| Objectives' To determine if diagnostic study was conducted aPpropriately inpapers published in Chinese Journal of Radiology, 1l) we assessed the reportingquality of diagnostic test published in Chinese Joumal of Radiology from1998-2000; (2) we investigated the current sitwtion of diagnostic study viastudying how the radio1ogists understand the design, methodology and criticalaPpraisal standard of diagnostic test and how they use them in their research; (3)we provided some suggestions and reconunendations on how to imProve thequality of diagnostic tests in radiology.Mcthods: (1) We hand searched all the papers published from l998-2000 inChinese Joumal of Radiology and collected the total number of published paPers,nuxnber of diagnostic paPers Which were determined if the study was used todiagnose the patients' disease. Two reviewer indePendently selected thediagnostic study. (2) A questionnaire based, on globally accePted standards ondiagnostic test was designed to assess the reporting quality of diagnostic test. ApiIot study on l0 papers frOm issues l-3 of 200l was used to assess the reliabiIityof the questionnaire. TWo reviewers independently assessed the papers anddisagreement was resolved by discussion. The agreement of the two reviewerexpressed as Kappa is 0.85. The questionnaire was revised based on the pilot study.Formal assessmel1t of the reporting quality of the selected diagnostic test wasperformed by one of tl1e two reviewers involved in the pilot study. Furtherproblem during the assessn1el1t was discussed with a clinical ePidemiologist. (3) Aquestionnaire was designed to investigate 500 editors, joumal peerreviewers andreaders how they understand diagnostic test on imaging. The validity andreliability, the pilot study were also conducted. The questionnaire was revisedbased on the pilot studyResults: (l) About l095 papers were published on Chinese Journal ofRadiology from l998-2000, among l29 papers on diagnostic test (l2%), 437papers relevant to clinical research (30%). (2) For the l29 papers on diagnostictest, 87 (67.4%) paPers performed the goId standard on aIl the study subjects, 92(7l .3%) paPers comPared the diagnostic tests and the gold standard, l8 (l4.4%)paPers blindly compared the diagnostic tests and the gold standard, 55 (42.6%)paPers aPpropriately selected the stUdy subjects, 23 (l7.8%) clearly described thesources of study subjects, 6l (47.3%) paPers caIcuIated at least one parameters ofdiagnostic test. Only 8 (6%) paPers met the 3 most important methodologicalcriteria: gold standard, blindly comParison the diagnostic test and gold standardand aPpropriately selected subect. The sample size arranged frOm 7-l000 with themedian of 38, but none of the paPer caIcuIated the sample size. Thirty paPersconsidered the interobserver Weement, 6 paPers considered the measurementvariation, but only 2 paPers calculated the coefficient vari8tion (CV). (3) Fivehundreds questiotmaires were seni to investigate the editors, joumaIpeerreviewers and readers, 43l(86%) questiotmaires were coIlected. Among 43lparticipants, 229(54.6%) had accePted training on research methodology withmost in 30-40 years old (43.4%), and 42.0% were doctors with bachelor, master ordoctoral degree. Seventy-five of l78 (42.l%) bachelors, l03 of l l9 (86.6%)master and 5l of 59 (86.4%) doctoraI degree accepted tr4ining on researchmethodology. 84.3%, 87.9%, 78.7%, 66.6%, 58.7%, 52.2%, 20.9% and l8.3%participants know sensitivity, specificity, accuracy, prevalence, positive andnegative predictive value, positive and negative likelihood ratio respectively.Conclusion: Study on diagnostic test is not appropriately conducted. Both the number and quality of diagnostic test cannot fulfill the need of clinical medicine. Radiologists and health professionals of imaging must accepted training on research methodology in order to conduct high quality study and provide valid, reliable information. Editors...
|
PDF Full Text Request