Time-to-event analyses in clinical trial publications: incomplete and not standardized

Many analyses in clinical research include so-called time-to-event (TTE) data, which indicate how much time has elapsed before an event occurs. A well-known example of such TTE analyses are survival curves estimated using the Kaplan-Meier method, which show how the proportion of surviving patients in the intervention and control group of a randomized controlled trial develops over time. It has been known for some time that TTE outcomes in studies are often reported incompletely and unsystematically. In an article published in the Journal of Clinical Epidemiology, a team of researchers – including Ralf Bender, Head of Medical Biometry at the Institute for Quality and Efficiency in Health Care (IQWiG) – has now investigated whether these shortcomings are also found in clinical trial publications included in systematic reviews.

Urgent need for reporting standards

Using 50 recent systematic reviews from the 119 most important medical journals (core clinical journals), the researchers were able to show that in the 235 clinical trials that were included in the meta-analyses for these reviews, for example, events were not clearly defined or methods and follow-up times were not well reported. However, such information is necessary to assess how reliable the results of TTE analyses are. In addition, the reporting of the TTE data varied considerably between the studies.

The researchers therefore call for authors of study publications to strictly adhere to available guidelines for TTE analyses. The way in which meta-analyses of TTE outcomes are reported in systematic reviews is being investigated in another article that has not yet been published.