Retention and sustained viral suppression in HIV patients transferred to community refill centres in Kinshasa, DRC

Abstract

Setting:

In 2010, Médecins Sans Frontières set up decentralised community antiretroviral therapy (ART) refill centres (“poste de distribution communautaire”, PODI) for the follow-up of stable human immunodeficiency virus (HIV) patients.

Objective:

To assess retention in care and sustained viral suppression after transfer to three main PODI in Kinshasa, Democratic Republic of Congo (DRC) (PODI Barumbu/Central, PODI Binza Ozone/West and PODI Masina I/East).

Design:

Retrospective cohort study using routine programme data for adult HIV patients transferred from Kabinda Hospital to PODIs between January 2015 and June 2017.

Results:

A total of 337 patients were transferred to PODIs: 306 (91%) were on ART for at least 12 months; 118 (39%) had a routine “12-month” viral load (VL) done, 93% (n = 110) of whom had a suppressed VL <1000 copies/ml. Median time from enrolment into PODI to 12-month routine VL was 14.6 months (IQR 12.2–20.8). Kaplan-Meier estimates of retention in care at 6, 12 and 18 months after enrolment into PODIs were respectively 96%, 92% and 88%.

Conclusion:

Retention in care and viral suppression among patients in PODI with VL results were better than patients in clinic care and national outcomes.

Worldwide, 37.9 million people were living with human immunodeficiency virus/acquired immune-deficiency syndrome (PLWHA) by end of 2018, with sub-Saharan Africa carrying most of the burden of this disease and accounting for nearly 70% of all PLWHA.¹ The adoption of the “triple 90” targets and the implementation of ‘universal test and treat’ strategy has resulted in a significant increase in the proportion of PLWHA on antiretroviral therapy (ART), with 62% of adults and 54% of children receiving lifelong ART in 2018.² Good adherence and retention in care are cornerstones for a successful HIV treatment outcome, reducing morbidity, mortality, transmission and HIV drug resistance at individual and population level. To improve and sustain ART retention in care, adherence and viral suppression, new interventions and models of health care delivery need to be explored.³

In 2010, Médecins Sans Frontières (MSF), an international non-governmental organisation (NGO) providing free medical care, in collaboration with the association of PLWHA “Réseaux national des personnes vivant avec le VIH (RENOAC)”, started a community-based ART supply programme at Kabinda Hospital (Centre hospitalière Kabinda, CHK) in Kinshasa. This centre, built in 2002 by MSF, is specialised in HIV/AIDS treatment and care for in- and out-patients. Its attendance has increased steadily with the introduction of ART. In order to relieve congestion, MSF set up a decentralised ART distribution strategy, through community-based refill centres (“poste de distribution communautaire”, PODI). In doing so, MSF hoped to reduce attrition by increase treatment availability and convenience. PODI have been included in the Ministry of Health (MOH) 2011–2015 strategic plan as an approach to reduce workload in health facilities with high numbers of patients on ART and increasing retention in care and adherence by providing care closer to patient’s home in a stigma-free environment. PODI employees are PLWHA who work at PODI centres to provide ART drugs to patients, perform voluntary testing for partners and family members and provide treatment adherence counselling. In the community, they also provide voluntary testing. There is no medical care in PODI.

At the beginning of the programme, the criteria for transferring stable patients from CHK to PODI were 1) being on ART for at least 6 months, 2) without active opportunistic infections, and 3) with a CD4 cell count >350 cells/μl. In 2016, the CD4 cell count criterion was replaced by a suppressed HIV viral load (VL) (⩽1000 copies/ml) at 6 months. Patient who achieved those criteria were considered stable on treatment, transferred to PODI for the three monthly ART supply and assigned to the closest health facility for the yearly clinical and virological monitoring. In January 2017, this model was modified, with patients transferred to PODI after a minimum of 3 months on successful ART, defined as a VL ⩽1000 copies/ml 3 months after ART initiation (“early” transfer).

The PODI model and outcomes have been described in an earlier publication in 2017,⁴ revealing low rates of loss to follow up and death after 6 months (2.2% and 0.1%, respectively), 12 months (4.8% and 0.2%) and 24 months (9% and 0.3%). This suggested good retention.

The present study aims at assessing sustained viral suppression and retention in care at 6, 12, 18 months after transfer to the three main PODIs in Kinshasa between January 2015 and June 2017.

