Who is in your trial? Improving the reporting of participant characteristics in trial protocols and results.

<p dir="ltr">‘Without data, you’re just another person with an opinion.’</p><p dir="ltr"><br></p><p dir="ltr">So said W. Edwards Deming, whose work on industrial quality-control methods helped post-war Japan’s economic recovery and that of the American car industry in the 1980s. It is doubtful that any seasoned trialist would argue with Dr. Deming. After all, generating data is trialists’ bread and butter.</p><p dir="ltr"><br></p><p dir="ltr">We do trials because we believe that our data will help patients, healthcare professionals, guideline developers, policymakers and others to make informed, evidence-based healthcare decisions. Trial data, ideally from more than one trial, reduces clinical uncertainty and improves confidence in a decision.</p><p dir="ltr"><br></p><p dir="ltr">But what if these data say little or nothing about who was in the trial beyond the fact that everyone met the inclusion criteria? Imagine a trial in the UK that aimed to reduce maternal mortality. Future users of the trial results will likely need to know (at least) the age, socioeconomic status and ethnicity of the people in the trial because maternal mortality is higher in older, socioeconomically disadvantaged and, in particular, Black women [1]. Older or socioeconomically disadvantaged or Black women have a great deal to gain from improved care, with those sharing all three characteristics standing to benefit the most.</p><p dir="ltr"><br></p><p dir="ltr">If this trial showed a benefit for those receiving the intervention, a UK policymaker could be expected to ask whether the intervention also worked for Black women and/or women who are older or experiencing socioeconomic disadvantage. Does it work for them too? An intervention that doesn’t help all women will widen, not reduce, inequity. Without explicit data on who is in the trial, users of the results are left to speculate. We are back to opinions.</p><p dir="ltr"><br></p><p dir="ltr">It is not difficult to find examples of trials that do not say much about who is in the trial beyond stating that participants met the clinical eligibility criteria [2,3,4,5,6]. Buttery and colleagues for example reported that of the 24 trials in their sample, only 12 reported ethnicity and only four reported a measure of socioeconomic status [4]. Of those not reporting ethnicity, none reported the lack of these data as a limitation and only one of the 21 trials not reporting socioeconomic status did the same. Fortunately, the editors of some journals have taken positive steps to help reduce the likelihood of this happening in the future [7, 8]. Today, we announce some changes Trials is making to do the same.</p>