Readers will be familiar with the CONSORT guidelines1 for the reporting of randomised controlled trials. These guidelines were designed to improve the reporting of RCTs and are of considerable use to researchers designing and reporting studies as well as to editors and to systematic reviewers. However, a major limitation of CONSORT is that they are limited to RCTs and there are far more research designs than RCTs. Some topics just cannot be investigated using RCTs and yet it may often be desirable to try to identify whether or not there is a causal link between two factors. It is therefore very welcome to see the development of the TREND statement (www.trend-statement.org/asp/trend.asp) — the Transparent Reporting of Evaluations with Non-Randomised designs.2 As the authors state these guidelines are developing — there is room for alteration and improvement but their prime focus is to encourage the transparency or clarity of reporting.

In view of the evolutionary state of these guidelines, I would encourage researchers, editors and reviewers to become familiar with them and to start using them. As they are used, areas, primarily of omission, will become apparent. The first area that needs expanding, in my opinion, is that more detail is required describing the characteristics of each population subgroup. I consider that more prominence of the potential for confounding factors needs to be acknowledged.

Table 1

Table 1 The TREND Checklist (Version 1.0)2
Full size table

As CONSORT has helped both investigators and reviewers improve the quality of RCTs and systematic reviews there is the possibility for TREND to do the same for non-randomised interventions. Who should then use TREND? Anyone designing a non-randomised intervention. However, if you are well into such an intervention it would be well worth your time revising the protocol against TREND to see if there are any more data that you need to collect. As CONSORT does for RCTs TREND will be particularly useful when it comes to writing up and reviewing these interventions. Editors should be encouraged to adopt TREND.

As someone who has systematically reviewed questions examined by RCTs and non-randomised interventions the process of finding the necessary information can be difficult. Although it is usual practice to contact authors it is rare to be given much additional useful information. TREND achieves a useful step forward in the reporting the of literature. Researchers and reviewers should be encouraged to use TREND which should improve the quality of work undertaken.