MATERIALS AND METHODS

Setting

Kinshasa, the capital of the Democratic Republic of Congo (DRC), has an estimated 13 million habitants.⁵ The HIV prevalence in DRC was reported to be 1.2% (in Kinshasa 1.6%) in 2016.⁶ Stigmatisation of PLWHA is high and access to care is poor, with low coverage of testing and treatment. In 2017, out of 516 617 PLWHA in DRC, only 46% knew their status, 41% of whom were on ART and of these, only 10% had access to VL testing.⁷

Kabinda Hospital Centre

The Kabinda Hospital Centre (CHK) is an HIV care referral hospital started in 2002 in the South-Western part of Kinshasa with high patient attendance and turnover, where free HIV care is provided by MSF. By the end of 2016, 1500 patients were on ART and an additional 547 patients were initiated in 2017.⁷ Patients stable on ART and with suppressed VL (VL ⩽1000 copies/ml) were offered to be transferred out to a PODI to access community-based ART supply. The number of HIV-related consultations decreased in CHK since the implementation of the PODI.

Poste de distribution communautaire des antirétroviraux

These are centres staffed by lay community health workers who are HIV-positive, act as counsellors to HIV patients and dispense ART. Before 2016, PLWHA were transferred from all HIV health centres in Kinshasa to the PODI. In 2016, the Ministry of Health recommended to apply a zonal approach (one PODI for all HIV health centres from the same health district). The PODI referred back to these health centres for routine clinical and virological monitoring every 12 months and in case of patients’ illness. The MOH provides ART in PODI and ART stock outs are rapidly resolved with the support of MSF and RENOAC.

Design and study population

This was a retrospective cohort study using routinely collected data from all HIV patients >15 years old and stable on ART (as defined in Table 1), transferred from CHK to the three principal PODIs in Kinshasa (PODI Barumbu/Central, PODI Binza Ozone/West and PODI Masina I/East) between January 2015 and June 2017 (Table 1).

TABLE 1.

Criteria of Kabinda Hospital, to define a “stable” patient on ART, who can be transferred to a “poste de distribution communautaire”, January 2017, Kinshasa, DRC

Minimum 3 months on ART No opportunistic infections VL ⩽ 1000 copies/ml Not pregnant or breastfeeding (or with HIV-exposed baby) >15 years old Not on second-line drugs for <3 months (transfer possible after undetectable VL check) Not having been hospitalised for OIs in the last 3 months No change in ART in the last 3 months No ART side effects BMI >16 kg/m2

ART = antiretroviral treatment; DRC = Democratic Republic of Congo; VL = viral load; OI = opportunistic infection; BMI = BODY MASS INDEX.

Data collection

HIV data are routinely collected and entered in electronic databases: FUCHIA (Follow-up and care of HIV infection and AIDS) at CHK and Tier.Net (Three Interlinked Electronic Registers. Network) at the PODI. Routine data include sex, age, education level, marital status, ART regime and duration on ART, HIV history and at each visit (presence/absence of opportunistic infections (OIs), blood test results (e.g., CD4, VL), ART regime changes and treatment outcome).

As patients were assigned to an HIV health centre for yearly clinical check-up and VL testing, we collected VL results in those centres from the Tier.Net database. In centres lacking updates from the Tier.Net database, data were collected directly from individual patient records. Quality control was done on the Tier. Net database in the PODIs. All data for this study were encoded in Epidata (Epidata Association, Odense, Denmark).

Data management and analysis

Data were analysed with Stata software (15.0 Copyright 1985–2017). Baseline sociodemographic characteristics were described using frequencies and medians. We compared proportions with χ² or a Fisher-adjusted χ² test for frequencies below five. We estimated routine VL coverage and the proportions of sustained viral suppression at 12 months after transfer to PODI. ‘Routine VL’ was any VL done in the window of 9 and 18 months after transfer to PODI. Viral suppression was defined as VL <1000 copies/ml.

The outcomes were as follows: 1) death, 2) loss to follow up (LTFU), 3) transfer out, and 4) active on ART in PODI. Patients were classified as LTFU if they did not return to the PODI for ART pick-up within 90 days after the date on which they were expected. Death from any cause was recorded.

The probability of survival on ART was estimated using the Kaplan-Meier method and log-rank test. Patients were considered to come under observation on the date they transferred to PODI. Death and LTFU were failure events, while transfer-out was a censoring event. Observation of time at risk for ART attrition ended either at date of death, LTFU or transfer-out. Observations were censored at 18 months after transferring to PODI if patients were still alive and on ART. Level of significance was set at 5%.

Ethical considerations

This protocol was approved by the Ethical Committee of the HIV/AIDS programme in DRC (Kinshasa). All analyses were done with anonymous data extracted from Fuchia and Tier.Net.

