STARD for registration: establishing guidance on where and how to register diagnostic accuracy studies prospectively

2016 Seoul

Korevaar D¹, Cohen J¹, Askie L², Faure H³, Gatsonis C⁴, Hunter K², Kressel H⁵, McInnes M⁶, Moher D⁷, Rifai N⁸, Hooft L⁹, Bossuyt P¹

¹Department of Clinical Epidemiology, Biostatistics and Bioinformatics, Academic Medical Center, University of Amsterdam, The Netherlands

²Australian New Zealand Clinical Trials Registry (ANZCTR), NHMRC Clinical Trials Centre, University of Sydney, Australia

³ISRCTN registry, BioMed Central, London, UK

⁴Department of Biostatistics, Brown University School of Public Health, Providence, USA

⁵Professor of Radiology Harvard Medical school, Radiology Editorial Office, Boston, USA

⁶Department of Radiology, University of Ottawa, Canada

⁷Clinical Epidemiology Program, Ottawa Hospital Research Institute, Ottawa, Canada

⁸Clinical Chemistry Editorial Office, Washington, USA

⁹The Netherlands Trial Register and Cochrane The Netherlands, Julius Center for Health Sciences and Primary Care, University Medical Centre Utrecht, University Utrecht, The Netherlands

Background:

The advantages of prospective registration are multiple, and include the identification of unpublished studies. Many diagnostic accuracy studies remain unpublished, but so far these studies are rarely registered. This could be caused by the existing guidance for registering trials, which mainly focuses on comparative trials of therapeutic interventions or systematic reviews.

Objectives:

To develop guidance on where and how to register diagnostic accuracy studies, thereby facilitating and encouraging informative registration.

Methods:

Two surveys were developed based on multiple-choice questions, each with the option for further clarification in an open comment box. In survey 1, a representative of each Primary Registry in the World Health Organization’s Registry Network (n = 15) and of ClinicalTrials.gov were invited to comment on their registry’s policy for registering diagnostic accuracy studies. In survey 2, the STARD group members (STAndards for Reporting Diagnostic accuracy; n = 85) were invited to indicate whether or not 20 proposed protocol elements that specifically apply to diagnostic accuracy studies should be included in the registry record. A majority vote was defined as ≥ 2/3 agreement.

Results:

In survey 1, still open at the time of writing, 10/16 (63%) invitees replied; eight agreed that registration of diagnostic accuracy studies in existing trial registries is preferred over developing a registry specifically designed for these studies; five registries always accept registration of these studies, whereas five do so in some cases; one registry already provided guidance for registering these studies while eight would be willing to consider implementing a guidance document for registering these studies. In survey 2, 71/85 (84%) invitees responded. A majority vote was reached for 14 of the 20 proposed protocol elements but additional elements were also proposed.

Conclusions:

Many trial registries accept registration of diagnostic accuracy studies. The collected responses will help the development of a guidance document for registering such studies.