Preliminary analysis of associations of prior TEs with disaster-related PTSD showed prior TEs involving exposure to sectarian, interpersonal, or sexual violence were the only ones consistently associated with increased risk of disaster-related PTSD controlling model 3 predictors. (See supplementary material.) The most parsimonious characterization of these associations used a single dichotomous variable for whether the respondent was previously exposed to any such TE (reported by 25.1% of respondents; OR 16.4, 95% CI 2.6–101.6). Preliminary analysis of the associations of CAs with disaster-related PTSD showed numerous significant positive univariate associations that could best be summarized with a 0–3+ count for number of MFF CAs (15.4%, 1; 6.5%, 2; 13.3% 3+; OR 2.9, 95% CI 1.4–6.2). (See supplementary material.) The multivariate ORs of both TEs and CAs were larger than the univariate ORs and had wide CIs. In addition, the OR of serious injury/death of someone close became markedly higher in model 3 than model 2.
