Due to the high prevalence of some outcomes, odds ratios obtained from logistic regression models were converted to model-adjusted risk ratios using predicted marginals.52,53 Results are presented as adjusted risk ratios (RR) and 95% confidence intervals (CI)s. All analyses were conducted to account for the complex survey design of the NCS-A and included sample weights to account for differential probability of selection and to be nationally representative based on sociodemographic characteristics. Analyses were conducted using SAS-callable SUDAAN version 11.0 (Research Triangle Institute, Research Triangle Park, NC, 2012).
