"Patients are not the same, so we cannot treat them the same" - A qualitative content analysis of provider, patient and implementer perspectives on differentiated service delivery models for HIV treatment in South Africa.

Introduction In 2014, the South African government adopted a differentiated service delivery (DSD) model in its "National Adherence Guidelines for Chronic Diseases (HIV, TB and NCDs)" (AGL) to strengthen the HIV care cascade. We describe the barriers and facilitators of the AGL implementation as experienced by various stakeholders in eight intervention and control sites across four districts. Methods Embedded within a cluster-randomized evaluation of the AGL, we conducted 48 in-depth interviews (IDIs) with healthcare providers, 16 IDIs with Department of Health and implementing partners and 24 focus group discussions (FGDs) with three HIV patient groups: new, stable and those not stable on treatment or not adhering to care. IDIs were conducted from August 2016 to August 2017; FGDs were conducted in January to February 2017. Content analysis was guided by the Consolidated Framework for Implementation Research. Findings were triangulated among respondent types to elicit barriers and facilitators to implementation. Results New HIV patients found counselling helpful but intervention respondents reported sub-optimal counselling and privacy concerns as barriers to initiation. Providers felt insufficiently trained for this intervention and were confused by the simultaneous rollout of the Universal Test and Treat strategy. For stable patients, repeat prescription collection strategies (RPCS) were generally well received. Patients and providers concurred that RPCS reduced congestion and waiting times at clinics. There was confusion though, among providers and implementers, around implementation of RPCS interventions. For patients not stable on treatment, enhanced counselling and tracing patients lost-to-follow-up were perceived as beneficial to adherence behaviours but faced logistical challenges. All providers faced difficulties accessing data and identifying patients in need of tracing. Congestion at clinics and staff attitude were perceived as barriers preventing patients returning to care. Conclusions Implementation of DSD models at scale is complex but this evaluation identified several positive aspects of AGL implementation. The positive perception of RPCS interventions and challenges managing patients not stable on treatment aligned with results from the larger evaluation. While some implementation challenges may resolve with experience, ensuring providers and implementers have the necessary training, tools and resources to operationalize AGL effectively is critical to the overall success of South Africa's HIV control strategy.