Stratifying risk of retinopathy in diabetes and other stories . . .

A systematic review by the OPEN Project (To Overcome failure to Publish nEgative fiNdings)—an initiative funded by the European Commission—looked at 39 studies covering more than 20 000 clinical trials (PLoS One 2014;9:e0114023, doi:10.1371/journal.pone.0114023). The authors estimate that only 53% of registered clinical trials have been published. They also found that trials with significant or positive results were about three times more likely to be published than those with non-significant or negative results. How much longer is clinical decision making going to be based on a biased sample of half the evidence?

Every person with diabetes in the UK is offered retinal screening at least once a year. When 354 549 patients were observed for up to four years, 16 196 (4.6%) of them progressed to referable retinopathy (Diabetes Care 2014, doi:10.2337/dc14-1778). After examining factors such as disease duration and the initial pattern of retinal changes, the investigators conclude that “It may be possible to stratify …