Pragmatic predictive model of placental dysfunction to guide key clinical decisions

/ABSTRACT Placental dysfunction causes more than half of preventable stillbirths in the United States. Placental dysfunction also causes preeclampsia and fetal growth restriction, both of which can represent near misses for stillbirth. Tools for preventing stillbirth due to placental dysfunction include heightened fetal surveillance and recommending early delivery for known at risk pregnancies. However, these tools are only effective if we can accurately detect which pregnancies are at increased risk of placental dysfunction and stillbirth. Thus, accurately predicting risk of placental dysfunction is key to stillbirth prevention. The current national guidelines to identify high-risk pregnancies are based on presence of individual risk factors and do a poor job of accurately categorizing pregnancies. This is in part due to lack of quantitative assessment of cumulative risk or use of biomarkers of placental health. A predictive model can both calculate cumulative risk from multiple factors and incorporate biomarkers into risk calculations. If we can accurately calculate individual pregnancy risk of placental dysfunction, we will be better able to prevent stillbirths using fetal surveillance and early delivery. Our long-term goal is to improve prediction and prevention of stillbirth. We propose using a predictive model to determine individual pregnancy risk of placental dysfunction. The main objective of this project is to externally validate and optimize our existing clinical predictive model of placental dysfunction to enable individual risk assessment for preventable stillbirth. Aim 1: Validate and update a novel predictive model of individual risk of placental dysfunction. Aim 2: Measure incremental benefit of adding biomarkers of placental health to the predictive model and determine the optimal combination of biomarkers to guide fetal testing and timing of delivery. Aim 3: Conduct cost-effectiveness analysis (CEA) of clinical implementation of predictive models of placental dysfunction, both with and without key biomarkers. Predictive models calculate a predicted probability, or risk, of a specific outcome. With knowledge of individual risk, use of tools to prevent stillbirth can be customized by degree of predicted risk (i.e., recommending the most intense surveillance and earlier delivery to those at highest risk). To test Aims 1 and 2, we will enroll 640 pregnant people and follow them throughout pregnancy, checking biomarkers (ultrasound and blood tests) at three time points. We will then identify the combination of biomarkers and clinical variables that best predict placental dysfunction. We have selected biomarkers that are already clinically available to ease future translation to real world use. To achieve Aim 3, we will build decision analytic models for using predictive models to inform decisions regarding both use of antenatal surveillance and timing of delivery. Models will be prioritized based on performance, cost- effectiveness, and practicality. Upon completion of this project, we will have an updated, optimized predictive model of placental dysfunction including clinically meaningful biomarkers along with CEA to inform the economics of implementation in future projects. Project Number: 1R01HD115604-01A1 | Fiscal Year: 2025 | NIH Institute/Center: Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) | Principal Investigator: Karen Gibbins | Institution: OREGON HEALTH & SCIENCE UNIVERSITY, PORTLAND, OR | Award Amount: $525,531 | Activity Code: R01 | Study Section: Reproductive, Perinatal and Pediatric Health Study Section[RPPH] View on NIH RePORTER: https://reporter.nih.gov/project-details/1R01HD11560401A1