Publications

CONTEXT

Traditional measures of pregnancy intentions that are dichotomous and retrospective do not fully capture the complexity surrounding women's plans to become pregnant.

METHODS

During January-June 2008, 249 women aged 15-44 awaiting pregnancy test results at family planning clinics in Pittsburgh completed a survey containing both single- and multi-item measures of pregnancy intentions. Chi-square analyses were used to assess differences between subgroups of women.

RESULTS

Few women were trying to become or planning for pregnancy (11% on the single-item measure; 20% on the multi-item measure), while approximately one-third of the sample were not trying to become or planning for pregnancy (31% on the single-item and 36% on the multi-item measure). The single-item measure categorized more women as ambivalent about pregnancy (58%) than did the multi-item measure (44%). Of women categorized as ambivalent by the single-item measure, 62% were also categorized as ambivalent by the multi-item measure. Overall, 68% of responses to the two measures were concordant. With both measures, women who were not planning or trying for pregnancy were more likely than those who were planning for pregnancy or who were ambivalent to indicate that they planned to have an abortion if their test was positive (27-29% vs. 0-2%).

CONCLUSIONS

Prospective assessment of pregnancy intention with either a single- or a multi-item measure may allow for a more nuanced assessment of pregnancy intention than current measures. The multi-item measure may reduce the number of women categorized as ambivalent and aid the development of targeted contraceptive and preconception counseling interventions.

We present a case study using a multilevel modeling approach to determine whether depressive symptoms are affected by genetic factors. Existing studies examining this question have focused on twins. The present study built on the literature by conducting a preliminary study of the heritability of depressive symptoms within extended families. At the same time, this study assessed the need for adjustment of a heritability measure in a family study using a multigenerational sample. The sample consisted of 230 community-dwelling extended families that included 431 adult offspring, comprising full siblings, half siblings and cousins that participated in the University of Southern California Longitudinal Study of Generations. All participants filled out the Center for Epidemiologic Studies Depression (CES-D) scale. The multilevel analysis allowed us to model the natural hierarchy of the extended family. Results indicate that the proportion of the phenotypic variance for CES-D that occurs due to genetic differences is not significantly larger than zero among these participants [h(2) = 8.6%, 95% confidence interval (CI) = 0-57%, p = 0.71]. Our findings suggest that future studies examining depressive symptoms in this sample can focus on non-genetic explanatory factors without the necessity to control for genetic variation. However, our study may be limited by measurement of prevalent depressive symptoms, which may not generalize to lifetime depressive symptoms.