The Lablite project: A cross-sectional mapping survey of decentralized HIV service provision in Malawi, Uganda and Zimbabwe

Most HIV-infected individuals on antiretroviral therapy (ART) in low and middle-income countries are treated following the World Health Organization (WHO) public health approach [1]. The public sector provides one standard first-line regimen, with alternative drug substitutions for anti-tuberculosis (TB) co-therapy or first-line regimen toxicity [1]. When first-line failure occurs, the patient switches to a standard boosted-protease inhibitor (PI)-based second-line regimen [1]. National guidelines specify simple ART formularies with few combination drugs in order to facilitate procurement and supply-chain management [2‐4]. Decentralization (delivery of services outside specialist centres) and task shifting (delegation of routine services to lower level relevant cadre) are key components of equitable public sector roll-out to ensure access to ART beyond ART facilities in tertiary centres and those operating in well-supported research programs [5‐13]. Access to HIV testing and ART has been prioritized over laboratory services for monitoring ART toxicity and identifying treatment failure and the need to switch to second-line [14]. This has enabled large numbers of individuals to access ART and to remain on therapy [15]. It will remain the bedrock for further service expansion towards universal ART access, particularly in sub-Saharan African countries with generalized HIV epidemics, constrained health budgets and fragile health systems.

Funding for HIV and AIDS programs is generally static and has even recently declined in some resource constrained countries [16]. The “first-line/second-line” paradigm and the ensuing simplified formularies are not in question and decentralization remains a core policy objective in most countries. Despite robust clinical trials evidence demonstrating that ART may be given in both adults and children without routine laboratory monitoring [17‐19], there remains an on-going debate about whether to prioritize limited funding towards laboratory capacity building for monitoring or expanding access to treatment [20, 21]. The 2013 WHO guidelines [22] promote both wider access to viral load monitoring and expanded entry into care (through raising the CD4 threshold for ART initiation and the ‘B plus’ approach to prevention of mother-to-child transmission whereby pregnant women start ART for life), without specific guidance on how to prioritize across these different aspirational approaches.

As these debates continue, many countries continue to decentralize their ART programs and policymakers face the challenge of prioritizing equity in access to ART with laboratory services and other health systems gaps. Although there are numerous studies describing decentralized ART programs in sub-Saharan African countries [5, 6, 9, 10, 23‐27], few have described the state of national ART roll-out at the lowest level of decentralized primary service delivery. Moreover these studies [28, 29] have assessed epidemiologic program outcomes (mortality and lost to follow-up) but have generally not described HIV service decentralization through a health systems lens. There also remains a need to gather evidence on the safety, efficacy and cost-effectiveness of decentralization of services as clinically directed monitoring is operationalized in the context of limited laboratory support, outside of national centres of excellence.

Malawi, Uganda and Zimbabwe account for about 11% of people living with HIV globally and national ART coverage rates from 2012 UNAIDS data were 70%, 64% and 80% respectively, increased from 67%, 54% and 77% in 2011 [15]. Public sector scale-up of ART in these countries began about a decade ago, but decentralization and ART coverage at the lowest level of health services remain variable [30]. The Lablite project (http://​www.​lablite.​org; 2011–2015) is a multi-country implementation project in these three countries to evaluate whether decentralized ART care can be delivered effectively at lower level health centres with limited laboratory services, and to assess the economic implications of decentralization employing a low technology, task-shifting strategy. Lablite began with a cross-sectional survey of representative health facilities to describe and compare national and inter-country delivery of training, clinical care and use of laboratories and monitoring in ART roll-out, providing a baseline for the project; this baseline survey is described here.

The Lablite project teams, working with MoHs in Malawi, Uganda and Zimbabwe, conducted a comparative cross-sectional survey of HIV service provision in a sample of health facilities. This survey, carried out at the start of the Lablite Project during 2012, gives an overview of the operational realities for patients and frontline health workers at the facility level during a period where countries were in various stages of transitioning [2, 3, 31] from previous guidelines for ART provision to the WHO 2010 update [35].

In this multi-country survey, HIV services were available comprehensively within referral facilities. Primary health facility coverage of HTC and basic PMTCT services was also comprehensive, although stock-outs of HIV test-kits, and drugs would clearly have limited the services at times. Nevertheless, the findings show that coverage of ART and paediatric HIV services (in particular) to the decentralized health facility level remained incomplete even though roll-out to lower level primary health care facilities of ART had already been underway. At the time of the study, integration of ART services into ANC or Maternity was also not documented at any site.

