JMIR Publications

ABSTRACT

Background:

A majority (>80%) of maternal deaths in the United States are preventable. Generation of machine learning models from electronic medical records (EMRs) is considered a promising approach to predict the risk of adverse maternal outcomes, which may enable proactive intervention to prevent maternal mortality. Current evidence syntheses of such machine learning approaches either focus only on specific maternal outcomes, aspects other than risk prediction, or do not consider the full pipeline of studies from the development of machine learning models to factors affecting implementation of clinical applications incorporating the models in practice.

Objective:

The goal of this scoping review is to document evidence for the use of machine learning models for predicting the risk of maternal morbidity and mortality outcomes (Research Objective, RO1), the translation of such models into applications in clinical use by providers (RO2), and factors associated with the implementation of clinical applications in practice (RO3).

Methods:

The review was limited to studies in healthcare settings, using data from EMRs. A detailed search string was developed in collaboration with an health sciences librarian, and implemented on February 20, 2023 on PubMed, CINAHL Plus, Scopus, Embase, and IEEE Xplore. Two reviewers independently reviewed titles and abstracts for inclusion, and conflicts were resolved by a third reviewer. Only full-length journal articles published in English were included. Studies using non-EMR data exclusively were excluded. Two reviewers independently reviewed full texts for inclusion, and conflicts were resolved by a third reviewer. A structured template was used for data extraction and findings were summarized descriptively.

Results:

In total, 734 studies were identified from the search, and 142 studies were included for full text review. Finally, 39 studies were included in the review. More than half of the included studies were conducted in 2022, and 34 studies were from just three countries (US, China, Israel). More studies focused on identifying the risk of pregnancy and delivery outcomes compared to post-partum outcomes. The top three most common outcomes for risk prediction were cardiovascular risks and hypertensive disorders of pregnancy (9 studies), gestational diabetes (7 studies), and post-partum hemorrhage (6 studies). Data were labeled with computable phenotypes in 30 studies, and the most often used method in machine learning models was boosting methods (18 studies). The most common metric used to assess model performance was Area Under the Curve, Precision-Recall (AUC/PR-AUC; 33 studies). No studies described clinical applications of machine learning models for providers (RO2) or associated implementation factors (RO3).

Conclusions:

Key recommendations for future research and practice are to expand efforts to examine less commonly studied maternal morbidity and mortality outcomes in the postpartum period, increase transparency and reproducibility of studies through consistent and detailed reporting of methodology (e.g., through implementation of reporting checklists) and to expand efforts to translate, implement, and evaluate the use of the machine learning models in clinical practice.