Artificial Intelligence-Assisted Clinical Decision Support for Non-Communicable Disease Management in Family Medicine Clinics in Nigeria: Development, Testing, and Ethical Analysis

Background: Artificial intelligence-assisted clinical decision support systems have demonstrated efficacy in improving diagnosis and treatment guidance for non-communicable diseases in high-income settings. Their development, contextual adaptation, and ethical implications in Nigerian family medicine represent a frontier research domain. Methodology: A multi-phase study developed and tested an AI-assisted clinical decision support system for hypertension and diabetes management at family medicine clinics across Abuja, Lagos, and Kano. Phase one conducted data needs assessment and model training using de-identified records from 2,000 patients. Phase two piloted the system with 20 family medicine practitioners at six clinics and evaluated decision concordance, time efficiency, and patient safety. Phase three applied a locally adapted AI ethics framework to assess algorithmic bias and accountability implications. Results: The AI-assisted system demonstrated high concordance with evidence-based treatment guidelines and reduced average consultation decision time significantly. Minor algorithmic bias was detected in medication dosing recommendations for elderly patients, requiring model recalibration. Practitioner acceptance was high among younger clinicians. Conclusion: AI-assisted clinical decision support is feasible, acceptable, and effective in Nigerian family medicine settings, with ethical safeguards essential to ensure equitable and safe deployment. Original ethical governance frameworks for AI in Nigerian primary care are urgently needed and are proposed in this study. Keywords: artificial intelligence, clinical decision support, non-communicable diseases, family medicine, Nigeria