Abstract
Background: High attrition rates in community intervention trials limit study generalizability, threaten internal validity and decrease statistical power. Attrition can introduce selection bias in randomized controlled trials (RCT); however, cluster randomization minimizes this effect because cohorts formed at baseline are balanced on known covariates. Nonetheless, attrition is a challenge for research with traditionally difficult-to-follow populations, including African Americans. The purpose of this paper is to determine attrition rates for the Educational Program to Increase Colorectal Cancer Screening (EPICS) cluster RCT, describe challenges encountered during a study pilot period and provide solutions to overcome threats to full-scale trial implementation.
Methods: We intend to enroll 7,200 individuals in this study, with 1,800 in four dissemination arms (in-person with technical assistance, in-person without technical assistance, web-access with technical assistance and web-access without technical assistance). Based on data from a previously published study, we calculated an intracluster correlation coefficient to equal 0.0911 (9.1%). The effective sample size was 5,645, which still has high power (>99%) 320 with design effect of 1.2754. Quantitative and qualitative measures were used to obtain data from community coalitions targeting African Americans ages 50-74 years with no personal or family history and not current on colorectal cancer (CRC) screening for EPICS implementation. Community coalition participation, facilitator-training outcomes and participant enrollment were analyzed to determine attrition rates. Key informant interviews were conducted to identify challenges to trial implementation.
Results: Thirty-three community coalitions were approached for participation; 13 did not meet inclusion criteria, eight were not available when the trial was offered and two declined participation. At baseline (Time 1), 20 community coalitions were randomized to passive or active study arms. At Time 2 (facilitator training), one community coalition was lost from each study arm (attrition rate=10%). Two hundred and fifty individuals were approached to complete facilitator training; 204 were trained and certified as EPICS facilitators (attrition rate=18%). At pilot testing (Time 3) 665 participants were enrolled; 436 completed all three EPICS sessions at Time 4 (attrition rate=34%).
Conclusions: We previously reported a participant attrition rate for the Colorectal Cancer Screening Intervention Trial (CCSIT). We attribute the 9% difference in participant attrition between CCSIT (25%) and EPICS cRCT (34%) to research conducted in an academic versus real world settings. Difficulty recruiting and engaging participants may be overcome by offering additional locations and more convenient days/times for EPICS sessions. For community coalitions, results suggest that enhancing capacity to conduct research would limit attrition in trials such as EPICS. Interviewing potential facilitators, describing the cRCT and describing roles and responsibilities in greater detail may result in greater training participation.
Citation Format: Selina A. Smith, Ernestine Delmoor, Joyce Q. Sheats, Mechelle D. Claridy, Francesca Damus, Ernest Alemah-Mensah, Daniel S. Blumenthal. Attrition in a cluster randomized controlled trial: Lessons learned from the Educational Program to Increase Colorectal Cancer Screening (EPICS) pilot study. [abstract]. In: Proceedings of the Twelfth Annual AACR International Conference on Frontiers in Cancer Prevention Research; 2013 Oct 27-30; National Harbor, MD. Philadelphia (PA): AACR; Can Prev Res 2013;6(11 Suppl): Abstract nr C10.
RSS Feeds