Comparison models were run without county fixed effects to reduce the computational complexity of the models during estimation. Sensitivity analyses compared adjusted breastfeeding initiation prevalence estimated from linear regression, logistic regression, and propensity score logistic regression models using inverse probability weighting to compare the robustness of our findings to potential imbalances in covariates or choice of estimation approach between WIC participants and nonparticipants. A 2-sided P < 0.05 was considered statistically significant. Differences in WIC status per year were compared using the β coefficient and t test from the LPM model fit for each year. An interaction term was included to assess differences in linear trends per year by WIC status. Changes over time in predicted probabilities of breastfeeding initiation were examined. Models were adjusted for maternal sociodemographic and maternal and infant health characteristics, which included maternal age, marital status, educational level, birth order, the timing of prenatal care initiation, prepregnancy BMI, smoking, route of delivery, birthweight, NICU admission, and infant sex.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |