What does it take to retain substance-abusing adolescents in research protocols? Delineation of effort required, strategies undertaken, costs incurred, and 6-month post-treatment differences by retention difficulty.

Research retention rates vary widely due to practical difficulties that can be exacerbated when participants are minors. This article describes: (1) the range of effort required and type of follow-up strategies used to complete face-to-face follow-up interviews with substance-abusing adolescent research participants; (2) common locations of follow-up interviews; and (3) characteristics of difficult- versus easy-to-retain adolescent participants. Diverse contact strategies and numerous contact attempts were needed to obtain a 94% 1-month and 92% 6-month retention rate among substance-abusing adolescent research participants. About half of the youth did not respond to basic telephone tracking and required enhanced tracking efforts. Approximately 40% of the youth required 6 or more contacts prior to interview completion. The majority of follow-up interviews (60%) were conducted in community settings such as fast food restaurants, constituting the adolescent's preferred interview location. Telephone interviews were infrequent since adolescents wanted privacy and were concerned that a household member would listen to their answers. Those youth proving difficult-to-retain were significantly more likely to report serious problem behavior and poorer outcomes 6-months post-treatment within the alcohol/drug, juvenile justice, family, and educational domains. It was estimated that an additional $85 per participant per follow-up wave (over and above project budgets) was needed to adequately track, locate and interview an adolescent research participant. This expenditure appears reasonable to ensure a reliable/valid data set. Assessing the cost/benefit of different methods used in preventing attrition, identifying the minimum standards that avoid response bias and examining the impact of interviewer/participant alliances on data reliability/validity is discussed.