Comprehensive Summary
This paper, from Spiel, highlights the potential of AI to improve health outcomes in cardio-obstetrics, particularly in pregnant individuals with cardiovascular conditions. Many physiological changes occur during pregnancy, which may make it difficult to identify these conditions in early stages, increasing the mortality rate. Preexisting risk models can fall short of detecting these risks, but AI could be trained to more effectively assess risk. AI could be very helpful in detecting many conditions at their early stages, such as arrhythmias, peripartum cardiomyopathy, preeclampsia, hypertensive disorders, spontaneous coronary artery dissection, pulmonary hypertension, valvular disease, and acute coronary syndrome. Spiel also calls for great care to be taken when AI solutions are developed for cardio-obstetrics, particularly when it comes to including data surrounding underrepresented groups and avoiding bias, and not doing so could diminish the positive effects of AI. Spiel additionally identifies barriers such as fragmented data and the process of testing and implementing AI solutions into the healthcare workforce.
Outcomes and Implications
AI could greatly improve early detection or cardiac issues, especially when they are masked by pregnancy, and help improve risk analysis. This would be particularly true if special care is taken in considering the diversity of patient populations and using data from underrepresented groups in addition to from the majority. AI has the potential to greatly improve outcomes, but must be extensively tested with care before this is done.