ABSTRACT

When diagnostic testing studies are examined, it is seen that due importance is not attached to planning and conduct-ing studies that evaluate the discriminating power of tests and reporting the results of such studies. Recent increase in the number of meta-analysis studies has indicated the need to create standards for the reporting of results. Based on this need, two statements making up such standards were developed and published in 2003 under the names of QUADAS and STARD. STARD criteria were adapted to Turkish by Genç et al. in 2012. The aim of our study is to identify the quality of the diagnostic accuracy studies published in Turkish Medical Directory by using the Turkish version of the STARD statement and to deter-mine the rate of satisfaction of each criterion by year.

To find appropriate studies, the articles that were published on ULAKBİM Medical Database [Turkish Medical Directory] between 2007 and 2010 and included the words, “sensitivity, specificity, discrimination receiver operating characteristic, ROC, diagnostic accuracy” were searched, and 459 articles were determined. As a result of the electronic search, 459 articles were determined. Then those studies were re-evaluated based on the specified criteria. Finally, 75 appropriate articles were chosen. Those studies were evaluated according to the Turkish STARD statement, and percentage of satisfaction was found for each criterion.

The most compliant articles with the STARD statement satisfied 21 items out of 25 whereas the least compliant articles satisfied only 6 items. Moreover, 9% of 75 articles have less than 50% relevance to the STARD statement while 81% of 75 articles are in 50-75% relevance range. A great majority of the studies (91%) provide measures related to diagnostic accura-cy, but only 32% give confidence intervals regarding them. In the evaluated articles, the criteria concerning the subjects included in the studies were mostly satisfied at a high rate (84% in average). These rates of satisfaction are similar over the years. In 96% of the evaluated studies, the clinical applicability of the results was discussed. This is the criterion satisfied most.

To conclude, the rate of satisfaction of the STARD criteria by the diagnostic accuracy studies published on the ULAKBİM Medical Database [Turkish Medical Directory] between 2007 and 2010 is adequate, though not very high. The intro-duction of the Turkish version of the STARD statement may serve as a very important tool for the standardization of the diag-nostic accuracy studies published in Turkish. This may lead to an improvement in the quality of the diagnostic accuracy stud-ies in the Turkish literature.

Keywords: STARD Statements, Diagnostic Accuracy Studies, Systematic Review