Determinants of research follow-up participation in an alcohol treatment outcome trial

This study examined factors associated with research attrition in a long-term follow-up study (48 months). Researchers attempted to contact all randomized participants, not just those who completed treatment. The processes by which baseline characteristics, early treatment-research experiences, and short-term outcome affected subsequent participation were examined using logistic regression. The analyses deal primarily with refusal, the main reason for attrition. Baseline characteristics had small effects on likelihood of refusal; research engagement had some impact; but treatment participation had a strong effect. Short-term outcome did not predict refusal. These findings, if generalizable, have implications for the conduct and reporting of outcome studies. By directly studying bias, rather than presuming its absence on skimpy evidence, researchers can achieve a better understanding of the strengths and limitations of outcome results.