Time for registration of observational research?
Much progress has been made over the past few years in ensuring that clinical trials are better designed, reported and published. Initiatives such as trial registration, requirements for protocols to be made available for reviewers, editors, and even readers of trials, and publishing initiatives to ensure that all trials make their way into the published literature – even if they don’t have the results that the trials sponsors hoped for – are hopefully the start of a trend that will provide more confidence in the results of trials.
What about studies other than clinical trials? After all, most clinical research is not comprised of clinical trials. Several initiatives have recently begun to address some of the same issues of study design and registration in other types of clinical research. PLoS Medicine argued back in September 2009 that it was essential that the growing enthusiasm for comparative effectiveness research did not result in ‘cherry picking’ of research to serve hidden agendas and that registration of all research was one way to stop this happening. More recently, in January 2010 the Lancet announced its intention to have protocols for observational studies registered, and this week the BMJ also indicated that it would take a more rigorous approach to scrutinising observational studies and is “actively supporting the registration of observational study protocols and results in publicly accessible registries” In the same issue of the BMJ Harry Hemingway and colleagues discuss a particular type of research – on prognosis. Their first figure starkly illustrates the “mile wide, inch deep” research literature typical of many prognostic markers – shown by looking at circulating biomarkers in neuroblastoma “in which the median number of publications per biomarker was 1”. Noting that “Stemming the tide of low quality, low impact, prognosis research is an urgent priority for the medical and research community.”, they go onto describe 10 areas where specific actions could make prognosis research more likely to generate reliable new knowledge and more likely to generate redundant or misleading research. Solutions include the development of protocols before research is done, development of better methodology, and reduction of publication bias — including registration of studies.
The news that journals and researchers are now turning the spotlight on the deficiencies of clinical research other than trials can only be very welcome.