Background: Implementation science offers valuable frameworks for evaluating digital health interventions in real-world settings, yet application of these frameworks in low-resource contexts remains limited. This paper synthesizes lessons from three digital health evaluations conducted in Kenya between 2022-2025 to guide future implementation research and inform health system decision-making.
Methods: We conducted a comparative analysis of three digital health evaluations: (1) a community health worker (CHW) mHealth tool in Machakos County (n=85 CHWs, 12 community health units); (2) a clinician decision support system in five county hospitals (n=120 clinicians, 5 hospitals across 5 counties); and (3) an electronic medical record (EMR) implementation in 12 primary health facilities in Embu County. For each evaluation, we systematically documented: implementation science frameworks used, adaptations made for the local context, data sources and measurement approaches, stakeholder engagement strategies, and lessons learned.
Results: Key lessons emerged across five domains: Framework Selection and Adaptation (RE-AIM most commonly used, with necessary adaptations for local context); Measurement Adaptations for Low-Resource Settings (fidelity measured through combined automated logs and sampling, sustained adoption metrics developed); Qualitative Methods Integration (rapid approaches including "quick interviews" and SMS-based experience sampling proved valuable); Stakeholder Engagement for Sustainability (quarterly feedback sessions and co-interpretation workshops critical); and Sustainability Factors (integration with existing systems, county budget allocation, designated technical support).
Conclusions: Implementation science frameworks can be effectively applied in low-resource settings with appropriate adaptations. Key success factors include simplified measurement approaches balancing rigor with feasibility, integrated rapid qualitative methods capturing real-time learning, and sustained stakeholder engagement positioning evidence for policy uptake.