A bibliometric analysis of the conversion and reporting of pilot studies published in six anaesthesia journals.

Pilot and feasibility studies are preliminary investigations undertaken before a larger study. We hypothesised that only a small proportion of pilot or feasibility studies published in anaesthesia journals were correctly labelled as such. We searched for papers published between 2007 and 2017 in six anaesthesia journals using the text words 'pilot' OR 'feasibility' and included 266 original articles with 26,682 human participants. Only 34 (12.8%) were correctly labelled as a pilot or feasibility study. They were more likely to: have more median (IQR [range]) participants, 73 (40-130 [4-2716]) vs. 27 (15-60 [2-3305], p < 0.001; report feasibility outcomes, 82.4% vs. 4.3%, p < 0.001; and report an intention to convert, 100% vs. 39.7%, p < 0.001. They were less likely to test the efficacy of the primary outcome, 50% vs. 72.8%, p = 0.009; and report firm clinical conclusions 41.2% vs. 67.7%, p = 0.004. Of the studies published more than 5 years ago, correctly labelled pilot or feasibility studies were more likely to precede a published conversion study, 53.8% vs. 16%, p = 0.004. There was no difference between the number of citations 18 (9-44 [2-216]) vs. 20 (7-47 [0-251]), p = 0.865. These results have important consequences for patients, trialists, researchers and funders. We argue that correctly labelled pilot studies enhance the quality of scientific research by encouraging methodological rigour, ensuring scientific validity and reducing research waste. Authors, reviewers, editors and publishers should ensure they adhere to the contents of the 2016 CONSORT extension for pilot and feasibility studies.