Human-Centred Trust in AI-Supported Risk Communication for High-Risk Pregnancy Care
International Journal of Drug Delivery Technology · 19.06.2026 · 16 (56s)
DOI: 10.25258/ijddt.16.56s.91 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Drug Delivery Technology, online erschienen am 19.06.2026. Verfasst von Wang Chunxiang und 2 weiteren Autor:innen.
Abstract (Originalsprache)
Digital prediction tools are progressively being utilized in the field of high-risk pregnancy care to diagnose preeclampsia, gestational diabetes and preterm birth and intrapartum fetal compromise. However, discrimination statistics are not the sole determinants of the clinical value of artificial intelligence. Development of trust in risk communication is based on how a prediction is explained, verified, governed and connected to collective action. This is a secondary fast review that analyzes the human-centred trust in AI-supported risk communication during high-risk pregnancy care. Thematic analysis coding was the primary means of coding evidence across patient-facing pregnancy digital tools, preeclampsia prediction, prediction of gestational diabetes, preterm prediction and cardiotocography/acidemia detection. Four themes were created: a high level of trust exists when AI assists rather than substitutes for clinicians; explainability and communication translate risk scores into actionable interpretations; prediction models have excellent potential but weak methodology; and equity, privacy and workflow suggest trust to levels beyond AUROC. This review concludes that AI can reinforce risk communication through providing earlier review, personalised counselling and regular escalation. Prospective impact, calibration, subgroup performance, clinician override, anxiety effects and patient understanding are not well reported in model-centred evidence, however. It is not an anti-AI argument based on the argument here. Instead, responsible obstetric AI needs to be developed as a communication and care-relationship technology. Human-centered implementation, as part of it, should include local validation, explainable output, clinician accountability, privacy protection, an equity audit, and plain-language shared decision-making.
Zitieren
APA
Chunxiang, W., Nasution, A. N., & Liena, L. (2026). Human-Centred Trust in AI-Supported Risk Communication for High-Risk Pregnancy Care. International Journal of Drug Delivery Technology, 16(56s). https://doi.org/10.25258/ijddt.16.56s.91
Vancouver
Chunxiang W, Nasution AN, Liena L. Human-Centred Trust in AI-Supported Risk Communication for High-Risk Pregnancy Care. International Journal of Drug Delivery Technology. 2026;16(56s). doi:10.25258/ijddt.16.56s.91
BibTeX
@article{Chunxiang2026,
author = {Chunxiang, Wang and Nasution, Ali Napiah and Liena, Liena},
title = {Human-Centred Trust in AI-Supported Risk Communication for High-Risk Pregnancy Care},
journal = {International Journal of Drug Delivery Technology},
year = {2026},
volume = {16},
number = {56s},
doi = {10.25258/ijddt.16.56s.91}
}
EUVIMED