Related Experiment Videos
[Preliminary assessment of articles reporting cancer diagnostic tests]
1Department of Clinical Epidemiology, First Clinical College, Chinese Medical University, Shenyang 110001, China.
Objective:
To improve the quality of papers on clinical diagnostic tests.
Methods:
Articles of diagnostic tests published in the Chinese Journal of Oncology in the year 1990 were surveyed by cluster sampling method.
Results:
In the 23 articles surveyed, immunoassays constituted the largest category, accounting for 39.1% (9/23), followed by biochemical markers and radiological tests, accounting for 17.4% (4/23) and 13.0%(3/23), respectively. Of the 23 articles, 19 (82.6%) employed a well-defined "gold standard", 14 (60.9%) correctly calculated sensitivity and specificity. The interpretation of the test was stated to be "blind" in only 39.1%(9/32). In only one paper were all the seven methodologic criteria adhered. Less than 60% of the articles followed more than four of the seven methodologic criteria, and in 2 articles only one of the criteria was met with.
Conclusion:
Although due attention has been paid to the methodologic criteria of assessing diagnostic tests, the quality of the articles surveyed needs to be improved.