Pregnancy and childbirth are complex physiological processes.While most pregnancies progress without serious issues, some result in complications that can endanger both pregnant person and baby. Maternal complications—such as severe hemorrhage, hypertensive disorders, and preterm birth—are among the leading causes of maternal and neonatal morbidity worldwide. Early identification of high-risk pregnancies is crucial for timely interventions that can reduce these risks, but accurately predicting which patients will develop complications remains a challenge.

Clinical risk factors like maternal age, BMI, diabetes, and hypertension play a role in determining maternal health outcomes, but their interactions are intricate, and traditional screening methods often fall short in capturing the full complexity of risk. Moreover, some pregnancies that appear low-risk can still result in unexpected complications, making it difficult for healthcare providers to allocate resources and tailor care appropriately. This challenge is further compounded by disparities in healthcare access, which can leave some patients at higher risk of preventable complications.

Historical perinatal data can be leveraged to develop predictive models that assess maternal complexity and stratify risk, enabling clinical teams to make well-informed decisions about patient care. Our predictive model builds upon the Maternal Complexity Score (MCS), which has already been linked to key maternal and neonatal outcomes, including C-section rates, severe maternal adverse events, and neonatal morbidity. By refining this scoring system, we incorporate a broader range of factors, enhancing its predictive accuracy.

A major challenge in developing predictive models for maternal health is ensuring their clinical relevance. While increased data can improve prediction accuracy, it is crucial that predictions are available early enough in pregnancy to guide care decisions. To address this, we collaborate closely with clinicians to identify data that is both useful and accessible at different stages of pregnancy, striking a balance between predictive power and practical application.