Nonuse and dropout attrition for a web-based mental health intervention delivered in a post-disaster context.

Web-based mental health interventions are an excellent means to provide low cost, easily accessible care to disaster-affected populations shortly after exposure to an event. However, the extent that individuals will access and use such interventions is largely unknown. We examined predictors of nonuse and dropout attrition for a web-based mental health intervention in 1,249 randomly selected adults in two Texas counties--Galveston and Chambers--that were hardest hit by Hurricane Ike in 2008. Participants completed a structured telephone interview to assess demographics, impact of disaster exposure, history of traumatic events, mental health symptoms, and service utilization. Following the interview, participants were oriented and invited to access a web-based intervention and then contacted four months later to evaluate their use of the website and mental health functioning. Separate logistic and Poisson regressions were used to determine baseline predictors of nonuse attrition, predictors of dropout attrition, and predictors of completing intervention modules. Results suggested that the strongest buffer against nonuse attrition and dropout attrition was having considered seeking formal mental health treatment. Results of this study inform the development and dissemination of web-based interventions in future disaster affected areas.