). Moreover, sex and level of care were included based on their clinical and biological relevance in HIV retention. Second, stepwise backward elimination was used until all remaining variables were significantly associated with attrition (p-value < 0.05). In an additional analysis we estimated the correlation between year of ART initiation and attrition until 12 months of ART. Patients who started ART in 2015 were excluded since their maximal follow-up time was too short (S1 Table).

Ethical considerations

The evaluation protocol received ethics approval from Medical Research Council of Zimbabwe (MRCZ/A/2033). The protocol was also sanctioned by the MoHCC. Confidentiality and anonymity of ART clients was protected use their unique ART numbers as no names were abstracted. Permission to conduct the evaluation was sought at various levels of the tiered health delivery system. All data abstractors signed confidentiality forms.

Results

A total of 3,993 (99.8%) records were abstracted out of a target of 4,000. On data cleaning 3,810 (95.4%) records were found to be of quality standard for the analysis (Fig 1).

Fig 1

Study population.

Of 3,810 patients included, 38.7% and 48.6% started ART in 2013 and 2014, respectively, and 59.4% were female. The median age in years was 35 [interquartile range (IQR):28–42]. Most patients were adults (71.8%), married (55.8%), and had a normal functional status (73.4%). Few (3.1%) were in WHO stage IV, and 27.1% started ART at a primary health care (PHC) facility. At time of ART initiation 5.5% of women were confirmed to be pregnant (Table 1).

Fig 2 shows that the proportion of patients receiving ART at PHC facilities increased gradually from 21.6% in 2012 to 34.5% in 2015.

Fig 2

Level of care where the 3,810 patients started ART between 2012–2015, in Zimbabwe.

Overall, retention at 6, 12, 24 and 36 months was 92.4%, 86.5%, 79.2% and 74.4%, respectively. Table 2 shows a comparison between adult patients who started ART between 2007–2010 and 2012–2015. Retention at 6, 12, 24 and 36 months was 1.3%, 7.8%, 9.9% and 9.3% higher among patients who started between 2012–2015. Differences were statistically significant except at 6 months.

Table 2

Retention in care of patients who started antiretroviral therapy between 2007–2010 and 2012–2015 in Zimbabwe.

	Months since ART initiation	2012–2015 Evaluation			2007–2010 Evaluation			Difference
		N	Retention (%)	95% CI	N	Retention (%)	95% CI	%	95%CI	aP value
Overall	6	3476	92.4	(91.5–93.2)
	12	3066	86.5	(85.3–87.5)
	24	1427	79.2	(77.7–80.6)
	36	222	74.4	(72.5–76.1)
Children < 15 years	6	250	97.3	(94.4–98.7)
	12	230	94.2	(90.5–96.4)
	24	114	86.3	(81.0–90.2)
	36	17	82.3	(75.5–87.4)
Adults > 15 years	6	3222	92.0	(91.1–92.9)	3739	90.7	(86.1–93.8)	1.3	(-0.02–2.6)	0.060
	12	2832	85.9	(84.7–87.0)	3641	78.1	(67.7–84.7)	7.8	(5.9–10.0)	<0.001
	24	1311	78.7	(77.1–80.1)	2003	68.8	(58.5–77.5)	9.9	(6.9–12.9)	<0.001
	36	204	73.7	(71.8–75.6)	806	64.4	(55.7–72.3)	9.3	(2.4–16.2)	<0.001

a: Two-sample test of proportions

The maximum follow-up time for the patients who started ART in 2012, 2013, 2014 and 2015 was 40.4, 37.2, 25.2 and 13.4 months respectively (S1 Table). We assessed 12-month retention by year of ART initiation with those who started ART in 2015 excluded. Retention decreased during the study period. Those who started ART in 2013 and 2014, as compared to those who started ART in 2012, were 2.18 and 3.78 times at risk of attrition at 95% confidence interval (CI), respectively (Fig 3).

Fig 3

Comparison of retention rates of patients who started antiretroviral therapy in 2012, 2013 and 2014. in Zimbabwe.

