OBJECTIVES: A retrospective naturalistic evaluation was undertaken to identify if pre- and post-disaster factors may predict the likelihood of those considered "at risk" of post-traumatic stress disorder (PTSD) entering a post-disaster clinical treatment program. METHODS: The intake data of 881 people referred to the program following the Queensland (Australia) natural disasters of 2010-11 was evaluated. Those referred scored >2 on the Primary Care PTSD scale. Assessment included the disaster exposure experience, demographic and clinical information, and measures of coping and resilience. Descriptive analyses and a Classification Tree Analysis (CTA) were undertaken to ascertain which factors may predict treatment participation. RESULTS: The treatment group (TG) in comparison to the non-treatment group (NTG) were more likely to perceive their life was threatened (85.1% vs 8.1%), less able to cope (67% vs 25.8%) and less resilient (4.2% vs 87.5%). The CTA using all the assessment variables found the Connor-Davidson (2-item scale) (P < 0.001), degree of property damage (P < 0.001), financial losses (P < 0.001), perception their life was threatened (P < 0.001) and insurance claims (P < 0.003) distinguished the TG from the NTG. CONCLUSIONS: The study identified factors that distinguished the TG from the NTG and predicted the likelihood of participation in a post-disaster mental health treatment.