Project Details

Co-Designing Implementation Strategies for Diagnosing Pediatric Acute Lower Respiratory Infections using Artificial Intelligence-Enabled Digital Auscultation in Rural Bangladesh

Nadia Hoekstra, MD

Summary

BACKGROUND: Acute lower respiratory infections (ALRIs), including pneumonia, cause approximately 700,000 deaths annually in children under five, with the largest burden in low- and middle-income countries (LMICs). Many frontline providers in LMICs lack reliable diagnostic tools, leading to delayed treatment or unnecessary antibiotic use and contributing to poor outcomes and antimicrobial resistance. GAP: While AI-enabled digital stethoscopes show promise for improving ALRI diagnosis, key questions remain about how to effectively integrate these tools into routine pediatric care across diverse LMIC healthcare sectors, including unconventional access points like private pharmacies. Sector-specific determinants, workflow constraints, and adoption readiness are poorly characterized, and co-designed implementation strategies have not been prospectively piloted in resource-limited settings. HYPOTHESIS: We hypothesize that implementation determinants and preferred strategies in rural Bangladesh will differ by healthcare sector, with pharmacies requiring simpler workflows and clear referral triggers. We also hypothesize that co-designed strategies will be acceptable, appropriate, and feasible, achieving a fidelity threshold of ≥60% of eligible encounters across multiple sectors. METHODS: This mixed-methods study is embedded within the NIH-funded BLAAAST study in rural Bangladesh. We will use the Consolidated Framework for Implementation Research (CFIR) to identify sector-specific barriers and facilitators, conduct participatory co-design workshops with healthcare workers and caregivers, and pilot two-month implementation modules in six facilities (public clinics, private clinics, and pharmacies) using an AI-enabled digital stethoscope. RESULTS: Pending. IMPACT: This work will generate a customized, replicable toolkit for integrating AI-enabled diagnostics into frontline pediatric care in LMICs, informing the responsible scale-up of emerging pediatric diagnostic technologies in resource-limited settings globally. Optional/Additional Comments: This project addresses a critical implementation gap for AI-enabled pediatric diagnostics in LMICs and is among the first to develop and evaluate sector-adapted implementation strategies for digital diagnostic tools in rural Bangladesh. The study leverages established research infrastructure and partnerships through the Projahnmo Research Foundation and builds on over a decade of work developing AI-powered lung sound analysis for children.