Use of Machine Learning for Early Detection of Maternal Cardiovascular Conditions: Retrospective Study Using...

Area of Science:

• Cardiology
• Obstetrics
• Data Science

Background:

• Cardiovascular conditions are the leading cause of maternal mortality in the US.
• The US has a high maternal mortality rate, disproportionately affecting minority groups.
• Early detection of cardiovascular risks in pregnancy is crucial.

Purpose of the Study:

• To evaluate the Healthy Outcomes for all Pregnancy Experiences-Cardiovascular-Risk Assessment Technology (HOPE-CAT) ML algorithm.
• To assess the capability and timing of HOPE-CAT in detecting maternal cardiovascular conditions.

Main Methods:

• Retrospective analysis of deidentified EHR data using the HOPE-CAT ML algorithm.
• Algorithm trained on expert-selected risk factors and current standards.
• Risk profiles generated and compared with clinical outcomes and diagnoses.

Main Results:

• HOPE-CAT identified risk factors an average of 56.8 days before diagnosis or intervention.
• Strongest performance in early detection of myocardial infarction.
• Preeclampsia was the most common condition identified.

Conclusions:

• ML enhances early detection of cardiovascular conditions in obstetrics.
• ML synthesizes patient data to support provider decision-making.
• This approach may help reduce maternal health disparities.