Promoting Early Access to Antenatal Care to Prevent Congenital Syphilis: Evidence from a randomized experiment and behavior-based risk prediction

Syphilis is resurging in many countries, including high-income countries, making congenital syphilis a public health concern. Most cases are preventable through early access to maternal health services, highlighting the importance of reaching those likely to delay obtaining such access. We construct a risk score based on actual pregnancy notification submission behavior of individuals who had experienced childbirth and extrapolated it to participants in a randomized information experiment to examine effects of information provision according to predicted risk of delayed submission. A congenital syphilis prevention message produced the largest average effect, whereas a message explaining treatment became more effective as delay risk increased, with the treatment-message intervention yielding larger effects in the highest-risk group. These findings highlight the importance of considering effects on high-risk groups alongside average effects when designing information interventions. Using risk predicted from past behavior to assess treatment effect heterogeneity may also inform other policy settings where target groups are small and difficult to identify in advance.