Practical Guidance for Using Behavioral Risk Factor Surveillance System Data: Merging States and Scoring Adverse Childhood Experiences.

Introduction: The Behavioral Risk Factor Surveillance System is a national health-related survey with an optional adverse childhood experience (ACE) module. States use varying methodologies, question formats, and sampling frames, and little guidance exists for conducting multistate explorations of adverse childhood experiences. In this study, 6 adverse childhood experience scoring approaches are compared, and practical recommendations are offered for when and how each approach can be utilized most effectively.

Methods: This study used 2015 Behavioral Risk Factor Surveillance System data from the adverse childhood experience module administered by 6 states. Data were merged and analyzed between 2018 and 2021. To understand how adverse childhood experience scoring may impact estimates of association, concordance/discordance among 6 approaches (continuous versus categorical, states that collected all adverse childhood experiences versus those that collected any adverse childhood experiences, and normalized versus standard scores) was evaluated. Using separate weighted multivariable logistic regression models controlling for confounders, the relationship between adverse childhood experiences using each approach and the presence of 10 chronic health conditions was also assessed.

Results: Comparisons revealed discordance for categorical versus continuous approaches (30%) and all-ACEs versus any-ACEs (20%) but full concordance for standard versus normalized approaches. Discordance occurred more frequently with low-prevalence outcomes (≤7.0%) and lower-exposure samples (any-ACEs).

Conclusions: Results revealed general concordance across adverse childhood experience scoring approaches when outcomes commonly occurred and when the sample was limited to just states that asked the full array of adverse childhood experiences. However, on a deeper exploration of discordant findings, specific nuances were uncovered that may help guide researchers when deciding on which approach to use on the basis of the research question and conceptual model driving study objectives.