Improving Maternal Health Equity and Outcomes Through the Development of a Clinician-Informed Algorithm: A...

Increasing proportions of adverse maternal health outcomes occur in the 12-month postpartum period and could be addressed in outpatient settings. Our objective was to develop and test an algorithm to support a population health tool to identify high-risk prenatal patients served by federally qualified health centers (FQHCs).

We leveraged human-centered design to develop and test the population health tool and algorithm. We conducted focus groups and a literature search to identify risk criteria for the tool. To evaluate the tool, we conducted structured interviews and predictive modeling to compare the recall between the original tool and the refined algorithm. The population health tool was initially tested using electronic health record (EHR) data at six pilot FQHCs. To test the model's predictive capacity, we expanded to 18 FQHCs. Focus group participants included FQHC clinicians and staff. Data to evaluate the population health tool were queried from prenatal patients receiving care at participating FQHCs. The primary outcomes were adverse outcomes addressed in outpatient settings and health care utilization within 12 months postpartum.

Two focus groups (N = 7) were conducted to inform the implementation. In follow-up interviews (n = 6), users highlighted the tool's utility for identifying high-risk patients. In the predictive models (N = 82,829), the adverse outcome recall increased by 16%, but the algorithm only correctly predicted 42% of adverse outcomes experienced. The postpartum visit recall increased by 45%, with the algorithm correctly predicting 96% of visits utilized.

Results of this project highlight the importance of a deep understanding of EHR data capture and the involvement of clinicians when developing, testing, and evaluating interventions aimed at optimizing care for vulnerable patient populations. Future research should incorporate inpatient, outpatient, and social determinants data to develop a more comprehensive understanding of maternal health risk in the postpartum period.