Over 6489 years of patient follow-up; 77.4% were active on treatment, 2.4% transferred out and 20.2% were lost through attrition (LTFU 19.8%, and died 0.4%) The median follow-up time per patient was 1.7 years (IQR 1.1–2.4). The overall attrition rate was 11.8, [95% confidence interval (CI):11.0–12.7] per 100 person years (PY). The other attrition rates are shown in Table 3.

Table 3

Bivariate and multivariate analysis of factors associated with attrition in the 2012–2015 ART cohort.

Variable	Categories	Total	PT years	Attrition$	Attrition %	Attrition Per 100PY	HR	p	95%CI	aHR	P	95%CI
Total		3810	6489	768	20.2	11.8
Sex	Female	2262	3804	440	19.5	11.6	1
	Male	1548	2685	328	21.2	12.2	1.07	0.377	(0.92–1.23)		NS	NS
Age group	Adults	2735	4650	555	20.3	11.9	1			1
	Children	254	467	35	13.8	7.5	0.64	0.009	(0.45–0.89)	0.64	0.011	(0.46–0.91)
	Adolescents & young adults	385	562	103	26.8	18.3	1.47	<0.001	(1.19–1.82)	1.41	0.002	(1.14–1.74)
	Elderly	424	784	75	17.7	9.6	0.81	0.096	(0.64–1.04)	0.83	0.123	(0.65–1.05)
	Missing	12
Marital Status	Married	2125	3619	419	19.7	11.6	1
	Single & divorced	863	1382	202	23.4	14.6	1.25	0.010	(1.05–1.47)
	Widowed	434	772	85	19.6	11.0	0.97	0.766	(0.76–1.22)		NS	NS
	Child N/A	318	584	49	15.4	8.4	0.74	0.042	(0.55–0.99)
	Missing	70
Level of Care	Central & province	1229	2119	230	18.7	10.9	1			1
	Primary health care facility	1032	1658	213	20.6	12.8	1.15	0.134	(0.96–1.39)	1.23	0.039	(1.01–1.49)
	District	1549	2712	325	21.0	12.0	1.11	0.242	(0.93–1.31)	1.21	0.032	(1.01–1.44)
Year ART Initiation	2012–13	1765	3994	377	21.4	9.4	1			1
	2014–2015	2045	2495	391	19.1	15.7	1.45	<0.001	(1.24–1.69)	1.45	<0.001	(1.24–1.69)
WHO Stage	I-III	3513	5944	675	19.2	11.4	1			1
	IV	120	194	43	35.8	22.2	1.95	<0.001	(1.43–2.66)	2.06	<0.001	(1.51–2.81)
	Missing	177
Functional status	Normal	2796	4774	538	19.2	11.3	1			1
	Impaired	811	1320	177	21.8	13.4	1.18	0.058	(0.99–1.40)	1.25	0.013	(1.04–1.49)
	Missing	203
Pregnant when starting ART	Not confirmed	3602	6219	718	19.9	11.5	1
	Confirmed	208	270	50	24.0	18.5	1.48	0.007	(1.12–1.98)		NS	NS

NS = Not significant (not included in final model); HR = Hazard Ratio; aHR = adjusted Hazard Ratio; CI = Confidence Interval; ART = Antiretroviral Therapy; WHO = World Health Organisation

$ either death or LTFU

Variables which had p -value < 0.1 (age group, marital status, year of ART initiation, WHO stage, functional status and pregnancy) were included in multivariable analysis. Sex and level of care were included based on their clinical and biological relevance in HIV retention. In multivariable analysis, being an adolescent or a young adult [multivariable-adjusted hazard ratio (aHR)(vs adult; 1.41; 95% CI:1.14–1.74], receiving care at PHC facility and district level (vs central and provincial facility; aHR 1.23; 95% CI:1.01–1.49) and aHR 1.21; 95% CI:1.01–1.44), having initiated ART between 2014–2015 (vs 2012–2013; aHR 1.45; 95% CI:1.24–1.69), having WHO Stage 4 (vs Stage I–III; aHR 2.06; 95% CI:1.51–2.81) and having an impaired functional status (vs normal status; aHR1.25; 95% CI:1.04–1.49) were associated with increased attrition in our setting. Risk of attrition among children (<15years) in our setting was about 40% lower(vs adults; aHR 0.64; 95% CI:0.46–0.91 (Table 3).

