Psychosocial, Sociodemographic, and Biomedical Predictors of Prenatal Risk Behaviour
The negative health effects of women's drinking, smoking, other drug use, and inadequate diet during pregnancy on the fetus have been well documented. Another body of research suggests that such psychosocial factors as stress, anxiety, social interactions, and maternal attitude towards pregnancy may also affect reproductive outcomes. This may be due, at least in part, to the effects of these psychosocial factors on health behaviours such as those just mentioned. The present study brings together these two previously separate lines of research, by comprehensively examining the relationship of psychosocial and biomedical factors to prenatal risk behaviour. The study, based on baseline data from a multi-wave longitudinal study, focused on five prenatal risk behaviours: alcohol drinking, smoking, psychoactive drug use, caffeinated beverage consumption, and inadequate diet. The purpose of this study was three-fold: (a) to examine the psychosocial, sociodemographic, and biomedical predictors of change in prenatal risk behaviour from pre-pregnancy to first trimester, (b) to examine the independent predictors of prenatal risk behaviour during pregnancy, and (c) to test an integrated, a priori model of prenatal risk behaviour. A face-to-face computer-assisted interview during the second trimester of pregnancy collected data from 452 prenatal clients of a city-wide public health unit (80% participation rate). This sample consisted of mostly primiparous women (59%), ranging in age from 15 to 39 years (56% 20-29 years),60% of whom had a high school education or less, and 17% of whom were Native. The questionnaire, administered by trained female interviewers, covered four broad areas--psychosocial, biomedical, sociodemographic, and behavioural variables. Significant change in risk behaviour was reported from pre-pregnancy to the first trimester, with 41 % giving up alcohol and 39% quitting drug use; while only 9% of smokers and consumers of caffeinated drinks gave up these habits. On the other hand, sizable numbers of women continued to engage in risk behaviours during early pregnancy: 47% continued to smoke, 46% drank alcohol, 23% consumed caffeine equivalent to three or more cups a day, and 13% used psychoactive drugs. Three types of multivariate analysis were carried out to address the study objectives. First, in terms of significant predictors of behaviour change, discriminant function analysis found that experiencing fewer negative life events in the preceding 12 months and higher desirability of pregnancy were associated with quitting drinking, smoking, and using drugs. Also, women whose partners did not smoke were more likely to give up drinking and smoking. In addition to these variables, each type of behaviour change had unique predictors. Second, logistic regression analysis found that the strongest two predictors of risk behaviour during the first trimester were the respondent's level of education attained and the smoking status of the father of the baby. Those women who did not complete high school were about 3.6 times more likely to engage in three or more risk behaviours, and those who reported that their partner smoked were 3.4 times more likely to engage in a higher number of risk behaviours than their respective counterparts. Other significant predictors of risk behaviour were: Native ethnicity, having experienced a previous miscarriage, drinking as a social activity, conflicted support network, and higher number of negative life events. Third, structural equation modeling tested an a priori model of prenatal risk behaviours. Two types of stressors, negative life events and social stressors (operationalized as level of income adequacy, years of education, and marital status) were found to have a strong direct impact on the number of risk behaviours, as well as an indirect effect, mediated by such psychosocial factors as mastery, emotional social support, self-esteem, and desirability of pregnancy. These psychosocial factors in tum influenced perceived stress, which also had a direct effect on behaviour. The third hypothesized type of stressor, reflecting biomedical risk based on problems in previous pregnancies, had only an indirect impact on behaviour, by increasing perceived stress. The findings of this study illuminate the needs of particular groups of pregnant women (e.g., Native, socioeconomically disadvantaged), and the importance of understanding risk behaviour within the structural, cultural, and psychosocial realities of their lives. What are the implications of the study findings for health promotion during pregnancy? The finding that prenatal risk behaviours are not randomly distributed across the population, but are socially patterned and interconnected, suggests that approaches to health promotion during pregnancy need to be broad-based and include multiple strategies. These strategies should focus not only on individual women and their immediate environment (i.e., family and friends), but also involve the broader community and public policy levels.