This paper reports a multi-dimensional approach to minimize drop-outs from a two-year follow-up of a clinical trial designed to reduce initiation of tobacco use in 16,915 adolescent orthodontic patients. A hierarchical approach to data collection and tracking was employed. Seventy percent of participants were reached and interviewed at home by telephone. Strategies used to survey remaining participants included calling parents' work numbers and directory assistance, reviewing orthodontists' charts, sending surveys by mail, offering incentives, and using reverse telephone directories. More than 92% of the participants completed follow-up surveys. Multivariate analyses showed that baseline tobacco and alcohol use predicted loss to follow-up. Similarly, the number of procedures used to track each participant predicted presence of risk behaviors at post-test, demonstrating that an organized tracking hierarchy curtailed even greater compromises to internal and external validity. Evaluation and costs of individual strategies are discussed.