Second, we used survey-weighted logistic regression models (Stata, version 17 logit; StataCorp) of the same BRFSS ACEs small area estimates data by US state to estimate adjusted odds ratios (AORs) for adults’ ACE count and selected self-reported current health outcomes (arthritis, asthma, cancer [excluding skin cancer], chronic obstructive pulmonary disease, depression, diabetes, heart disease, kidney disease, and stroke) and risk factors for ill health (heavy drinking, overweight or obesity, and smoking) (eTable 1 in Supplement 1 reports health condition definitions in source data).6,8,9,15 Models controlled for respondent characteristics as self-reported in BFRSS (sex, age, race and ethnicity, educational level, marital status, current employment status, and metropolitan status) and models of health outcomes (eg, cancer) controlled for the analyzed risk factors (eg, smoking). Models controlled for race and ethnicity because previous research has reported different prevalence of both ACEs and chronic health conditions among adults by race and ethnicity.
