Managing maternity: Moving care, not patients, using artificial intelligence ( <scp>AI</scp> ), internet‐of‐things ( <scp>IOT</scp> ) and point‐of‐care testing ( <scp>POCT</scp> ) devices

Gen Lin Foo, Mahesh Choolani, Aris T. Papageorghiou, Hema Divakar, Hassan Shehata, Nir Melamed, Gabriel Davis Jones, Vyta Senikas, Eline M. Van der Beek, Nandita Palshetkar, Gabrielle Saccone, Vincenzo Berghella, Justin Konje, B.c. Bhatt, Augusto Cam, Moses Obimbo, Anne Beatrice Kihara, Moshe HodInternational Journal of Gynecology & Obstetrics 2026 · DOI

TL;DR

This paper explores how Artificial Intelligence (AI), Internet of Things (IoT), and Point-of-Care Testing (POCT) devices, collectively termed "FemTech," can revolutionize maternity care. It proposes a shift towards decentralized, personalized, and augmented care models that extend beyond the clinic, aiming to improve access, precision, and support for the healthcare workforce.

What Problem Does This Paper Address?

The current model of maternity care, characterized by rigid, in-person clinic visits, often fails to address social and structural health determinants, leading to disparities in access and outcomes. The COVID-19 pandemic highlighted the potential of telemedicine, but its adoption in obstetrics has been fragmented. Furthermore, existing AI applications in healthcare often face challenges related to data bias (historical exclusion of pregnant women from trials), the need for explainability, and the potential to widen the healthcare gap if not implemented equitably. This paper addresses the need for a strategic integration of AI-powered technologies to create a more accessible, personalized, and efficient maternity care system that complements, rather than replaces, human expertise.

Methods

The paper outlines a conceptual framework for AI-allied maternity care, focusing on three key domains:

  • Changing where women are cared for: This involves decentralizing care through AI-powered IoT devices and POCT, enabling remote prenatal monitoring and extending care into patients' homes. This approach aims to reduce reliance on clinic visits and improve access, particularly for underserved populations.
  • Changing how women are cared for: This domain focuses on personalized, adaptive care using predictive and reinforcement learning algorithms. These algorithms can analyze complex data to predict complications, project disease progression, and tailor treatment plans, moving beyond static risk models.
  • Changing the maternity workforce: AI can augment the workforce through generative tools for patient engagement, clinical decision support, and automation of tasks like ultrasound imaging. This aims to address staffing shortages and enhance efficiency while maintaining clinician oversight.

The paper also discusses the role of "agentic AI," which can act autonomously to search for information and trigger workflows, potentially reducing reliance on centralized cloud infrastructure.

Datasets

The paper mentions the "Oxford Maternity Dataset" as an example of a curated, high-quality training dataset that can help mitigate gender bias in AI models. It also refers to the need for "pregnancy-specific ML datasets" that are longitudinal and encompass pre-conception to postpartum periods. No new datasets are introduced in this paper.

Key Insights and Ideas

  • Decentralized Care is Key: Moving maternity care from the clinic to the patient's home via AI-powered IoT and POCT devices is a critical step towards equitable access.
  • Personalization Over Standardization: AI enables a shift from one-size-fits-all approaches to highly personalized care plans, adapting to individual patient needs and disease trajectories.
  • AI as a "Digital Colleague": AI technologies are positioned as augmentative tools for the maternity workforce, addressing staffing shortages and enhancing efficiency without replacing human empathy and clinical judgment.
  • The "FemTech" Revolution: The integration of AI, IoT, and POCT into women's health innovations ("FemTech") is poised to transform pregnancy care by addressing long-standing disparities.
  • Global Adoption Dynamics: The USA and China are leading in AI adoption in healthcare, driven by patents and publications, but equitable global integration requires careful consideration of economic and geopolitical factors.
  • Explainability and Governance are Crucial: For widespread adoption, AI in maternity care needs to be explainable, transparently validated, and governed by robust frameworks for safety and consent.

