Implementation fidelity to HIV assisted partner services (aPS) during scale-up in western Kenya: a convergent mixed methods study

BACKGROUND: HIV assisted partner services (aPS) is an intervention to improve HIV status awareness among sex and drug-injecting partners of people newly diagnosed with HIV (index clients). Implementation fidelity—the degree to which an intervention is conducted as intended – is critical to effectiveness, but there are limited data about aPS fidelity when delivered by HIV testing service (HTS) providers. We explored factors affecting implementation fidelity to aPS in two high-HIV prevalence counties in western Kenya. METHODS: We used convergent mixed methods adapting the conceptual framework for implementation fidelity within the aPS scale-up project. This was an implementation study examining scale-up of APS within HTS programs in Kisumu and Homa Bay counties that recruited male sex partners (MSPs) of female index clients. We defined implementation fidelity as the extent to which HTS providers followed the protocol for phone and in-person participant tracing at six expected tracing attempts. Quantitative data were collected from tracing reports in 31 facilities between November 2018 and December 2020, and in-depth interviews (IDIs) were conducted with HTS providers. Descriptive statistics were used to describe tracing attempts. IDIs were analyzed using thematic content analysis. RESULTS: Overall, 3017 MSPs were mentioned of whom 98% (2969/3017) were traced, with most tracing attempts being successful (2831/2969, 95%). Fourteen HTS providers participated in the IDIs—mostly females (10/14, 71%) with a median age of 35 years (range 25–52), who all had post-secondary education (14/14, 100%). The proportion of tracing attempts occurring by phone ranged from 47 to 66%, with the highest proportion occurring on the first attempt and lowest on the sixth attempt. Contextual factors either enhanced or impeded implementation fidelity to aPS. Positive provider attitudes towards aPS and conducive work environment factors promoted implementation fidelity, while negative MSP responses and challenging tracing conditions impeded it. CONCLUSION: Interactions at the individual (provider), interpersonal (client—provider), and health systems (facility) levels affected implementation fidelity to aPS. As policymakers prioritize strategies to reduce new HIV infections, our findings highlight the importance of conducting fidelity assessments to better anticipate and mitigate the impact of contextual factors during the scale-up of interventions. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12913-023-09541-1.