The 2013 WHO Guidelines [22] advocating for adoption of Option B+ [36, 37] in many low resource, high prevalence operational settings necessitates that ART services become available and integrated into primary health centres where there are ANC and MCH services provided [38]. Early programmatic data shows considerable variation in retention on ART in Option B + women between health facilities [39, 40] as well as association of attrition with different approaches for timing of ART initiation relative to knowledge of HIV test result, place of ART initiation (antenatal or ART clinic), place of follow-up postnatally and post-breastfeeding, and type of support women receive [41]. As numbers of people on therapy are increasing, many health systems bottlenecks remain, in particular around inadequate support of human resources for health capacity, supply chain management and laboratory services infrastructure.

From a human resources perspective, successful ART roll-out relies on task shifting of clinical services in primary care facilities [42‐45]. This study highlights further constraints beyond clinical tasks around laboratory technical capacity as well as administrative tasks. Malawi, which has one of the lowest professional health cadres-to-population ratios in the world, has uniquely adopted mitigation strategies which include the utilization of a formalized community health worker cadre, and use of NGO partners to implement a standardized supportive supervision and mentorship program led by the MoH national program.

Despite maturity of national ART roll-out programs with many programs operating for nearly a decade or longer, stock-outs of HIV test kits and drugs remains a critical bottleneck to access [46, 47], and needs urgent attention in the light of wider access being strived for in the WHO 2013 guidelines update. Mechanisms to forecast drugs through quantification of retention in care is particularly challenging for alternative first line and second line regimens [48, 49]. The data from this study highlighted that even at the primary care level, due to policy transition (e.g. d4T phase-out, introduction of TDF, and moving from NVP to efavirenz (EFV), various first-line regimens were available. The complexity of movement of large quantities of people in and out of regimen categories can be problematic during the period of transition, as the number of regimens used becomes more complicated with the evolution of the WHO guidelines. Accurate forecasting and ensuring availability of drugs is of key importance as patients may be triaged and delayed from substituting treatment regimens due to stock-out threats. Patients who may meet clinical criteria to be on certain regimens may be triaged out based on need and limited quantities. Downstream, ‘on the ground’ decisions made due to limited resources will misclassify groups of people, and contribute to lack of sensitivity in traditional methods of forecasting (which are based on monitoring numbers of people retained on specific regimens at a cross sectional point in time).

In our study, frequent stock outs of drugs, particularly cotrimoxazole, were reported in Malawi and Uganda, although less so in Zimbabwe. Stock outs emphasize the need to prioritize development of stock management capacities centrally and at the frontlines of delivery. Planners need to acknowledge and respond to the potential impacts of expanding ART eligibility criteria without adequate resourcing for drugs.

Maintaining supply of HIV test kits seems to be particularly problematic [50]. Already prioritization and quantification can be affected by challenges in coordinating multipronged HTC strategies supporting both community based screening of asymptomatic individuals, and facility based testing through PITC. Expanded eligibility, without concomitant supply side support for HTC kit logistics may contribute to ongoing stock out issues for facility based PITC programs targeting the patients who are most likely to need it the most e.g. patients with severe immunodeficiency, TB/HIV patients, HIV+ pregnant women eligible for Option B or Option B+ strategies.

As with other literature [51], our study reinforces that laboratory testing for drug toxicity and treatment effectiveness monitoring is available in these African countries, but generally not by patients accessing care at lowest level health facilities. Even haemoglobin (a very basic laboratory test) was available in less than half of primary health facilities. With respect to treatment monitoring, fewer lower level facilities had CD4 testing on-site, and in Malawi, few facilities used CD4 to monitor HIV patients (recommended for pre-ART [4]). On average the number of CD4 tests conducted per month was lower than reported or recommended in National guidelines, given numbers of adults on and initiating ART, even in sites which had access to CD4 testing. Overall and particularly in Malawi and Zimbabwe, very few facilities had any access to viral load monitoring, which is highlighted as the preferred monitoring method in WHO 2013 guidelines. Very few facilities have been able to adopt the recommendations for use of viral load monitoring for treatment failure, suggesting that in the immediate future, reinforcement of good clinical decision-making seems to be the most practical approach for monitoring patients on ART.

