Research and Development of AI-Powered Task-Shifting for High-Quality Obstetrics Ultrasound Services in Community Healthcare Settings: Real-World Implementation Experiences from an LMIC
Oxford Open Digital Health · 13.11.2026
DOI: 10.1093/oodh/oqag025 ↗Kurz gefasst
Fachartikel in der Zeitschrift Oxford Open Digital Health, online erschienen am 13.11.2026. Verfasst von Ranjana Koirala und 12 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Blind Sweep Obstetric Ultrasound (BSOU) paired with artificial intelligence (AI) is increasingly proposed as a scalable approach to expand antenatal screening in low- and middle-income countries (LMICs). However, limited evidence exists on how to implement BSOU-AI systems within real clinical environments, where infrastructure, workflows, and human resource constraints differ substantially from research settings. This study reports implementation experiences from a multi-center BSOU-AI project across tertiary hospitals and community health facilities in Nepal. We collected more than 3000 BSOU datasets using minimally trained providers and handheld ultrasound devices, developed and iteratively refined a nine-sweep BSOU protocol, and coordinated with national stakeholders to ensure alignment with the Rural Obstetrics Ultrasound Program. Real-world deployment revealed major challenges including machine incompatibility, inadequate digital infrastructure, annotation bottlenecks, workflow mismatches in high-volume hospitals, and difficulty capturing rare complications which required continuous adaptation of protocols, staffing, and data-management systems. Engagement with Nepal’s Ministry of Health prompted the addition of an evaluation arm comparing BSOU-AI performance not only to radiologist scans but also to nurse-led rural ultrasound, reflecting the true standard of care. Our experience demonstrates that AI readiness in LMICs depends as much on implementation science as on algorithm development. We provide methodological insights and practical recommendations for designing BSOU-AI studies that are scalable, contextually grounded, and responsive to health-system realities.
Zitieren
APA
Koirala, R., Bhandari, P., Luitel, N., Poudel, K., Acharya, B., Ghimire, A., Karki, R. K., Chaulagain, A., Sherpa, M. G., Poudel, A., Dhakal, P., Regmi, P. R., & Khanal, B. (2026). Research and Development of AI-Powered Task-Shifting for High-Quality Obstetrics Ultrasound Services in Community Healthcare Settings: Real-World Implementation Experiences from an LMIC. Oxford Open Digital Health. https://doi.org/10.1093/oodh/oqag025
Vancouver
Koirala R, Bhandari P, Luitel N, Poudel K, Acharya B, Ghimire A, et al. Research and Development of AI-Powered Task-Shifting for High-Quality Obstetrics Ultrasound Services in Community Healthcare Settings: Real-World Implementation Experiences from an LMIC. Oxford Open Digital Health. 2026. doi:10.1093/oodh/oqag025
BibTeX
@article{Koirala2026,
author = {Koirala, Ranjana and Bhandari, Prasiddha and Luitel, Nishant and Poudel, Kanchan and Acharya, Bishram and Ghimire, Angelina and Karki, Rina Kumari and Chaulagain, Amrita and Sherpa, Mingmar G and Poudel, Atit and Dhakal, Prajwal and Regmi, Pradeep Raj and Khanal, Bishesh},
title = {Research and Development of AI-Powered Task-Shifting for High-Quality Obstetrics Ultrasound Services in Community Healthcare Settings: Real-World Implementation Experiences from an LMIC},
journal = {Oxford Open Digital Health},
year = {2026},
doi = {10.1093/oodh/oqag025}
}
EUVIMED