We sought to identify whether patterns exist in the Burn Injury Rehabilitation Model Systems' database among participants lost to follow-up at 6, 12, or 24 months after injury and to define characteristics that reliably discriminate between persons who are lost to follow-up and those who are not. All participants met the American Burn Association criteria for major burn injury, were 18 years of age or older, received care from one of four burn model systems, and consented to participate in a 2-year prospective data-collection process. Step-wise logistic regression was used to develop three prediction models for the probability of loss to follow-up. The percent of individuals successfully contacted for follow-up were 64% at 6 months, 54% at 12 months, and 42% at 24 months after injury. Individuals who were younger, not employed at time of burn, with less than a high school level education, a history of drug abuse, circumstances of injury involving suspected assault, and having no insurance for care were lost to follow-up. Longer stay in the hospital, on the other hand, increased the likelihood of follow-up. The same risk factors remained significant with or without adjusting for site indicating that these factors are independent and significant in spite of any potential site differences. Successful follow-up at 6- and 12-month intervals increased the likelihood of achieving a follow-up at 24 months after injury. The sociodemographic risk factors for attrition identified in this study represent significant enduring vulnerabilities. The findings necessitate a close examination of several factors and the use of strategies to reduce the risk of attrition.