We identified the years from 2018 to 2022 in which states had a child abuse policy and a mandated reporting policy for prenatal substance use using information from the Child Welfare Information Gateway, official state policy text from the Guttmacher Institute and available online from state agencies, and published studies that have reconciled differences across sources, including through direct contact with CPS agencies.^18-20 We categorized each state and year as having a child abuse policy only, a mandated reporting policy only, both policies, or neither policy (Appendix Table A, available as a supplement to the online version of this article at http://www.ajph.org).

For our primary analyses, we considered states to have a child abuse policy if prenatal substance use was criminalized or defined as child abuse or neglect under civil child welfare statutes, considered to be grounds for termination of parental rights, used as evidence to substantiate reports of child abuse or neglect, or resulted in placement on a CPS registry.¹⁹ We considered states to have a mandated reporting policy if health care providers were required to report suspected or confirmed prenatal substance use to CPS agencies.¹⁰

Congenital syphilis is a reportable condition in all 50 states and the District of Columbia and is a nationally notifiable disease. All US jurisdictions voluntarily provide case notifications to the Centers for Disease Control and Prevention (CDC) through the National Notifiable Diseases Surveillance System using the Council of State and Territorial Epidemiologists case definition. We used 2018 to 2022 National Notifiable Diseases Surveillance System data on the annual number of congenital syphilis cases for all 50 states and the District of Columbia.

We created a conceptual diagram of factors likely operative in associations between these state policies and congenital syphilis case rates using existing empirical evidence and our subject matter expertise (Appendix Figure A, available as a supplement to the online version of this article at http://www.ajph.org). On the basis of our conceptual diagram, we adjusted analyses for timevarying indicators of the state health policy context (state Medicaid expansion, income thresholds for Medicaid eligibility for pregnant and postpartum women, percentage of adult women 19 to 64 years of age without health insurance, number of obstetricians/gynecologists per 100 000 live births, and policies specifying priority access to substance use disorder treatment for pregnant women) and state demographics (percentage of the population unemployed, percentage of the population in poverty, and race and ethnicity).

We examined congenital syphilis case rates per 100 000 live births by year and state policy category. We calculated rates using live birth data from the National Center for Health Statistics. We also examined the distribution (mean, median, 25th percentile, and 75th percentile) of indicators of the state health policy context and state demographics over the entire study period by state policy category.

To compare congenital syphilis case rates in states with a child abuse policy only, a mandated reporting policy only, and both policies to rates in states with neither policy, we conducted linear regression with a generalized estimating equation approach and an exchangeable correlation matrix to account for repeated measures within states over time. Generalized estimating equations are a set of equations applied to regression models to estimate model coefficients and their standard errors when the outcomes are not independent, for example when there is clustering of data (e.g., repeated measures within states over time).²¹ They are used to obtain population average estimates while accounting for the fact that outcomes within clusters (e.g., within states) are likely to be more similar than those between clusters (e.g., between states).²¹ We adjusted regression models for the confounders noted previously and a linear time trend to account for secular trends over time.

We did not use a quasi-experimental study design, such as a difference-indifferences approach, as there was no prepolicy time for most states. Fortyeight of the states and the District of Columbia first implemented child abuse or mandated reporting policies before 2018 (Appendix Table A). Prepolicy time is necessary for conducting a differencein-differences analysis as this approach leverages policy changes both over time and between states to generate estimates.²²

The CDC uses National Notifiable Diseases Surveillance System case notification data and a stepwise analytic framework²³ to classify congenital syphilis cases by type of missed prevention opportunity. This framework classifies missed prevention opportunities as follows: (1) among all congenital syphilis cases, the number that had a nondocumented or nontimely test or late identification of seroconversion in pregnancy; (2) among cases with a timely test, the number with inadequate treatment or no or nondocumented treatment; and (3) among cases in which there was adequate treatment, the number with clinical evidence of congenital syphilis despite adequate treatment or with an unclear indication for inclusion as a reportable congenital syphilis case. We used information on missed prevention opportunities to create Sankey diagrams visualizing the number of missed prevention opportunities by state policy category and to calculate the unadjusted percentage of cases by type of missed prevention opportunity.

We conducted multiple sensitivity analyses. First, given that some state child abuse policies have less severe consequences (e.g., placement on a CPS registry), we conducted analyses in which we classified only states with policies that criminalized prenatal substance use, defined it as child abuse or neglect, or considered it grounds for termination of parental rights as having a child abuse policy (Appendix Table A). Second, given that some state mandated reporting policies are less stringent and require reporting of prenatal substance use to CPS only if there is also suspicion of child abuse or neglect, we conducted analyses in which we classified states as having a mandated reporting policy if prenatal substance use alone was sufficient to require reporting (Appendix Table A).

Third, given the onset of the COVID-19 pandemic and acute disruptions to health care access in 2020, we conducted analyses excluding 2020. Fourth, we adjusted for the rate of primary and secondary syphilis infections among women 15 to 44 years of age per 100 000 population as a proxy for the incidence of syphilis in each state (Appendix Figure A).

Fifth, we adjusted for whether states required syphilis testing during the third trimester of pregnancy, overall or only for populations at increased risk, as of 2018.²⁴ Although this requirement is unlikely to affect whether states have a child abuse policy or mandated reporting policy for prenatal substance use, it is likely to affect receipt of syphilis testing and treatment during pregnancy. Finally, to assess whether our results were disproportionately influenced by states with high congenital syphilis case rates, we conducted analyses excluding California, a state with a child abuse policy only, and Florida and Texas, states with a mandated reporting policy only.

This study has limitations. First, this was an ecological study conducted at the state level. The results do not allow for casual inferences regarding the potential impact of living in a state with a child abuse policy or a mandated reporting policy for prenatal substance use on congenital syphilis as a pregnancy outcome at the individual level. Second, although we adjusted for a comprehensive set of potential confounders, there may be unmeasured confounding. Specifically, because we were not able to use a quasi-experimental study design such as a difference-in-differences approach, there is the potential for unmeasured confounding resulting from differences in state contexts.

Third, there is some variation across states in the specifics of their child abuse policies and mandated reporting policies. However, our results were robust to sensitivity analyses incorporating narrower definitions of these policies. Fourth, how states implement child abuse policies and mandated reporting policies likely varies. Future research can further our understanding of differences in policy implementation across states and the potential implications of these differences for behaviors and receipt of health care services among pregnant women.

Finally, pregnant women may not necessarily be aware of specific state policies. However, pregnant women using substances report that they avoid health care interactions because they are afraid of detection and punishment, including CPS involvement and loss of infant custody.^12-16 They often note that these fears are based on their own experiences during previous pregnancies and knowledge of their family’s and friends’ experiences,^12-16 suggesting an overall awareness of the context and consequences created by child abuse policies and mandated reporting policies for prenatal substance use. Existing evidence aligns with these fears, showing higher rates of CPS reports¹⁸ and foster care entries³¹ among infants in states with these policies than in states without them.

Congenital syphilis is a serious condition that is preventable though timely syphilis testing and adequate treatment of infection during pregnancy. Rapidly increasing congenital syphilis case rates indicate an urgent need to reduce barriers to prenatal care, a critical opportunity for testing and treatment. Leading medical and public health organizations have issued statements recommending against child abuse policies and mandated reporting policies for prenatal substance use, noting that these policies can create barriers to care and subsequent harms to maternal and infant health.^32-34 Our results align with these statements, demonstrating elevated congenital syphilis case rates in states with both a child abuse policy and a mandated reporting policy for prenatal substance use.