Finally, it is clear that paediatric HIV services are lagging behind management of adults on ART, and our study reflects the low coverage of paediatric ART for those in need in sub-Saharan Africa. Roll-out of DBS-PCR, which is important for early infant diagnosis and treatment has only been modestly successful [52‐55] despite the availability of this modality in many countries for the last 5 years. The availability of paediatric ART is also inconsistent, as we show that in Uganda, stock outs of paediatric medications are relatively common and occur more frequently than adult ART. The paucity of child-friendly formulations [56] in most national program supply chains should be acknowledged and calls for better harmonisation across adult and paediatric formulations. The current implementation of more aggressive PMTCT strategies should certainly lead to a decrease in paediatric HIV incidence, but until issues of infrastructure, health care worker capacity, and supply chain constraints are managed, gaps in paediatric HIV services will remain. Acceleration of ART roll-out to children is key and can be done without routine toxicity monitoring [18].

One limitation of this study is that the numbers were relatively small, but purposeful sampling allowed selection from different regions and levels of health care to allow us to capture within country and between country differences. An additional limitation is that sampling may not necessarily be generalizable as representation of each country was based on sites selected by the country MoH partners. In Uganda, health centre IIIs and health centre IVs were selected from all four regions of the country, thus results are most generalizable nationally. In Zimbabwe facilities were included in 1 district each of 4 provinces (all in the North-East), namely, Mashonaland East, West, Central and Manicaland. Two urban facilities in the capital city were included for comparison. In Malawi, 3 facilities in the Northern region (2 rural, 1 periurban referral hospital in Chitipa District), 3 facilities in the Central region (1 urban tertiary hospital, 2 periurban facilities in Lilongwe District) and all 14 health facilities in the Phalombe District Health Office in the Southern Region (1 referral hospital, and a mixture of mission and public primary health facilities) were selected thus some inter-regional variations may have been missed. Data collected applied to the day of interview or, for stock-outs of HIV test kits and drugs, to the 3 months prior to the interview. Whereas the most recent data may be most reliable, this approach cannot capture trends, seasonal variations or unusual fluctuations. However, overall this survey is likely to provide a reasonable snapshot of health facilities by country and by level at one point in time. Most data were collected by interview and it is possible that the interviewees provided answers with a social desirability bias, in the belief that funding might be forthcoming if they highlighted the shortages and needs in services. A final limitation is that some data were extracted from the existing MoH monitoring and evaluation tools for HIV, and there may be some concerns regarding accuracy and consistency of such operational data. In Malawi, where facility level is available electronically through the national M and E system, data was harmonized with the national data to confirm accuracy and generalizability.

In the context of the recent release of the 2013 WHO guidelines [22], we can anticipate that the variability in practice and challenges on the ground documented in this study, can be expected to continue to be implementation bottlenecks to operationalizing expanded eligibility criteria. Care should be taken to quantify economic and operational feasibility especially in high prevalence, resource-limited countries where the approach to treatment is standardized and not individualized by clinicians. This survey was undertaken before roll-out of the Option B-plus strategy and we plan to repeat it in 2014. As Malawi and Uganda have since started Option B-plus and Zimbabwe is planning to start shortly, we will have the opportunity to compare the effect on ART roll-out before and after adoption of Option B-plus, as well as across the three countries.

At the time of this survey, provision of HTC and PMTCT services was comprehensive across facility levels in Malawi, Uganda and Zimbabwe, although limited by stock-outs of supplies at some facilities. Decentralization of ART to primary care was ongoing but incomplete, being furthest ahead in Malawi; in Zimbabwe, ART provision in primary care was mostly by outreach teams, and in Uganda, provision was limited at primary care level. We found clear evidence of task shifting of clinic services in Malawi, but laboratory services were very limited in primary care, reinforcing the importance of good clinical management of patients. Stock-outs of drugs, particularly cotrimoxazole, and HIV-test kits, were reported in a number of facilities particularly in Uganda and Malawi. As demand for ART increases with adoption of Option B + (lifelong ART for pregnant and breastfeeding HIV-positive women) and higher threshold for ART initiation in the WHO 2013 guidelines, challenges of service provision on the ground may well increase if the health service barriers identified in this survey are not addressed concurrently.