RESULTS

Sociodemographics characteristics

A total of 337 patients were transferred to PODI for ART services. Of these, 242 (72%) were female and the median age at transfer was 47 years (interquartile range [IQR] 41–54). Most of the patients, 235 (70%), had secondary education. Sixty-four per cent were transferred to PODI after 24 months of ART and 86% were on first-line ART (Table 2).

TABLE 2.

Characteristics of PLWHA transferred from Kabinda Hospital to PODI East, West and Central, January 2015–June 2017, Kinshasa, DRC

Characteristics	n	%
Sex
Female	242	72.0
Male	95	28.0
Age, years, median [IQR]	47 [41–54]
Age groups, years
15–18	2	0.6
19–24	14	4.1
25–44	115	34.1
45–64	197	58.5
⩾65	9	2.7
Educational level (9 missing)
None	9	2.6
Primary school	42	12.5
Secondary school	235	69.7
University	9	2.7
ART regimen at the time of transfer
First-line	289	85.8
Second-line	48	14.2
Community-based centres (PODI)
East	160	47.5
West	130	38.6
Central	47	13.9
Duration of ART before transfer, months, median [IQR]	34 [14–54]
3–5	8	2.4
6–11	54	16.1
12–23	56	16.8
24–35	50	15. 0
⩾36	166	49.7

PLWHA = people living with human immunodeficiency virus/acquired immune-deficiency syndrome; PODI = poste de distribution communautaire; DRC = Democratic Republic of Congo; IQR = interquartile range; ART = antiretroviral therapy.

Coverage of routine VL and maintenance of VL suppression

Of 337 patients transferred, 329 (97.0%) had a VL at transfer. Among these, 306 (93.0%) patients were alive and in care 1 year after transfer and thus required a routine 12-month VL (Figure 1). Overall, 189 (61.8%) patients had at least one VL load test during the follow-up. The median time after transfer to PODI of the first routine VL done was 14.6 months (IQR 12.2–20.8). A total of 118 (38.5%) patients had a VL done at 12 months after being transferred to PODI and the proportion of those who maintained their VL suppressed was 93% (n = 110). For 71 (23.2%) patients, routine VL was performed before 9 months or after 18 months in PODI (Table 3).

FIGURE 1.

Number and proportion of PLWHA with routine VL done and VL suppressed after transfer from Kabinda Centre to PODI East, West and Central, January 2015–June 2017, Kinshasa, DRC. PLWHA = people living with human immunodeficiency virus/acquired immune-deficiency syndrome; VL = viral load; PODI = poste de distribution communautaire; DRC = Democratic Republic of Congo.

TABLE 3.

Time of first routine VL testing after transfer of PLWHA from Kabinda Hospital to PODI East, West and Central, January 2015–June 2017, Kinshasa, DRC

Time of first VL after transfer months	n	%
<9	3	1.6
9–18	118	62.4
19–24	42	22.2
25–36	21	11.1
>36	5	2.7
Total	189	100

VL = viral load; PLWHA = people living with human immunodeficiency virus/acquired immune-deficiency syndrome; PODI = poste de distribution communautaire; DRC = Democratic Republic of Congo.

Retention in PODI

The overall probability of being in care after transfer to PODI at 3, 6, 12 and 18 months was 98.5% (95% confidence interval [CI] 96.5–99.3), 96.1% (95%CI 93.5–97.7), 91.7% (95%CI 88.2–94.2) and 88% (95%CI 84.3–91.4) respectively (Figure 2; Table 4).

FIGURE 2.

Kaplan-Meier plot of retention in care in PODI at time of transfer of PLWHA from Kabinda Centre to PODI East, West and Central, January 2015–June 2017, Kinshasa, DRC. PODI = poste de distribution communautaire; PLWHA = people living with human immunodeficiency virus/acquired immune-deficiency syndrome; DRC = Democratic Republic of Congo.

TABLE 4.

Retention in PODI by time of follow-up of PLWHA transferred from Kabinda Centre to PODI East, West and Central, January 2015–June 2017, Kinshasa, DRC

Duration on follow-up months	n	Death and LTFU	Retention (95%CI)
3	338	5	95.8 (96.5–99.4)
6	328	8	96.2 (93.5–97.7)
12	298	15	97.1 (88.2–94.8)
18	128	10	88.3 (84.3–91.4)

PODI = poste de distribution communautaire; PLWHA = people living with human immunodeficiency virus/acquired immune-deficiency syndrome; DRC = Democratic Republic of Congo; LTFU = loss to follow-up; CI = confidence interval.

