Client and healthcare worker experiences with differentiated HIV treatment models in Eswatini

Abstract

Introduction Universal access to antiretroviral therapy (ART) is a cornerstone of Eswatini’s national HIV strategy, and the number of people on ART in the country more than tripled between 2010 and 2019. Building on these achievements, the Ministry of Health (MOH) is scaling up differentiated service delivery, including less-intensive differentiated ART (DART) models for people doing well on treatment. We conducted a mixed-methods study to explore client and health care worker (HCW) perceptions of DART in Eswatini. Methods The study included structured site assessments at 39 purposively selected health facilities (HF), key informant interviews with 20 HCW, a provider satisfaction survey with 172 HCW and a client satisfaction survey with 270 adults. Results All clients had been on ART for more than a year; 69% were on ART for ≥ 5 years. The most common DART models were Fast-Track (44%), Outreach (26%) and Community ART Groups (20%). HCW and clients appreciated DART, noting that the models often decrease provider workload and client wait time. Clients also reported that DART models helped them to adhere to ART, 96% said they were “very satisfied” with their current model, and 90% said they would recommend their model to others, highlighting convenience, efficiency and cost savings. The majority of HCW (52%) noted that implementation of DART reduced their workload, although some models, such as Outreach, were more labor-intensive. Each model had advantages and disadvantages; for example, clients concerned about stigma and inadvertent disclosure of HIV status were less interested in group models. Conclusions Clients in DART models were very satisfied with their care. HCW were also supportive of the new approach to HIV treatment delivery, noting its advantages to HF, HCW and to clients. Given the heterogeneous needs of people living with HIV, no single DART model will suit every client; a diverse portfolio of DART models is likely the best strategy.