Association Between Recommended Preconception Health Behaviors and Screenings and Improvements in Cardiometabolic Outcomes of Pregnancy

Kaitlyn K. Stanhope, PhD, MPH; Michael R. Kramer, PhD, MMSc


Prev Chronic Dis. 2021;18(1):e06 

In This Article


Study Population

We used data from all sites participating in phase 8 of the Pregnancy Risk Assessment Monitoring System (PRAMS), comprising a cross-sectional, representative sample of births for each included site (state or US territory) for years 2016–2017. CDC releases sites with at least 55% response rates; 39 sites during 2016–2017 had a sufficient response rate. Each site participating in PRAMS randomly selects women who have had a live birth in the past year to complete a survey about their preconception, prenatal, and postpartum behaviors, health outcomes, and experiences.[21] PRAMS is weighted to give a representative sample of all women who gave birth in participating sites in those years. For this analysis, we included all women who had complete information on gestational diabetes, gestational hypertension, or both, as well as all included covariates In total, 8.1% of observations were excluded because of missing data (n = 1,105 excluded for missing information on GDM or HDP; n = 4,945 excluded for missing race/ethnicity or insurance). The final analytic sample of unweighted observations was 68,493. Compared with women who were included in the analysis, women who were excluded were more likely to have public insurance before their pregnancy (39.1% [SE, 0.26] vs 29.1% [SE, 1.84]) and less likely to report non-Hispanic white race/ethnicity (42.7% [SE, 1.32] vs 56.9% [SE, 0.27]). For analyses of GDM, we excluded an additional 2,617 participants who reported pregestational diabetes (analytic data set: n = 65,876 unweighted observations). For analyses of associations with HDP, we exclude an additional 4,350 participants reporting pregestational hypertension (analytic data set: n = 64,143 unweighted observations).


Our outcomes were 1) whether a woman had experienced GDM during this pregnancy and 2) whether a woman had experienced HDP during this pregnancy. Previous research has shown low positive predictive value (~50%) for self-report of pregnancy complications on PRAMS.[22] Thus, for GDM and HDP, we used data from the birth certificate, which is linked to PRAMS survey responses for all participants. Both GDM and HDP have acceptable (>80%) positive predictive values on the birth certificate when compared with medical record diagnoses.[23]

CDC considers 38 preconception health indicators, 23 of which are measured on PRAMS.[3] Of these, we excluded 8 indicators that applied only to multiparous women or reflected postpartum recommendations and 2 for which we did not have data (ie, emotional abuse and use of assisted reproductive technology) resulting in 13 indicators across 7 domains: health care (health-care coverage in the month before pregnancy; and routine check-up, teeth cleaning, and advice to improve health before pregnancy in 12 months before pregnancy); pregnancy intention; tobacco and alcohol use (any smoking or drinking in 3 months before pregnancy); nutrition and physical activity, including prepregnancy body mass index (BMI: underweight [<18.5 kg/m2], normal weight [18.5–24.9 kg/m2], overweight [25.0–29.9 kg/m2], or obese [≥30.0 kg/m2]), folic acid supplementation, and exercising at least 3 times per week in 3 months before pregnancy); mental health (clinical care for depression or anxiety in 12 months before pregnancy); emotional and social support (physical abuse in 12 months before pregnancy); and chronic conditions (diabetes or hypertension diagnosis at any time before pregnancy).


All analyses were conducted using SAS survey procedures version 9.4 (SAS Institute) or SUDAAN (RTI International) to account for survey weights. To understand sources of variation in the prevalence of GDM and HDP, we first calculated the prevalence of GDM and HDP across maternal demographic and socioeconomic characteristics and the prevalence of each preconception recommendation across maternal race/ethnicity. Because of the small number of Hawaiian/Pacific Islander women who met inclusion criteria, we do not present estimates on them, although they are included in the total number of women in the final sample. For multivariable analyses, we grouped American Indian/Alaska Native women, Hawaiian/Pacific Island women, Asian women, and Other/Mixed women.

To estimate associations between recommended preconception indicators and GDM/HDP, we fit logistic models between each indicator and each outcome (GDM/HDP) separately. We present adjusted odds ratios (ORs) and 95% CIs. For GDM, we excluded women who reported a diagnosis of diabetes before pregnancy. For HDP, we excluded women who reported a diagnosis of hypertension before pregnancy. Because maternal preconception health indicators and perinatal risk are correlated with women's socioeconomic status, including demographics and access to care, we selected certain variables a priori for adjustment. In each model, we adjusted for maternal age (≤17, 18–19, 20–24, 25–29, 30–34, 35–39, ≥40), race/ethnicity (non-Hispanic white, non-Hispanic black, Hispanic, non-Hispanic other), prepregnancy insurance (private, public, or none), prepregnancy BMI, and whether the woman reported attending a check-up with an obstetrician/gynecologist (OB/GYN) or primary care physician before pregnancy.