Methods for modeling gestational weight gain: empirical application using electronic health record data from a safety net population

Background: Understanding the risks and effects of gestational weight gain (GWG) is a prominent area of perinatal research but approaches for quantifying GWG are evolving and remain underdeveloped, especially in clinical settings for underserved demographic subgroups. To fill this gap, we demonstrated and compared six GWG metrics across pre-pregnancy BMI classifications: total GWG, trimester-specific linear rate of GWG, adherence to total and trimester-specific recommendations, area under the curve, and GWG for gestational age z-scores. Methods: We used clinical data on 44,801 pregnant people from community-based health care organizations with extensive longitudinal measures and substantial representation of understudied subgroups. Results: Total GWG was lower in individuals with higher pre-pregnancy BMI; yet more temporally resolved analyses revealed differences in trimester-specific weight change. Differences included common first trimester weight loss in people with pre-pregnancy class II or III obesity and substantial first trimester weight gain in people with pre-pregnancy underweight, with the greatest pre-pregnancy BMI-related variation in GWG occurring in the second trimester. These differences are reflected to varying degrees in the AUC and GWG z-score metrics. Conclusions: Our findings inform development of GWG guidelines within BMI categories, especially in obesity subclasses and underweight, and selection, refinement, and application of GWG metrics in future research. GWG metrics differ to varying degrees across BMI categories in a population consisting of several underserved subgroups: pregnant people of color, with larger body sizes, or with lower incomes. Stronger evidence on safe levels of first trimester weight loss and obesity class-specific recommendations is needed.