Quality of interventional radiology literature: a review of articles published in JVIR and CVIR.

PURPOSE

To evaluate the quality of reporting of clinical studies published in two interventional radiology journals.

MATERIALS AND METHODS

Two investigators reviewed all articles reporting the outcomes from therapies in 12 consecutive months of Journal of Vascular and Interventional Radiology (JVIR) (August 2007 to July 2008) and CardioVascular and Interventional Radiology (CVIR) (July/August 2007 to May/June 2008). The included studies were evaluated by means of a score sheet adapted from the Consolidated Standards of Reporting Trials criteria. The score sheet was comprised of 22 categories, with each given a score of 0-2. These scores were summed (maximum score, 44) and the comparative results analyzed by using the Wilcoxon rank sum and chi(2) tests.

RESULTS

A total of 129 articles were reviewed from JVIR and 86 from CVIR. JVIR's mean score was 23.3 +/- 4.9, which was significantly higher than CVIR's mean score of 19.8 +/- 5.7 (P< .0001). Prospective studies comprised 38% (49 of 129) of JVIR's articles and 35% (31 of 86) of CVIR's studies (P = .9076). The mean sample sizes were larger for JVIR than for CVIR (130.8 and 66.3, respectively) (P = .0173). Both journals primarily published case series (112/129 [86.8%] for JVIR and 76/86 [88%] for CVIR). Only six of the 129 articles (4.6%) in JVIR and seven of the 87 (8.1%) in CVIR were randomized studies. Key weaknesses in reporting include lack of randomization, blinding of outcome assessment, sample size analysis, and proper reporting of outcomes.

CONCLUSIONS

Articles published in both journals displayed substantial weaknesses that potentially limit the validity of their conclusions.