The use of online methods to recruit and follow a hard-to-reach population in the Peer Alternatives for Addiction Study 2021 Cohort

Abstract

Background: Although studies are increasingly adopting online protocols, few such studies in the addiction field have comprehensively described their data review procedures and successes in detecting low-quality/fraudulent data. The current study describes data collection protocols and outcomes of a large, longitudinal study (the PAL Study 2021) that implemented online design elements to study individuals seeking peer support for an alcohol use disorder. Methods: In 2021, the PAL Study collaborated with mutual-help group (MHG) partners and recovery-related organizations to recruit individuals attending a 12-step group, Women for Sobriety (WFS), LifeRing Secular Recovery, and/or SMART Recovery for an alcohol problem in-person and/or online in the prior 30 days. Participation was solicited both online and in-person. Individuals accessed baseline surveys via an open web link; follow-ups occurred at 6 and 12 months. Analyses included calculating the proportion of surveys eliminated in data quality review; comparing MHG subsamples to internal survey (benchmark) data for Alcoholics Anonymous (AA), WFS, LifeRing, and SMART; and examining response rates and attrition. Results: Although 93% of respondents who opened the baseline survey completed it, 87% of baseline surveys were eliminated in data quality review (final N = 531). Nonetheless, cleaned MHG subsamples were generally similar to benchmark samples on gender, age, race/ethnicity, and education. Follow-up rates for the cleaned sample were 88% (6 months) and 85% (12 months). Analyses revealed some differences in attrition by gender, primary MHG, and lifetime drug problems, but there was no evidence of greater attrition among those in earlier/less stable recovery. Conclusions: Study methods appear to have produced a valid, largely representative sample of the hard-to-reach target population that was successfully followed across 12 months. However, given the high survey elimination rate and need for extensive data review, we recommend that researchers avoid open-link designs and include comprehensive data review when incorporating online design elements.