Evidence for strategies that improve recruitment and retention of adults aged 65 years and over in randomised trials and observational studies: a systematic review

Rosie J Lacey,Ross Wilkie,Gwenllian Wynne-Jones,Joanne L Jordan,Emily Wersocki,John McBeth

DOI: https://doi.org/10.1093/ageing/afx057

2017-11-01

Abstract:Background: adults aged ≥65 years are often excluded from health research studies. Lack of representation reduces generalisability of treatments for this age group. Objective: to evaluate the effectiveness of strategies that improve recruitment and retention of adults aged ≥65 in observational studies and randomised controlled trials (RCTs). Methods: searches conducted in 10 databases for RCTs of recruitment and retention strategies in RCTs or observational studies. Two reviewers screened abstracts and full-text articles for eligibility and extracted data. Studies without separate data for adults aged ≥65 were discarded. Risk of bias assessed using the Cochrane Risk of Bias tool. Results were synthesised narratively. Results: thirty-two studies were included in the review (n = 75,444). Twelve studies had low risk of bias, of which 10 had successful strategies including: Opt-out versus opt-in increased recruitment (13.6% (n = 261)-18.7% (n = 36) difference; two studies); Advance notification increased retention (1.6% difference, OR 1.45; 95% CI 1.01, 2.10, one study (n = 2,686); 9.1% difference at 4 months, 1.44; 1.08, 1.92, one study (n = 753)); Hand-delivered versus postal surveys increased response (25.1% difference; X2 = 11.40, P < 0.01; one study (n = 139)); Open randomised design versus blinded RCT increased recruitment (1.56; 1.05, 2.33) and retention (13.9% difference; 3.1%, 24.6%) in one study (n = 538). Risk of bias was high/unclear for studies in which incentives or shorter length questionnaires increased response. Discussion: in low risk of bias studies, few of the strategies that improved participation in older adults had been tested in ≥1 study. Opt-out and advance notification strategies improved recruitment and retention, respectively, although an opt-out approach may have ethical limitations. Evidence from single studies limits the generalisability of other strategies.