Reference citations in radiology: accuracy and appropriateness of use in two major journals.

OBJECTIVE: Errors in reference citation and use are common in the medical and scientific literature. The prevalence of such errors in the radiology literature has not been reported. We did a study to assess the accuracy and appropriateness of use of references cited in two general radiology journals. MATERIALS AND METHODS: All references cited in the June 1993 issues of the American Journal of Roentgenology and Radiology were numbered consecutively. Fifty references were chosen at random from each journal, and copies of the original publications were obtained from the medical library at our institution or through interlibrary loan. Each reference was studied for accuracy and appropriateness of its citation in the June 1993 journal article (the "index article"). Errors were classified as major or minor in each category. Data were analyzed with the SAS statistical package. RESULTS: Forty-seven (94%) of 50 references were obtained from AJR, and 48 (96%) of 50 from Radiology. Of the 47 from the AJR, one (2%) had a major error and 21 (45%) had a minor error in accuracy. Of the 48 from Radiology, two (4%) had a major error and 11 (23%) had a minor error in accuracy. These values were significantly different for minor errors (p = .0188), but not for major ones (p = 1.000). When we adjusted for index article type, error rates for the two journals were not significantly different (p = .0612). We found four major errors (9%) and two minor errors (4%) in appropriateness of citation in the AJR references we studied. Three references (6%) from Radiology contained major errors in appropriateness of use; we found no minor errors of that type. These values were not significantly different (p = .232 for minor errors; p = .709 for major errors). One error in accuracy prevented location of the original reference. Errors were not related to the number of references cited in an index article (p = .528 for accuracy; p = .092 for appropriateness). CONCLUSION: The rate of minor errors in accuracy of references is fairly high in the two journals studied and is comparable to rates previously reported for other types of journals. The rate of major errors in accuracy of references is slightly lower than rates for other types of journals. The percentage of cited references that could not be located was also smaller than in previous reports. Errors in citation appropriateness were less common as well. Given the small number of errors that prevented references from being located, significant expenditure of time and money by journal staff members in checking references is probably not justified. However, authors should be encouraged to exercise greater care in checking all of their references for both accuracy and appropriateness of use. Differences in error rates between AJR and Radiology may have resulted in part from the random sampling method, which produced different mixtures of index articles for the two journals.