| The purpose of this dissertation is to determine which combination of risk factors from Community Care of North Carolina's Pregnancy Medical Home risk screening form is most predictive of preterm birth and to compare receipt of postpartum contraception for women who experienced a recent preterm birth with women who had a term birth. This retrospective cohort includes pregnant Medicaid beneficiaries screened by the program who delivered a live birth in North Carolina between September 2011 and September 2012. Data come from Community Care of North Carolina's Case Management Information System, Medicaid claims, and birth certificates. The first paper of the dissertation designed and internally validated a predictive model of preterm birth using variables from the Pregnancy Medical Home risk screen. Logistic regression models and bootstrapping techniques indicated that the optimal combination of risk factors for PTB prediction include non-Hispanic black race, smoking during pregnancy, underweight, multi-fetal gestation, chronic disease (diabetes, hypertension, asthma, renal disease, and other chronic conditions), cervical insufficiency, nulliparity, and previous adverse reproductive outcomes (history of preterm birth, low birth weight baby, fetal death, and second trimester loss). Salient risk factors were identified for subgroups by parity and race/ethnicity including obesity, food insecurity, unsafe or unstable housing, and a short interpregnancy interval. Evaluation of the Pregnancy Medical Home risk screen provides insight to increase its sensitivity and specificity to improve identification of women early in pregnancy at highest risk for preterm birth. The second paper compared the prevalence of contraceptive receipt and effectiveness in Medicaid claims among women with and without a recent preterm birth using logistic and multinomial logistic regression. Contraceptive receipt was 50% (25% for the most effective methods) for women who delivered preterm at 90 days postpartum compared to 52% (28% for the most effective methods) among women who delivered full term. Women with a recent preterm birth with more than two children had lower odds of receiving contraception compared to equal parity women who delivered at term (OR = 0.84, 95% CI: 0.74, 0.96). Care management should focus on the provision of effective methods of contraception to women with a recent preterm birth. |
