OBJECTIVES: Prenatal care is an established mechanism for identifying and managing risk factors impacting pregnancy outcomes. Despite aggressive efforts in the United States (US) to assure that all women begin care in the first trimester, every year about 70,000 women in the US receive no care prior to delivery. We hypothesized that US women receiving no prenatal care comprise clusters (subgroups) with distinctive behavioral, socio-demographic, and medical risks and that birth outcomes differ among the clusters. METHODS: White, Black, and Hispanic women (n = 126,220) receiving no prenatal care and delivering a live, singleton infant were identified from linked birth and death certificates for years 1995 through 1997. Cluster analysis was used to group women with similar characteristics, and cluster assignment was evaluated using discriminant analysis. Birth outcomes for any care and no-care women were then examined using logistic regression. RESULTS: Six replicable clusters of women with no care were identified. Birth outcomes varied significantly among clusters and were two to four times worse for no-care clusters compared to outcomes for women receiving any care. CONCLUSIONS: Cluster analysis is an effective alternative for grouping individuals for use in public health education, intervention, and outreach programming. Women receiving no prenatal care were characteristically different from women receiving any care in this study, but they did not represent a homogenous group. Findings suggest that interventions should target reducing the proportion of women receiving no care and should be tailored to specific no-care clusters.
OBJECTIVES: Prenatal care is an established mechanism for identifying and managing risk factors impacting pregnancy outcomes. Despite aggressive efforts in the United States (US) to assure that all women begin care in the first trimester, every year about 70,000 women in the US receive no care prior to delivery. We hypothesized that US women receiving no prenatal care comprise clusters (subgroups) with distinctive behavioral, socio-demographic, and medical risks and that birth outcomes differ among the clusters. METHODS: White, Black, and Hispanic women (n = 126,220) receiving no prenatal care and delivering a live, singleton infant were identified from linked birth and death certificates for years 1995 through 1997. Cluster analysis was used to group women with similar characteristics, and cluster assignment was evaluated using discriminant analysis. Birth outcomes for any care and no-care women were then examined using logistic regression. RESULTS: Six replicable clusters of women with no care were identified. Birth outcomes varied significantly among clusters and were two to four times worse for no-care clusters compared to outcomes for women receiving any care. CONCLUSIONS: Cluster analysis is an effective alternative for grouping individuals for use in public health education, intervention, and outreach programming. Women receiving no prenatal care were characteristically different from women receiving any care in this study, but they did not represent a homogenous group. Findings suggest that interventions should target reducing the proportion of women receiving no care and should be tailored to specific no-care clusters.
Authors: Bin Zhang; Rong Yang; Sheng-Wen Liang; Jing Wang; Jen Jen Chang; Ke Hu; Guang-Hui Dong; Rong-Hua Hu; Louise H Flick; Yi-Ming Zhang; Dan Zhang; Qing-Jie Li; Tong-Zhang Zheng; Shun-Qing Xu; Shao-Ping Yang; Zheng-Min Qian Journal: J Huazhong Univ Sci Technolog Med Sci Date: 2017-08-08