Development and validation of a community risk score for sexual and reproductive health in the United States

Abstract

Equitable access to sexual and reproductive health (SRH) care is key to reducing inequities in SRH outcomes. Publicly funded family-planning services are an important source of SRH care for people with social risk factors that impede their access. This study aimed to create a new index (Local Social Inequity in SRH [LSI-SRH]) to measure community-level risk of adverse SRH outcomes based on social determinants of health (SDoH). We evaluated the validity of the LSI-SRH scores in predicting adverse SRH outcomes and the need for publicly funded services. The data were drawn from more than 200 publicly available SDoH and SRH measures, including availability and potential need for publicly supported family planning from the Guttmacher Institute. The sample included 72 999 Census tracts (99.9%) in the 50 states and the District of Columbia. We used random forest regression to predict the LSI-SRH scores; 42 indicators were retained in the final model. The LSI-SRH model explained 81% of variance in the composite SRH outcome, outperforming 3 general SDoH indices. LSI-SRH scores could be a useful for measuring community-level SRH risk and guiding site placement and resource allocation.

Keywords: family planning; machine learning; reproductive health; sexual health; social determinants of health; spatial analysis.

Figure 1.

Mean LSI-SRH score by state. Source: Authors’ analysis of an author-developed composite risk score predicting a composite outcome consisting of 5 measures of SRH based on 42 indicators of social determinants of health (see Appendix Table A1). Shown are the mean LSI-SRH scores for each state, indicating relative risk of adverse SRH outcomes based on 42 indicators of social determinants of health. Darker shade indicates higher risk, on average. Abbreviations: LSI-SRH, Local Social Inequity in Sexual and Reproductive Health; SRH, sexual and reproductive health.

Figure 2.

Variance in composite SRH outcome explained by the LSI-SRH score vs other SDoH indices (n = 72 999 Census tracts). Source: Authors’ analysis of 3 general SDoH indices (Social Vulnerability Index, Area Deprivation Index, and Social Deprivation Index) and the LSI-SRH score, an author-developed composite risk score predicting a composite outcome consisting of 5 measures of SRH based on 42 drivers of health (see Appendix Table A1). Variance explained is a measure of model fit based on R² statistics. A higher variance explained is better. Abbreviations: LSI-SRH, Local Social Inequity in Sexual and Reproductive Health; SDoH, social determinants of health; SRH, sexual and reproductive health.