|| Checking for direct PDF access through Ovid
The purpose of the retrospective epidemiologic study was to determine smoking prevalence and its correlates among pregnant women. Data for the study were collected from 477 randomly selected medical records of pregnant women attending a Kentucky obstetrical clinic. Logistic regression analysis, when controlling for age and education, showed that significant predictors of smoking prior to pregnancy were being married and being a woman of color. Significant demographic predictors of smoking at initiation of antenatal care were being married, being a woman of color, early age at onset of smoking, and number of cigarettes smoked per day.