Post-it notes to improve questionnaire response rates in RCTs - findings from a randomised sub-study

Abstract

Background: Valid treatment effect estimates in the analysis of RCTs using patient reported outcomes depend on adequate response rates. Evidence from non‐RCT surveys suggests that inexpensive post‐it notes may improve response rates. Aim: To assess the effectiveness of a post‐it note attached to questionnaires on patient response rate and time to response in a mental health study population. Methods: A two‐arm RCT was embedded into the follow‐up of older adults as part of the UK Casper Plus trial and Casper Cohort. At 4 months followup, participants were randomised to receive either a questionnaire or a questionnaire + post‐it note requesting completion. Logistic regression and time‐to‐event analyses were used to assess attrition. Results were combined with those of a previous embedded RCT (ATLAS) in a meta‐analysis. Results: 266 of 297 (89.6%) participants returned their questionnaire in the post‐it note arm, compared with 282 of 314 (89.8%) in the control arm (OR 0.97, 95% CI 0.58 to 1.64, p=.92). The pooled OR of response rates from the meta‐analysis was 0.94 (95% CI 0.65 to 1.37, p=.76). An exploratory analysis showed that participants with major depression were more likely, and participants with sub‐threshold depression less likely to respond to post‐it notes (p of interaction =.02). There was no evidence for group differences for the time taken to respond (p=.54) or whether a reminder was required (p=.80). Conclusions: Post‐it notes were not found to be effective in improving retention in this trial. However, beneficial effects for sub‐groups of clinical populations may exist.