PostMom: An AI-Powered Culturally Tailored Persuasive App for Postnatal Support for Underserved Nigerian Mothers

Abstract

The World Health Organization (WHO) recommends that pregnancy, childbirth, and the postnatal period should be positive experiences for all women, yet many underserved women in countries including Nigeria still lack adequate postnatal support. Limited access to postpartum care, cultural barriers, and inadequate health communication continue to hinder positive postnatal outcomes for Nigerian mothers. Although artificial intelligence (AI)-powered mobile health (mHealth) applications exist for maternal health, culturally appropriate interventions that integrate persuasive strategies for postnatal education and support remain limited, especially in low-resource contexts. This thesis presents PostMom, a persuasive mHealth technology that integrates AI techniques, including natural language processing (NLP), to deliver culturally appropriate postnatal health education and support for underserved Nigerian mothers. This work followed a three-phase, user-centered process. First, a review of 62 studies on AI in maternal health identified gaps in existing solutions and informed a set of persuasive strategies using the Persuasive System Design (PSD) model. Second, guided by these insights and consultations with four Nigerian medical doctors, a medium-fidelity prototype of PostMom was designed and evaluated through surveys and semi-structured interviews with 36 Nigerian mothers, whose feedback shaped design refinements. Third, a refined version of PostMom was developed and deployed in a one-week field study during which 70 mothers used the application for daily postnatal guidance and support. Results from validated usability and user experience assessments demonstrated good usability and positive user experience, while pre-post knowledge assessments showed that PostMom's persuasive features significantly improved participants' postnatal health knowledge (p < .001) and enhanced health information literacy. This thesis offers design recommendations for postnatal health applications, underscoring the importance of a human-in-the-loop design process. Overall, this work contributes to understanding the design and development of culturally tailored, AI-powered persuasive technologies for improving postnatal health knowledge, and enhancing postnatal care experiences for underserved populations.