Reporting participation rates in studies of non-pharmacological interventions for patients with chronic obstructive pulmonary disease: a systematic review

UNLABELLED BACKGROUND Pulmonary rehabilitation (PR) and self-management (SM) support programmes are effective in the management of patients with chronic obstructive pulmonary disease (COPD), but these interventions are not widely implemented in routine care. One reason may be poor patient participation and retention. We conducted a systematic review to determine a true estimate of participation and dropout rates in research studies of these interventions. METHODS Studies were identified from eight electronic databases including MEDLINE, UK Clinical Trial Register, Cochrane library, and reference lists of identified studies. Controlled clinical trial studies of structured SM, PR and health education (HE) programmes for COPD were included. Data extraction included 'participant flow' data using the Consolidated Standards of Reporting Trials (CONSORT) statement and its extension to pragmatic trials. Patient 'participation rates' (study participation rate (SPR), study dropout rate (SDR) and intervention dropout rate (IDR)) were calculated using prior participation definitions consistent with CONSORT. Random effects logistic regression analysis was conducted to examine effects of four key study characteristics (group vs. individual treatment, year of publication, study quality and exercise vs. non-exercise) on participation rates. RESULTS Fifty-six quantitative studies (51 randomised controlled trials, three quasi-experimental and two before-after studies) evaluated PR (n = 31), SM (n = 21) and HE (n = 4). Reports of participant flow were generally incomplete; 'numbers of potential participants identified' were only available for 16%, and 'numbers assessed for eligibility' for only 39% of studies. Although 'numbers eligible' were better reported (77%), we were unable to calculate SPR for 23% of studies. Overall we found 'participation rates' for studies (n = 43) were higher than previous reports; only 19% of studies had less than 50% SPR and just over one-third (34%) had a SPR of 100%; SDR and IDR were less than or equal to 30% for around 93% of studies. There was no evidence of effects of study characteristics on participation rates. CONCLUSION Unlike previous reports, we found high participation and low dropout rates in studies of PR or SM support for COPD. Previous studies adopted different participation definitions; some reported proportions without stating definitions clearly, obscuring whether proportions referred to the study or the intervention. Clear, uniform definitions of patient participation in studies are needed to better inform the wider implementation of effective interventions.