HARNESSING AI FOR COMMUNITY HEALTH: DESIGN FRAMEWORK AND PILOT PROTOCOL FOR A HAUSA-LANGUAGE CHATBOT FOR MALARIA PREVENTION IN BAUCHI STATE, NIGERIA

ABSTRACT

This paper presents a design framework and study protocol; no empirical results are reported. Malaria is a critical health burden in Bauchi State, Northern Nigeria, driven partly by the absence of communication tools accessible in Hausa, functional under low-bandwidth conditions, and suitable for low-literacy users. We propose a dual-platform AI chatbot for malaria prevention via WhatsApp and SMS, using Google Dialogflow with a hybrid rule-based and retrieval-based NLP approach for low-resource Hausa. Content is developed by an interdisciplinary team of computer scientists, biomedical scientists, a Hausa language specialist, and community health workers. Three research questions guide the study: (RQ1) Is the chatbot feasible over eight weeks in rural Bauchi State? (RQ2) Does it improve malaria KAP scores? (RQ3) What shapes adoption among low-literacy users? A 200-participant pilot across four LGAs is planned, with ethics clearance from Federal Polytechnic Bauchi's Research Ethics Board.

Keywords: AI chatbot; Hausa NLP; malaria prevention; mHealth; Bauchi State; study protocol