Discussion

This follow-up evaluation of the Zimbabwe National ART programme, which included children, adolescents and adults started on ART between 2012 and 2015, showed an overall improvement of the performance of the ART programme compared to those started on ART between 2007–2010 [17]. Attrition was mainly explained by LTFU. Retention observed in this evaluation was above what is reported for children, adolescents and adults in other similar studies from low-resource settings [26-28]. Adolescents and young adults, patients with advanced HIV disease (WHO Stage 4, impaired functional status), those receiving care at PHC level and starting treatment after 2013, when the country switched treatment guidelines to the 500 CD4 cells/mL threshold, were at risk of attrition.

The increase in retention since the previous evaluation may be explained by strategies put in place by the MoHCC. Relying on a vast body of evidence showing that decentralization increases access to ART and retention in care [29], the prior evaluation recommended further decentralization. Moreover, other innovative strategies to improve patient retention, such as adherence clubs, food supplementation and mobile short messages service (SMS) reminders were recommended [17]. Our data show a substantial increase of the proportion of patients receiving ART at PHC level. This decentralization was supported by policy shifts (in particular, the provision for nurses to initiate ART in non-complicated cases), and by significant investments in training, supportive supervision and clinical mentoring [30]. In addition to decentralization, the improvement in retention could also be attributed to adoption of WHO guidance for earlier ART initiation among those tested HIV-positive with CD4<500 cells/mL compared to the previous 350 cells/mL CD4 threshold. For instance in this cohort 41.1% had WHO stage 3 or 4 compared to 87.6% in the previous study [17]. However, other recommended measures, such as food supplementation, adherence clubs, mobile short messages service (SMS) reminders, fast tracking of stable patients on ART, community and family ART refill groups, were not implemented on a wide scale in the public sector.

However, the increase of decentralized ART does probably not explain the higher level of retention, compared to the previous evaluation. Surprisingly, our study showed that receiving ART at PHC level was associated with attrition. This finding contrasts with the findings of the previous evaluation. In the previous evaluation, retention among patients receiving ART at PHC level was better than among those initiating ART at higher levels of care, in particular district/mission hospitals [17]. We speculate that the higher level of attrition in PHC facilities may be explained by the massive decentralization (down referral), which happened during the current evaluation period, whereby patients who started ART at district/provincial hospitals were referred to PHC level for follow-up. Implementation of policy changes may be abrupt and have an adverse effect. After assessment by a healthcare worker, stable patients may have little or no opportunity to object against the decision to be referred to another clinic. Patients referred to a clinic which is not of their choice are more likely to self-transfer to another facility of their choice. Such patients are then considered as LTFU in one clinic, while retained in another clinic. Moreover, at the beginning of the decentralization process, care at PHC level may not have been fully developed. Another study showed that abrupt down-referral may lead to a decrease in quality of care and resulting in worse health outcomes [31]. When PHC facilities are ill prepared, usually outcome monitoring is poor. The country has almost completed the decentralization process for ART services and the focus now should be on improving the quality of care at PHC facilities. The necessary health system support structures which include human resources for health (recruitment, training and capacity building), drug supply, monitoring and evaluation should be prioritised.

