Increasing value in diagnostic research: Publication and reporting of test accuracy studies

Diagnostic tests form an essential part of daily clinical practice. Unfortunately, most tests are not perfectly accurate; they usually produce a proportion of false positive and false negative results. Diagnostic accuracy studies evaluate a test’s ability to find out whether or not individuals have a specific target condition. After researchers complete their diagnostic accuracy study, they should publish a corresponding study report that not only contains the study findings, but also provides sufficient details about the study’s rationale, objectives, design, methods, and analyses. The studies presented in this thesis aimed to uncover problems and deficiencies in the process of publishing and reporting diagnostic accuracy studies, with the ultimate goal of increasing value in diagnostic research. We found that almost half of all initiated diagnostic accuracy studies never reach full-text publication in a peer-reviewed journal, and that studies that report more promising results on the diagnostic performance of a test seem to be published more rapidly than those with less promising results. Such practices are undesirable; not only does this represent a waste of research, but it may also lead to unjustified optimism about test performance. Among studies that are published, we found that the quality of reporting has improved gradually over the past years, but that reporting remains insufficiently complete for many studies. To assist authors in reporting informative study reports, we developed the Standards for Reporting of Diagnostic Accuracy Studies (STARD) 2015.