DISCUSSION

This is the first study to investigate VL coverage and viral suppression among ART patients who were transferred to PODI (community-based refill centres) in the DRC. The VL testing coverage at 12 months was low (38.5%), while viral suppression for those covered was optimal at 93%. Overall 12-month retention was high at 91.7% (95%CI 88.2–94.2).

In comparison, in MSF-supported clinics in the same area in 2017, retention in care at 12 months was 80% and viral suppression was 89%.⁷ VL completion was better, at 80%, but this is explained by direct MSF support to VL scale up in these clinics. Nationally, VL completion at 12 months was 25% and viral suppression 74% in 2017.⁸ Retention in care at 12 months in sites with the electronic monitoring system Tier.net was 84% for 2016.

The WHO and national guidelines in DRC recommend annual VL monitoring.^4,9 For patients in PODIs this is done in the health facility where the PODI is linked. VL monitoring among patients in this study was low at 62%. Potential reasons for this include insufficient or delayed referral from PODI to health facility, as well as the fact that the national scale up of VL testing was limited to the pilot sites at the time of the study. Only three laboratories were performing VL in Kinshasa, with financial support from MSF and US President’s Emergency Fund for AIDS Relief. Supply as well as demand for VL was therefore limited.

The high 12 months sustained viral suppression and retention in care suggest that PODI are an effective model of care for stable HIV patients on ART. Patient-friendly care provided by peer PLWHA, a lower potential for stigmatisation by health care workers or patients without HIV, inter-patient relationships and peer support might promote good adherence to ART. Moreover, the disadvantages of overcrowded health centres, namely long waiting times and poor reception, are reduced.

Assessment of viral suppression in another type of community models of care showed high proportions similar to those in our study. In a retrospective cohort analysis of patients enrolled in Adherence Clubs (AC) in Cape Town, South Africa, between 2011 and 2014, VL completion was at 80% and viral suppression was 97% (95%CI 96–98) at 4 months, 96% (95%CI 95–97) at 16 months and 94% (95%CI 92–96) at 28 months after AC enrolment.¹⁰

Our results provide additional evidence on the effectiveness of differentiated models of care for PLWHA. Retention in care at 12 months was high (91%) in PODIs. In an earlier analysis, in 2014, in the same PODI, Vogt et al. found low mortality and loss to follow up rates at 12 months, respectively at 0.2% and 4.8%. In our study, the attrition rate (death plus LTFU) at 12 months was 4.4%, demonstrating that the early positive outcomes were retained over time. Other studies of community models also show good retention in care. In Swaziland, a retrospective analysis of stable adult patients enrolled in three different community refill centres showed retention in care of 90.9% and 82.2% at 6 and 12 months, respectively.¹¹ In the Cape Town study mentioned above, retention in care was 95.1% (95%CI 94.2–95.9) at 12 months in the ACs.¹⁰

Limitations

This study is limited by its retrospective, non-comparative design, the use of routine data, a small sample size and missing data on the virological outcomes of 38% of the participants. While these exploratory results illustrating a real-world setting with very limited resources are encouraging, they are prone to selection bias, as patients in PODIs are likely to be different from patients in clinic care, and measurement bias.

The three PODI included in the study were chosen purposively to evaluate the outcomes of referred CHK patients, as CHK had early access to VL testing. Other PODI in Kinshasa receive patients from other health centres or hospitals. This study therefore does not reflect the reality of all the PODIs in Kinshasa.

Another limitation is that patients who have not done a VL can be different from those who did, e.g., they could have been more likely ill. This might have introduced a bias in our results.

Implications

High retention in care and viral suppression in patients in PODIs are factors in support of further exploration and potential scaling up of this model and better organization of patient follow-up, especially in achieving suppressed VL. It seems from these results that it is not necessary to medicalise the PODIs. This would carry the risk of replicating the negative conditions of health centres (long waiting times, stigmatisation by health workers, reduced interaction between patients and with peers) that stimulated the creation of PODI. The model in its current organisation seems effective. The MOH should look for means to increase VL coverage and mentor the NGOs involved in this activity.

This study adds data to the very limited evidence on the efficiency of PODIs. Additional research and evaluations are necessary to corroborate and add to our results.

CONCLUSION

This study provides the first data on sustained viral suppression in community refill centres (PODIs), a particular differentiated model of care that originated in Kinshasa and is well adapted to highly stigmatised, low-prevalence settings. Good retention in care and promising data on viral suppression—albeit limited by low coverage given the early stage of VL scale up—add to the evidence supporting PODIs as a successful HIV-differentiated model of care.