Predictors of pregnancy-associated change in physical activity in a rural white population.

OBJECTIVE: The objective of this study was to examine the relationship of sociodemographic and psychosocial characteristics and exercise prior to pregnancy to change in physical activity (PA) during pregnancy.
METHODS: This study was part of the Bassett Mothers' Health Project (BMHP), a prospective cohort study of women who enrolled for prenatal care in the Bassett Healthcare system between November 7, 1994, and November 15, 1996. Six hundred twenty-two women met the eligibility criteria and were enrolled in the study. Sociodemographic and psychosocial characteristics and exercise behavior were assessed prenatally using the medical record and an anonymous mailed questionnaire. Staged linear regression analysis was used to predict change in physical activity with the psychosocial and sociodemographic characteristics as the independent variables and prepregnancy exercise frequency as a control variable.
RESULTS: Women who were older, more educated, and had higher household incomes were more likely to exercise prior to pregnancy (p < .05). Change in PA after becoming pregnant was associated with prepregnancy exercise frequency (p < .001). Women who exercised frequently moderated their activity, while sedentary women maintained or increased their PA after becoming pregnant. Exercise self-efficacy (the belief in one's ability to exercise regularly) and body mass index (BMI) were positive predictors of change in PA (p < .05), while prepregnancy exercise frequency remained significant as a control variable.
CONCLUSION: Interventions that aim to help women maintain or increase their PA during the perinatal period should be designed to increase exercise self-efficacy.