Most of the reported attrition (98%) was due to LTFU. This finding was similar to the prior evaluation where LTFU also accounted for the larger proportion of attrition. Determining true LTFU is not always easy. A recent meta-analysis showed that, of patients LTFU and traced, 30% had self-transferred, 30% had stopped taking ART and the other 30% had died [32]. We therefore hypothesize that the increased LTFU in our study might be administrative, especially at PHC level where monitoring is less well developed. A substantial proportion of patients may seem LTFU, but be in care at another or even the same clinic with another identification number [31]. The probability of administrative reasons for LTFU is probably higher at PHC facilities, especially when paper-based tools are used to monitor and report treatment outcomes [31]. Poor documentation of clinic visits and transfers in medical records may result in administrative LTFU. There is need to determine the true nature of the outcome of patients LTFU within the Zimbabwe ART programme. A substantial proportion may be alive and receiving ART from another facility, while others may have died but were not reported as such [33,34]. This may be achieved through strengthening the current active patient tracking and tracing mechanisms [35]. The existing ART programme’s electronic patient monitoring systems (ePMS) should also be scaled up and optimised to bring efficiency in patient monitoring [36]. Tracing mechanisms include SMS reminders, phone calls and home visits by community health workers [37].

There was a change in the risk factors for attrition in comparison to the previous evaluation. Males were at risk of attrition in the prior evaluation but were not at risk in the current evaluation. Several studies have consistently showed males at high risk of attrition due to several reasons, including employment related constraints and poor health seeking behaviour leading to late presentation [13,15,38,39]. During the study period the country did not implement many specific interventions that targeted men. Decentralization of ART services may have also worked in favour of males for better retention as they could now access services at health facilities closer to where they reside. Another strategy that may have improved male ART uptake and retention was male involvement in PMTCT (Option B+), promoted during massive campaigns [19]. There is a need for further studies to assess whether this strategy suffices, or if other strategies targeting men are needed, such as flexible clinic hours to accommodate work, community-based ART delivery, and tracing of those who miss appointments [40,41].

The previous evaluation did not include data on children (< 15 years) but included adolescents (>=15 years) and young adults. Though age was included the previous evaluation did not assess the effect of age on attrition [17]. In our study we found children (< 15 years) to be at lower risk of attrition as compared to adults. Our findings seem to contrast with those from previous studies, which identified dependency of guardians, lack of palatable formulations, and challenges with disclosure as additional challenges experienced by children [42-44]. We speculate that the following interventions impacted positively on retention in our setting: PMTCT Option B+ targeting the mother-baby pair, integration of paediatric ART services with immunisation, decentralization of paediatric ART services, improved access to palatable paediatric ART formulations, and emphasis on secondary caregivers to support the primary caregivers with the child’s adherence in case of absence or other commitments [45-48].

We found that adolescents and young adults were more at risk of attrition when compared to adults. Our findings are consistent with other studies [49-52]. Adolescents and young adults have been shown to be at high risk of attrition due to several factors, which include lack of youth-friendly services, rigid scheduling not taking into account schooling, and unavailability of peer caregivers [53,54]. Addressing these challenges can lead to improvement in the retention of adolescents and young adults on ART. Locally, community adolescent treatment supporters (CATS) have been shown to improve retention among adolescents and scaling up of the initiative should be prioritised [55].

We found patients with advanced disease to be at higher risk of attrition. Patients with advanced HIV disease are prone to attrition mainly due to mortality and morbidity. The common causes of morbidity and mortality in low resource setting include cryptococcal meningitis, tuberculosis, sepsis, malignancy and wasting syndrome/chronic diarrhoea [56-58]. Advanced HIV disease was also a risk factor for attrition in the previous evaluation. Very minimal investment in building capacity at primary health facilities in the management of patients with advanced HIV disease contributes too poorer outcomes in this subgroup. Indeed, access to baseline CD4 testing, screening and management of opportunistic (tuberculosis, cryptococcal meningitis) and access to prophylaxis (isoniazid preventive therapy & pre-emptive fluconazole) remains a challenge in the Zimbabwe’s public sector [59,60]. Recent evidence has also shown that screening, prophylaxis and management of opportunistic diseases (mainly tuberculosis, bacterial sepsis and cryptococcal disease) significantly reduce morbidity and mortality [61,62]. However, evidence from this enhanced care package is currently being poorly implemented by most ART programmes in low resource setting [63]. There is a need to mobilise resources for training and capacity building of health workers, setting up the necessary infrastructure and procurement of the necessary commodities required in the management of patients with advanced disease. However, the process should be guided by a formal assessment of the current burden of advanced disease in the country.

