Women enrolled in this study were participants in The First 1,000 Days program, a systems-level initiative that engages stakeholders across clinical and public health sectors to reduce the prevalence of obesity and obesity risk factors among mother-infant pairs who are low-income by addressing the levels of individual, family, and socio-contextual factors that hinder progress in obesity prevention. The system-wide intervention begins when women initiate prenatal care in their first trimester of pregnancy and offers support for mothers, their partners, and eventually the mother-partner-infant triads, throughout pregnancy and the child’s first 24 months. The conceptual framework, intervention design, evaluation methods, and primary results have been described in detail elsewhere.15,16 A subset of women completed surveys in their first and third trimesters of prenatal visits to examine individual-level changes in behavior and psychosocial outcomes. For this secondary analysis, a quasi-experimental pre-post design was used to evaluate the changes in the behavior and psychosocial outcomes from the first to third trimester visit among participants.
Eligibility and Recruitment
The First 1,000 Days program was offered to women who initiated prenatal care between August 2016 and September 2017 in three community health centers in Boston, Revere, and Chelsea, Massachusetts that serve predominantly low-income, racially/ethnically diverse populations. Upon completion of the intake survey and consent, women were considered to be enrolled in the program. The intake survey was administered to women at their first prenatal care visit which for most women was during the first trimester. Of the 366 women who completed a first trimester intake survey, 286 (78%) also completed a third trimester survey, and 264 (72%) women were included in the analyses as 22 women had missing vital demographic data (Fig. 1). The First 1,000 Days study protocol was approved by Partners HealthCare Institutional Review Board and registered at ClinicalTrials.gov (NCT03191591).
Program Components in Pregnancy
The First 1,000 Days program has multiple components that aim to improve primary and secondary prevention of obesity. The program components that have been previously described in detail and include: staff and provider training emphasizing obesity prevention efforts; clinical decision support tools to track gestational weight gain; first prenatal visit universal screening for health behaviors and socio-contextual factors; patient navigation focused on healthy behavior change, social needs, and clinical and public health services; and health coaching for women at high risk of obesity.15,16
During pregnancy, the program focused on five behavior targets including: eating a balanced nutrition plan; drinking predominantly water and avoiding sugary-drinks; being physically active; getting recommended amounts of sleep; and reducing stress through social supports. Information regarding the behavior targets was delivered through printed materials including posters hanging in health centers and public health offices and individual booklets provided to patients. Booklets were available in English, Spanish, Vietnamese, and Arabic and contained customizable sections for gestational weight gain recommendations and behavior changed goal setting (Fig. 2). Women could also enroll in a text-messaging program to provide behavior change support and education and received 2–3 text messages during their pregnancy. Short informational videos (Vidscrips®) were also created in English and Spanish and available to women and their partners. The videos reinforced the behavior targets of the program, answered commonly asked questions, and provided recommendations.
Data Collection and Outcome Measurements
Information was collected through self-administered questionnaires during the first and third trimester of gestation and from electronic health records (EHR). Measures collected included behavior (i.e., diet, physical activity and screen time) and psychosocial outcomes, as well as WIC program enrollment. Surveys were available in English, Spanish, and Arabic.
Dietary behaviors, including fruit and vegetable, sugary-drinks, and fast food consumption were evaluated by asking women, “During the past 7 days, on average, how often did you eat …?”. Women responded by selecting: never; once per week, 2–4 times per week, nearly daily, 2–4 times per day, or 5 or more times per day. The fruit and vegetable consumption question included fresh, cooked, canned, or frozen fruits or vegetables and excluded juices or dried fruits; mean consumption was measured in times per day. The sugary-drink question included fruit-flavored drinks, juice from concentrate, punch, Kool-Aid, soda, sports drinks, sweet tea or coffee drinks, and sweetened milks; mean consumption was measured in beverages consumed weekly. The fast-food question asked about eating from a fast food restaurant; mean consumption was measured in weekly consumption. The items are from a validated food frequency questionnaire and have been previously used during pregnancy.17,18
To evaluate physical activity, women answered, “During the past 7 days, on how many days were you physically active for a total of at least 30 minutes per day? Add up all the time you spent in any kind of physical activity that increased your heart rate and made you breathe hard some of the time.” Responses to this question ranged from 0–7 days per week. This question was adapted from the Youth Risk Behavior Survey.19
To measure screen time, women were asked, “During the past 7 days, on average, how many hours per day did you usually spend watching TV or videos. Include time spent watching on a television, computer, phone or tablet.” Women selected: never, < 1 hour per day, 1, 2, 3, 4 or 5 hours per day, and 6 or more hours per day. This question was adapted from the Nurses’ Health study.20
We assessed pregnancy-related anxiety using the Pregnancy Anxiety Score, which focused on 5 topics: the extent of worry or concern about their health during pregnancy, about their baby’s health normal growth and developing, about losing the baby, about having a hard labor or difficult delivery, and about taking care of a new baby21,22 Possible responses to these questions were: never concerned, sometimes, most of the time, and almost all of the time concerned. A score was calculated by summing each of the five questions with a possible range from 0–20. Higher scores indicate a greater extent of worry or concern. If more than two questions were missing, a score was not calculated.
WIC program enrollment was offered to participants during pregnancy who reported they were not enrolled and met the income level criteria. To evaluate the number of women receiving WIC program support, we asked women, “Do you currently receive benefits from WIC?”. Women responded by answering: yes, no, or unsure.
Based on the previous literature, we used covariates for adjusted analyses.7,23,24 Socio-demographic variables collected in the first and third trimester survey were used and they included maternal age and race/ ethnicity. Pre-pregnancy body mass index (BMI) was collected from the EHR.
Data from the EHR and survey results were merged to allow for analyses. We compared participants’ responses at baseline (first trimester of gestation) with results from third trimester surveys; each participant was measured twice, resulting in pairs of observations. We used repeated measures design, including the paired t-tests for quantitative variables and the McNemar’s test for qualitative variables in unadjusted models. We also used multivariable regression models to adjust for possible confounders, including maternal age, race/ethnicity and pre-pregnancy BMI. Mixed-effect linear models were applied for continuous outcomes, with a time predictor to indicate the time points of first and third trimesters. The models accounted for clustering of observations within individuals. The adjusted mean differences between first and third trimester and 95% confidence intervals (CI) were calculated. For the dichotomous outcome, WIC enrollment, we applied logistic regression using generalized model, fitted with generalized estimating equation to address the repeated measurements. The adjusted odds ratio (OR) and 95% CI were generated for the WIC enrollment outcome. We performed analyses on complete cases and excluded participants with missing values in predictor, outcome, or covariates. A 2-sided alpha level of 0.05 was used to test for statistical significance in all analyses. Analyses were performed in RStudio 3.5.1 and SAS 9.4 (SAS institute, Cary, NC).