AI-based Clinical Decision Support for Primary Care

This technical report evaluates the impact of an AI-based clinical decision support tool, AI Consult, in primary care settings. Conducted in partnership with Penda Health, the study involved a quality improvement analysis of 39,849 patient visits across 15 clinics in Nairobi, Kenya. The objective was to assess the effectiveness of AI Consult in reducing clinical documentation and decision-making errors. Results indicated that clinicians using AI Consult made 16% fewer diagnostic errors and 13% fewer treatment errors compared to those without access to the tool. The study also highlighted that the implementation of AI Consult could potentially avert significant numbers of diagnostic and treatment errors annually. The report discusses the design and integration of AI Consult into clinician workflows, emphasizing the importance of preserving clinician autonomy while providing timely support. Findings suggest that LLM-based tools can enhance clinical decision-making and improve patient care quality in real-world healthcare settings.