We also found that patients who started ART in 2013 or later were at risk of attrition as compared to those who started in 2012. We speculate that this could be explained by decentralization and the fact that patients were now starting ART at a higher CD4 count. Decentralization happened gradually between 2012–2015. By the end of 2015 more patients were receiving care at PHCs, and our study found attrition to be higher at PHCs. The number of patients who were being initiated and followed-up on ART at PHCs increased during the period (Fig 1). We speculate that the gradual increase of patients starting ART at the possibly ill-prepared PHCs might have contributed to the increase in attrition over time during the study period. In 2013, the country switched ART guidelines to the 500 CD4 cells per cubic millilitres threshold from the 350 cells per cubic millilitres cut-off which was used in 2012. On top of the change in the CD4 cut-off point, test and treat was also introduced to specific sub-populations (children under 5 years, TB/HIV co-infected, HBV/HIV co-infected, the HIV-positive partner in HIV sero-discordant relationship and pregnant and breastfeeding mothers (Option B+). Similar findings were reported in previous studies, which showed that patients who started ART at a higher CD4 were at risk of attrition [8,9]. Patients starting ART at high CD4 are less sick and have low risk perception which might affect adherence to long term therapy [64]. This then calls for earnest implementation of the current country operational guidelines which recommends adequate psychosocial preparation and readiness assessment before ART initiation under Treat All [37].

Our study was a follow-up evaluation of the Zimbabwe ART programme. To our knowledge, no studies showing national data have been followed-up by a second study of similar magnitude. In a recent review, none of the studies included was a follow-up to a prior evaluation [5]. Despite the highlighted strengths our study had limitations. Most of the reported attrition was due to LTFU. Patients who are classified as LTFU might have been alive and still on ART (self-transferred), stopped ART or may have died [32]. Our LTFU definition (180 without visit) differed from the one used by the program (LTFU = 90 days late), due to data availability. However, considering that patients received up to 3 months ART refill, both definitions would result in similar findings. We could not report on immunologic and virologic patient outcomes due to missing information in the patient manual medical records. Information on viral load testing and CD4 testing was more than 30% missing. These variables were excluded. To preserve all observations for variables with less than 30% missingness the missing indicator method was used. We assumed that missingness was at random, and mainly explained by health workers inconsistently completing the medical records. However, the method can affect the residual variance of the regression [65]. Moreover, we did not assess differences between clinics of the same level of care.

Conclusion

Zimbabwe’s ART program shows good retention across different patient groups. The ART retention increased since the previous 2011 evaluation. However, adolescents and young adults, patients with advanced HIV disease (WHO Stage 4, impaired functional status), receiving care at PHC level and starting treatment after 2013 when the country switched treatment guidelines to the 500 CD4 cells per cubic millilitres threshold were at risk of attrition.

From the findings we recommend research into the following areas: reasons for higher attrition at PHC facilities, feasibility of screening for advanced disease at primary health facilities and ensuring access to referral clinical care, assessing retention within levels of care, qualitative research to explore the causal link between attrition in recent years, and being less sick at the start of treatment. To improve monitoring electronic patient monitoring systems should be prioritised to aid patient tracking. There is also a need for differentiated care strategies for adolescents and young adults to improve retention. Creation of youth-friendly services, flexible scheduling of visits and expansion of peer caregivers/treatment supporters should be considered.

Patient Retention, Clinical Outcomes and Attrition-Associated Factors of HIV-Infected Patients Enrolled in Zimbabwe’s National Antiretroviral Therapy Programme, 2007–2010.

(PDF)

Click here for additional data file.

Follow-up time of the 3810 patients who started ART in Zimbabwe between 2012–2015.

(DOCX)

Click here for additional data file.

