Economic evaluations of differentiated service delivery should include savings and ancillary benefits, not only health system costs

Abstract

We write in reference to the study by Nichols et al.[1] that evaluated costs and outcomes of community-based differentiated service delivery (DSD) models for HIV treatment in Zambia. The authors compared conventional, facility-based care to mobile antiretroviral therapy (ART), community adherence groups (CAGs), urban adherence groups (UAGs), and home delivery of ART under the Community HIV Epidemic Control (CHEC) model. The authors found that conventional care was least expensive in terms of direct clinical service and medication costs, whereas mobile ART, CAGs, UAGs, and CHEC were more expensive, in that order. We appreciate this detailed costing analysis of DSD models in Zambia. At the University of Maryland Baltimore, we have nearly two decades of experience in the provision of medical/technical service delivery in Zambia [2]. Based on community-based approaches demonstrated to improve HIV case-finding and linkage [3], provide high-quality care [4], and improve retention in adult ART programs [5], we developed and implemented the CHEC model [6], which provides home delivery of ART and was one of the models evaluated in this analysis. CHEC was primarily implemented in the PEPFAR/CDC-funded Stop Mother and Child HIV Transmission (SMACHT) project, which was conducted in the Southern Province of Zambia from 2015 to 2020. Under SMACHT, CHEC significantly improved maternal/child HIV outcomes [7] and achieved 90% linkage to ART and 91% viral suppression [8]. We would like to discuss four key considerations that are relevant to the economic impacts: the models compared, the outcome selected, how retention is defined, and the ancillary benefits and savings that were not included. First, the authors compared four DSD models to conventional care. It should be noted that all DSD models build on conventional care and essentially include conventional care at health facilities as part of their model. Without accounting for the savings created by DSD models, such as health facility decongestion or decreased patient costs, then by definition, all DSD models will appear to be more expensive than conventional care. Second, the manuscript evaluates the costs of the DSD models under the assumption that they provide similar services. However, CAGs, UAGs, and mobile ART only provide ART services, whereas the CHEC model provides HIV testing services in the community, escorted linkage to care, and home ART delivery. To fairly compare CHEC against other models, these extra services should also be considered. Under SMACHT, the CHEC model provided HIV testing services to nearly 1.4 million Zambians, with the identification of 46 138 PLHIV and linkage of 41 366 to ART (>90% linkage) [8]. While retention is a valuable outcome, viral load suppression is a more direct measure of individual health outcome and population HIV epidemic control, which was not assessed. Viral suppression among CHEC clients receiving home ART was 97%, and overall suppression in the SMACHT project was 91% [8]. The authors defined retention as engagement with the national health system occurring 9–15 months after being defined as 'stable-on-care' [1]. Under this analysis, they estimated CHEC 12-month retention as 79% [1]; however, over the same time period we recorded 92% retention at 12 months among CHEC clients [8]. This discrepancy is likely due to the fact that the design of CHEC intentionally decreases facility visits, and not all CHEC interactions were documented in the national health record. This bias contributes to a misclassification of outcome results and an underestimation of retention benefits of home ART. Finally, this evaluation only describes costs borne to the healthcare system by DSD models, but none of the savings or benefits to the patients for which DSD models were actually designed; this is a glaring omission. Indeed, other studies that included such benefits found that DSD models reduce costs compared to standard of care [9]. The CHEC model provided a client-centered approach that has multiple medical, economic, and social benefits not just to patients but to the systems needed to achieve and sustain epidemic control. These are summarized in Table 1. Table 1 - Benefits of the Community HIV Epidemic Control model. Benefits Individual Household Community Health facility Medical Check vital signs and screen for symptomsIndividualized care and counselingImproved adherence and retentionIdentification of other illnesses such as TB, HTN, malaria that need referral and treatment CHWs often identify illnesses or health conditions within families that need referralAlso help to refer pregnant women for early ANC CHWs lead health talks in the communityCHWs refer pregnant women for ANCAbility to rapid disseminate information and practices for novel health concerns, i.e. COVID Decongests health facility by task-shifting low acuity patients to community careAllows health facility staff to focus on more acute patientsDecreases wait times for patients Economic Saves patient transport costsEnables patient to continue working, rather than spend several days traveling back and forth to clinic Limiting costs and time associated with transport significantly reduce burden on family Whilst en route to patients, CHWs are often engaged by community members on health and prevention Minimizes need for additional human resources Social Creates dignity and respect for patient and their conditionFacilitates client centered services Mitigates stigma around HIV by normalizing patient and demonstrating how to care for them CHWs serve as an extension of the health facilityViewed as a trusted source for health information ANC, antenatal care; HTN, hypertension; CHWs, community health workers; COVID, coronavirus disease; TB, tuberculosis. In conclusion, the CHEC model approach provided significant benefits during its implementation, and we propose that any evaluation of this model (and other DSD models) should include a global assessment of benefits and savings, not just provider-initiated costs. Ultimately, cost-effective DSD models are critically needed to help Zambia achieve long-lasting and sustainable HIV epidemic control; as such we need comprehensive evaluations not only of DSD costs, but also their benefits. Acknowledgements Conflicts of interest There are no conflicts of interest.

Topics & Concepts