Using the Area Under the Curve Method to Model Gestational Weight Gain in a Safety Net Population

Presenter Biography

Anna Booman is a first-year student in the PhD Epidemiology program. After growing up in the Portland area, Anna attended undergraduate and graduate programs on the east coast and moved back to Portland for this program in 2020. She is interested in maternal health disparities and how events during pregnancy can have lasting impacts on the health of both mother and child.

Institution

OHSU

Program/Major

Epidemiology

Degree

PhD

Presentation Type

Presentation

Start Date

4-8-2021 12:02 PM

End Date

4-8-2021 12:13 PM

Persistent Identifier

https://archives.pdx.edu/ds/psu/35585

Keywords

gestational weight gain; area under the curve; trajectory modeling; methodology

Abstract

Introduction: Gestational weight gain (GWG) is a predictor of maternal and perinatal health outcomes, but standard techniques used to assess GWG do not provide an estimation of timing of weight gain. The area under the curve (AUC) is an uncommonly used method to model GWG that provides, in a single statistic, an understanding of both total GWG and the GWG trajectory.

Objective: To (a) describe feasibility and interpretation of the AUC method applied to EMR data in a safety net population and (b) compare results to commonly used GWG assessment techniques.

Methods: We applied the AUC method to EMR data from 40,085 low-income (up to 250% FPL) pregnant patients who received care at community health centers throughout the U.S, totaling 19,905,167 person-weeks. We compared interpretations with standard methods used to assess GWG.

Results: Mean AUC (pound-days) was 2064.9 with a range from 2702 among women with an underweight pre-pregnancy BMI status to 1131.8 among women with an obesity class III pre-pregnancy BMI status. Mean total GWG was 25.1 pounds with a range from 31.1 among women with an underweight pre-pregnancy BMI status to 14.3 among women with an obesity class II pre-pregnancy BMI status. While interpretation of results from the AUC method is difficult and comparability across studies is not yet possible, the characterization of GWG trajectories is a defining feature.

Significance: This project will provide a tutorial of applying the AUC method to pregnancy data, providing future researchers with the information needed to use this method and understand results.

This document is currently not available here.

Share

COinS
 
Apr 8th, 12:02 PM Apr 8th, 12:13 PM

Using the Area Under the Curve Method to Model Gestational Weight Gain in a Safety Net Population

Introduction: Gestational weight gain (GWG) is a predictor of maternal and perinatal health outcomes, but standard techniques used to assess GWG do not provide an estimation of timing of weight gain. The area under the curve (AUC) is an uncommonly used method to model GWG that provides, in a single statistic, an understanding of both total GWG and the GWG trajectory.

Objective: To (a) describe feasibility and interpretation of the AUC method applied to EMR data in a safety net population and (b) compare results to commonly used GWG assessment techniques.

Methods: We applied the AUC method to EMR data from 40,085 low-income (up to 250% FPL) pregnant patients who received care at community health centers throughout the U.S, totaling 19,905,167 person-weeks. We compared interpretations with standard methods used to assess GWG.

Results: Mean AUC (pound-days) was 2064.9 with a range from 2702 among women with an underweight pre-pregnancy BMI status to 1131.8 among women with an obesity class III pre-pregnancy BMI status. Mean total GWG was 25.1 pounds with a range from 31.1 among women with an underweight pre-pregnancy BMI status to 14.3 among women with an obesity class II pre-pregnancy BMI status. While interpretation of results from the AUC method is difficult and comparability across studies is not yet possible, the characterization of GWG trajectories is a defining feature.

Significance: This project will provide a tutorial of applying the AUC method to pregnancy data, providing future researchers with the information needed to use this method and understand results.