22 Oct 2019

PONE-D-19-23547

Retention and predictors of attrition among patients who started antiretroviral therapy in Zimbabwe’s National Antiretroviral Therapy Programme between 2012 and 2015

PLOS ONE

Dear Dr Makurumidze,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

We would appreciate receiving your revised manuscript by 30th November 2019. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Kwasi Torpey, MD PhD MPH

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at http://www.journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and http://www.journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. We note that you have stated that you will provide repository information for your data at acceptance. Should your manuscript be accepted for publication, we will hold it until you provide the relevant accession numbers or DOIs necessary to access your data. If you wish to make changes to your Data Availability statement, please describe these changes in your cover letter and we will update your Data Availability statement to reflect the information you provide.

3. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

Additional Editor Comments (if provided):

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: No

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Review

Retention and predictors of attrition among patients who started antiretroviral therapy in Zimbabwe’s National Antiretroviral Therapy Programme between 2012 and 2015

Dear authors,

thank you for submitting this interesting paper on factors influencing retention of the ART program in Zimbabwe.

I do have the following questions and remarks regarding your methodology and interpretation:

Sample size:

Your assumptions regarding the sample size does not include the comparison between the retention rates at 12 months between the actual and the former study.

Reading the section it seems you need 4000 people to estimate a single rate – but then you should maybe give the broadness of a confidence interval to motivate this sample size.

What do you mean by design effect 2.0?

Sampling criteria:

You sampled from 1.10.2012 to 31.12.2015. The analysis was done in 2016?

When I see this information I conclude that in the sub-cohort of those persons starting the ART-Program in 2014/15 there is nobody that can have a retention until 36 months (if starting 2014) and 24 months (if starting in 2015). I don’t find any information at the minimum/maximum length of observation in the (sub-) cohort, could you please add this.

Statistical analysis:

Line 203-5

“Time to attrition was calculated as the time between the date of ART initiation and the date of outcome or last patient clinic visit. Those active on ART and transferred out were censored on the date of data abstraction and the date they were transferred out.”

I don’t understand the second sentence. How can you be censored to two dates?

- As you don’t have information on the time of censoring you could think of using the AFT-Model (Accelerated-Failure-Time-Modell) instead of the Cox-Model. In the AFT-Model you don’t need a date but a period in which an event/censoring occurs.

- Did you check for the proportional hazards assumption of the Cox-Model?

- I suggest to look into subgroups of patients conditioning on the time having been in the ART-program:

e.g. the retention in those patients already having been in ART for 6 months/for 12 months. Maybe the risk of attritions is time dependent and decreases over time.

- Fig 3:

In the method section you describe an hierarchical variable selection for a multivariable model. I would have preferred a more knowledge-based model building – if at all. Could you please mention all the variables selected? I guess the reader is supposed to find out in Fig 3 – but I find bivariate p-values <0.1 and no effect estimate of this variable in the multivariable section – for example in the case of marital status. (what does the N.S. – non significant mean in this situation then?).

Acutally I don’t think you need the multivariable analysis as you want to describe risk groups for higher attrition. The risk groups are well described already in the bivariable analysis. A multivariable model can be used as a prediction model – but this is not what you wanted or what is wanted in dealing with ART-patients, I guess.

In the results section you discuss the single risk factors in comparison to the former analysis – you don’t refer to the a prediction.

- Fig 2 and Fig 3

Please refrain from reporting the p-values besides your main effect (please see the paper ‘Scientists rise up against statistical significance’ in Nature 2019 https://www.nature.com/articles/d41586-019-00857-9)

Fig 3:

- In the sub group 2012/2013 there is much more PT compared to the patients included in 2014/2015 yielding in lower attrition rate/PT.

I think this is because your sampling until 12/2015 limits the potential observation time differentially and therefore fewer persons can be present at 12 months or later in the later cohort. How can somebody in the later subcohort be observed for 36 months? The observation time for somebody starting in 2015 is <12 months.

- You state that the difference in attrition rate between patients starting in the two periods is because of the changing criteria for ART-therapy regarding the cell count. in the introduction and conclusion you write that the change in this criterium happened in 2013.

According to Fig 1 most of the patients in the cohort 2012/2013 should have had the same conditions for inclusion into the program than during 2014/2015.

