This blog is written by student members of the Summer 2021 Prenatal Team. The team consists of Claire Dawson, Dipra Debnanth, Stephanie Ganzi, Leena Ghrayeb, Meghana Kandiraju, Amanda Naccarato, and Harini Pennathur.
Hello! This is the Prenatal team here at CHEPS! We work with Dr. Alex Friedman Peahl, a practicing OB-GYN physician at Michigan Medicine, to improve prenatal care. The current prenatal guidelines involve a large number of prenatal visits, yet the U.S. still leads in maternal mortality rates amongst developed nations. Dr. Peahl highlighted this discrepancy and emphasized the idea that a large number of prenatal visits may not actually improve postpartum outcomes, and that we may need to refocus our efforts away from providing a “one-size fits all” pre-determined care framework and rather toward providing tailored care to our patient population (through telehealth, social work, or other non-traditional services).
Our team utilized data of patients who delivered and utilized prenatal care at Michigan Medicine in a given year. We then assigned a medical and psychosocial risk score to each patient based on previous diagnoses and demographic information and used these scores to allocate patients to one of four categories: low psychosocial risk/low medical risk, low psychosocial risk/high medical risk, high psychosocial risk/low medical risk, and high psychosocial risk/high medical risk. Our team performs data analysis and is constructing a simulation to test prenatal policy changes and identify trends based on patient quadrant assignment. We hope to use this data to learn how we can best meet the needs of all types of patients requiring prenatal care.
This semester on the data analysis team, we have been working to determine the proportion of patients that may begin prenatal care as medically low risk, but develop pregnancy complications that bump them into the high medical risk category. It is important to address this population as they may progress into requiring more enhanced medical care than was previously determined. Our team is currently looking into three conditions: gestational diabetes, gestational hypertension, and pre-eclampsia. We are calculating the median age of diagnoses for each of the aforementioned complications, specifically focusing on the percentage of low medical risk patients transitioning to high medical risk. However, we are also analyzing each quadrant of risk-level, determining trends associated with an early/late diagnosis by risk level. This data analysis helps to support the simulation and is also used by Dr. Peahl in some of her academic papers to further highlight the importance of this work.
On the simulation team, we use data from the data analysis team as inputs to the simulation code. Our simulation runs for an entire year with randomly generated patients of various patient types entering the system on a weekly basis. Each patient has an associated pathway that indicates the number of appointments and types of appointments based on their patient type. In addition, our simulation contains providers, which are specific to clinic locations and are randomly assigned to patients. With that, we output the unused capacity of providers, overbooking of providers, longest patient delays, average weekly patient type arrival rate to see what the results of various tailoring pathways would have on the patients and the providers. We are currently working on establishing “dynamic patient types”: using data about pregnancy complications to analyze patients moving from low medical risk to high medical risk over the course of their pregnancies.
Overall, the data analysis team compiles prenatal statistics from historical data which the simulation team uses to manipulate various prenatal care steps, demonstrating how adding more variety within prenatal care could benefit different types of patients. We work together to create a more equitable system for all!