Introduction

In 2003, the Standards for Reporting of Diagnostic Accuracy Studies (STARD) statement was published in 13 biomedical journals.1 ,2 Diagnostic accuracy studies provide estimates of a test's ability to discriminate between patients with and without a predefined condition, by comparing the test results against a clinical reference standard. The STARD initiative was developed in response to accumulating evidence of poor methodological quality and poor reporting among test accuracy studies in the prior years.3 ,4 The STARD checklist contains 25 items which invite authors and reviewers to verify that critical information about the study is included in the study report. In addition, a flow chart that specifies the number of included and excluded patients and characterises the flow of participants through the study is strongly recommended. Since its launch, the STARD checklist has been adopted by over 200 biomedical journals (http://www.stard-statement.org/).

Over the past 20 years, reporting guidelines have been developed and evaluated in many different fields of research. Although a modest increase in reporting quality is sometimes noticed in the years following the introduction of such guidelines,5 ,6 improvements in adherence tend to be slow.7 This makes it difficult to make statements about the impact of such guidelines. For STARD, there has been some controversy around its effect.8 While one study noticed a small increase in reporting quality of diagnostic accuracy studies shortly after the introduction of STARD,9 another study could not confirm this.10

Systematic reviews can provide more precise and more generalisable estimates of effect. A recently published systematic review evaluated adherence to several reporting guidelines in different fields of research, but STARD was not among the evaluated guidelines.11 To fill this gap, we systematically reviewed all the studies that aimed to investigate diagnostic accuracy studies’ adherence to …