Other:

- Maybe this paper could be of interest for you as well:

Alhaj M et al. Retention on antiretroviral therapy during Universal Test and Treat implementation in Zomba district, Malawi: a retrospective cohort study. J Int AIDS Soc 2019 Feb;22(2):e25239

- Please check the syntax of the sentences in lines 283 and 390

- Loss-To-Follow-Up not abbreviated correctly all the time – please check!!

Reviewer #2: This is a large retrospective study evaluating the risk of LTFU among individuals who started ART between 2012-2015 in Zimbabwe. We commend the efforts of working with program data to produce this manuscript and support it’s publication.

I believe this study will benefit from the following comments:-

1. The authors should be more specific about the purpose of this study, the comparisons and the hypotheses.

2. How would patients who were considered LTFU (>180 days since without a visit) but show up at the clinic over 6 months later were treated in the analysis? I.e. are there some cases that were LFTU based on this criteria but showed up after 6 months has lapsed?

3. Could the other program activities (Line 294) explain lower retention in the 2014-2015 cohort compared to 2012-2013? (Fig 3). Did these activities diminish over time? I recall reading that more individuals initiated ART over time so does it mean that although ART was scaled up, the retention was lower. Please correct my misinterpretations and kindly explain these results.

4. Discuss program adherence and retention interventions (if any) and consider citing a paper by Coker et al., Curr HIV/AIDS Reports 2015. Socio-Demographic and Adherence Factors Associated with Viral Load Suppression in HIV-Infected Adults Initiating Therapy in Northern Nigeria: A Randomized Controlled Trial of a Peer Support Intervention that highlights that indicate that ART adherence will improve significantly regardless of whether HIV-infected adults received adherence interventions or standard of care services.

Please see attachment for additional comments that could not fit in here.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: Yes: Modupe Coker

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

Submitted filename: Letter PLOS_submitted20191017.doc

Click here for additional data file.

7 Dec 2019

Dear Editor,

Many thanks for the review of “Retention and predictors of attrition among patients who started antiretroviral therapy in Zimbabwe’s National Antiretroviral Therapy Programme between 2012 and 2015". We have carefully read and discussed the recommendations made by the reviewers, and have adapted the paper accordingly.

A marked-up copy of the manuscript highlighting changes made to the original version labeled 'Revised Manuscript with Track Changes' and a clean version labeled 'Manuscript' have been uploaded.

In this response letter we respond to each point raised by the reviewers. Please find our responses below with revisions or additions to the manuscript in quotes. We shall of course be available to respond to any additional question you may have.

Kind regards,

Richard Makurumidze, on behalf of the co-authors

Submitted filename: Response to Reviewers.docx

Click here for additional data file.

18 Dec 2019

Retention and predictors of attrition among patients who started antiretroviral therapy in Zimbabwe’s National Antiretroviral Therapy Programme between 2012 and 2015

PONE-D-19-23547R1

Dear Dr. Makurumidze,

We are pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it complies with all outstanding technical requirements.

Within one week, you will receive an e-mail containing information on the amendments required prior to publication. When all required modifications have been addressed, you will receive a formal acceptance letter and your manuscript will proceed to our production department and be scheduled for publication.

Shortly after the formal acceptance letter is sent, an invoice for payment will follow. To ensure an efficient production and billing process, please log into Editorial Manager at https://www.editorialmanager.com/pone/, click the "Update My Information" link at the top of the page, and update your user information. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, you must inform our press team as soon as possible and no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

With kind regards,

Kwasi Torpey, MD PhD MPH

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

23 Dec 2019

PONE-D-19-23547R1

Retention and predictors of attrition among patients who started antiretroviral therapy in Zimbabwe’s National Antiretroviral Therapy Programme between 2012 and 2015

Dear Dr. Makurumidze:

I am pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

For any other questions or concerns, please email plosone@plos.org.

Thank you for submitting your work to PLOS ONE.

With kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Professor Kwasi Torpey

Academic Editor

PLOS ONE