Health and economic impact of decentralized health interventions in sub-Saharan Africa

Abstract

Abstract Health and economic impact of decentralized health interventions in sub-Saharan AfricaChair of the supervisory committee: Monisha Sharma, Associate Professor Department of Global Health In most low- and middle-income country (LMIC) settings, including Eastern and Southern Africa (ESA), health systems face persistent challenges in delivering accessible, quality care due to limited healthcare infrastructure, workforce shortages, and logistical constraints. Funding for critical health responses, such as HIV and TB, remains insufficient; UNAIDS estimates a shortfall of $9.8 billion to meet HIV targets in 2025, and 62 LMICs reported a combined funding gap of $1.5 billion for TB programs in 2022. Although innovative solutions such as decentralized service delivery models show promise in expanding healthcare access with a patient- and provider-centered lens, governments remain uncertain about the affordability and economic sustainability of these approaches. Economic evaluations are essential to help policymakers understand the health and economic implications of adopting decentralized models, which could increase coverage more efficiently and may reduce funding requirements in LMIC settings where health budget allocation is critical. This dissertation applies economic evaluation to assess the health and economic impact of decentralized care interventions across tuberculosis (TB), HIV, and cardiovascular disease among people living with HIV (PLHIV) in real world settings.Aim 1 evaluates the cost and cost-effectiveness of decentralized TB diagnostic testing using the Molbio Truenat MTB Plus and MTB-RIF Dx assays compared to hub-and-spoke GeneXpert MTB/RIF Ultra testing in Mozambique and Tanzania, using primary cost and effectiveness data from a pragmatic cluster-randomized trial. Facility-based diagnostic costs per participant tested were higher in the decentralized arm ($51 vs $40 in Mozambique; $46 vs $31 in Tanzania) from a modified societal perspective. However, the diagnostic cost per participant initiating TB treatment within seven days was comparable or lower with decentralized testing ($703 vs $853 in Mozambique; $592 vs $596 in Tanzania). The incremental cost per incremental seven-day treatment initiation was $422 (95% uncertainty range: –$114, $1,019) in Mozambique and $580 ($167, $1,638) in Tanzania. Cost-effectiveness was highly sensitive to testing volume, with per-test costs stabilizing when instruments were utilized at three or more tests per day. These findings demonstrate that decentralized molecular TB testing can be cost-effective when deployed in settings with sufficient demand and appropriate integration into service delivery. Aim 2 assesses the health and economic impact of hypertension screening, treatment, and an associated implementation strategy (Systems Analysis and Improvement Approach, “SAIA”) among PLHIV in Mozambique using a decision-analytic state-transition model informed by trial data. Scaling up hypertension screening and pharmacological treatment reduced acute myocardial infarction and stroke events by 25–31% over ten years and yielded an incremental cost-effectiveness ratio (ICER) of $212 per disability-adjusted life year (DALY) averted, at an additional cost of $4.61 per person per year. Incremental to the clinical intervention, the SAIA implementation strategy further reduced cardiovascular events by approximately 15% and had an ICER of $44 per DALY averted, at an additional cost of $0.79 per person per year. Average ten-year cardiovascular risk was reduced by 29.3% with the intervention alone and by 40.3% when the SAIA strategy was co-introduced. These results show that implementation strategies can generate substantial additional health gains at low incremental cost, but only when a minimum level of health system readiness is present. Aim 3 evaluates the cost-effectiveness of delivering HIV pre- and post-exposure prophylaxis (PrEP and PEP) through online pharmacies in Kenya using an agent-based HIV transmission model parameterized with empiric utilization data from an online PrEP/PEP pilot study. Online PrEP/PEP achieved low population coverage (0.7% for PEP and 0.3% for PrEP) but was projected to avert 13.9% of HIV infections over ten years. From the Ministry of Health perspective, the intervention was cost-effective with an ICER of $212 per DALY averted. In a scenario with online PEP only, 11.6% of infections were averted with a similar ICER of $210 per DALY averted, indicating that PEP accounted for the majority of health benefits due to high observed demand. The intervention remained cost-effective across sensitivity analyses and became cost-saving under scenarios of reduced HIV testing and treatment coverage. Across chronic and infection diseases, this dissertation demonstrates that decentralized care models can deliver substantial health benefits and represent good value for money in sub-Saharan Africa, but that their efficiency depends critically on implementation context, utilization, and system readiness. By embedding economic evaluation within implementation science, this work provides policy-relevant evidence to guide decisions about scaling, sustaining, and adapting decentralized health interventions in resource-constrained settings.