Data and code for: Quality, readability and guideline agreement of AI-generated cardiovascular prevention messages in a Kenyan community

This dataset contains the data and analysis code for a cross-sectional content analysis of 200 cardiovascular prevention messages. The messages were produced by a retrieval-augmented generation tool for adults aged 30 to 60 years in Embu County, Kenya, between February and March 2026. It includes the de-identified messages with each participant's profile, the independent scores of two clinical officers (modified DISCERN, Global Quality Score and eight guideline-based items), the final scores, the readability results and the Python code used for the analysis. See README.md inside the zip file for the contents and for instructions to reproduce the results.