Feasibility, acceptance, and workflow integration of an AI-enabled clinical decision support system for non-communicable diseases in Kiambu County, Kenya: A mixed-methods implementation evaluation

This repository contains the datasets generated during the development and evaluation of the NCDAI clinical decision support system for noncommunicable disease (NCD) management across primary health facilities in Kenya.

The files comprise the following datasets:

• De-identified demographic, medical history, and clinical parameters collected from participating health facilities with foundational patient characteristics used to evaluate the NCDAI platform’s clinical recommendations and performance. Expert physician ratings and qualitative comments on AI-generated recommendations used to assess clinical appropriateness, guideline adherence, and inter-expert agreement.
• Anonymised responses from healthcare workers evaluating usability, workflow integration, and perceptions of the NCDAI system.
• Supplementary materials
• Structured AI prompt framework used in the NCDAI study to support AI-assisted clinical decision support for hypertension and diabetes, including system-level context instructions and dynamic patient-data prompt construction aligned with evidence-based guidelines. This R script demonstrates the logic used to assemble AI prompts in the NCDAI study. All patient information shown is synthetic and provided for methodological illustration only.