Key Results

The paper is primarily a conceptual and forward-looking review, outlining potential applications and future directions rather than presenting empirical results from specific AI models. It discusses the potential for:

  • Reducing reliance on clinic visits through home-based monitoring.
  • Enhancing diagnostic precision and supporting clinical decision-making.
  • Augmenting the maternity workforce to address shortages.
  • The cost of querying AI models is rapidly decreasing, potentially making advanced AI more accessible to lower-resource countries.

Limitations and Future Work

  • Explainability Challenges: The "black box" nature of some AI algorithms poses challenges for clinician understanding and liability. While Explainable AI (XAI) is discussed, full implementation may be difficult.
  • Data Bias and Standardization: The need for high-quality, diverse, and pregnancy-specific datasets is emphasized, along with the importance of multinational consortia for creating benchmark datasets.
  • Regulatory and Legislative Frameworks: The paper highlights the lack of specific AI-healthcare legislation and the need for frameworks to govern data use, algorithmic fairness, and privacy.
  • Global Disparities: The uneven pace of AI development between high-income countries and LMICs could lead to the latter becoming "technological colonies." Strategies for equitable deployment in LMICs are crucial.
  • Maintaining Human Skills: There's a risk that over-reliance on AI could lead to a decline in essential non-technical clinical skills among caregivers.
  • Future Work: The authors suggest the need for hybrid medical oversight mechanisms, safety-net based interventions, and regulatory compliance. They also emphasize the importance of close collaboration between clinicians and industry to develop AI solutions that enhance clinical practice while preserving the human aspects of care. Future adoption will depend on global economic and geopolitical dynamics, and equitable integration will require explainable AI, transparent validation, multinational benchmark datasets, and robust governance.

Summary updated

Abstract

The integration of artificial intelligence (AI) into healthcare is accelerating and maternity care is at a pivotal moment for the strategic implementation of these technologies. This article explores how AI-assisted women's health innovations, often termed "FemTech," may transform pregnancy care by addressing long-standing disparities: enhancing diagnostic precision and supporting the obstetric workforce. We outline three domains in which AI is poised to drive change: where women are cared for, how they are cared for, and who delivers their care. First, decentralized AI combined with Internet of Medical Things (IoMT) devices can extend prenatal monitoring into homes, reducing reliance on clinic visits and expanding access for underserved populations. Second, predictive and reinforcement learning algorithms enable personalized, adaptive care across the reproductive continuum, from preconception to postpartum, moving beyond static risk models and uniform treatment approaches. Third, AI has the potential to augment the maternity workforce by offering generative tools for patient engagement, clinical decision support and automation of ultrasound imaging, while ensuring clinician oversight remains central. Future adoption will depend on global economic and geopolitical dynamics, with the USA and China currently leading in patents, publications, and model development. Equitable integration will require explainable AI, transparent validation, multinational benchmark datasets, and robust governance on safety and consent. Ultimately, AI-powered technologies should complement, not replace human expertise, embedding digital innovation within a model of maternity care that preserves empathy and clinical judgment.
BibTeX
@article{foo2026managing,
  author = {Gen Lin Foo and Mahesh Choolani and Aris T. Papageorghiou and Hema Divakar and Hassan Shehata and Nir Melamed and Gabriel Davis Jones and Vyta Senikas and Eline M. Van der Beek and Nandita Palshetkar and Gabrielle Saccone and Vincenzo Berghella and Justin Konje and B.C. Bhatt and Augusto Cam and Moses Obimbo and Anne Beatrice Kihara and Moshe Hod},
  title = {{Managing maternity: Moving care, not patients, using artificial intelligence ( <scp>AI</scp> ), internet‐of‐things ( <scp>IOT</scp> ) and point‐of‐care testing ( <scp>POCT</scp> ) devices}},
  journal = {International Journal of Gynecology \& Obstetrics},
  year = {2026},
  doi = {10.1002/ijgo.70942},
}
Figure