Trust, but verify. The accuracy of references in four anesthesia journals.

To determine the accuracy of bibliographic citation in the anesthesia literature, we reviewed all 1988 volumes of ANESTHESIOLOGY, Anesthesia and Analgesia, British Journal of Anaesthesia, and Canadian Journal of Anaesthesia and sequentially numbered all references appearing in that year (n = 22,748). One hundred references from each of the four journals were randomly selected. After citations to nonjournal articles (i.e., books or book chapters) were excluded, the remaining 348 citations were analyzed in detail. Six standard bibliographic elements--authors' names, article title, journal title, volume number, page numbers, and year--were examined in each selected reference. Primary sources were reviewed, unless our institution did not own the source or could not obtain it through interlibrary loan, in which case standard indexes, abstracting services, and computerized databases were consulted. Each element was checked for accuracy, and references were classified as either correct or incorrect. A reference was correct if each element of the citation was identical to its source. Of the examined references, more than half (50.3%) contained an error in at least one element. The elements most likely to be inaccurate were, in descending order, article title, author, page numbers, journal title, volume number, and year. No significant differences (P = 0.283) existed in the error rates of the four journals; the percentage of citations containing at least one error ranged from 44% (Anesthesia and Analgesia) to 56% (British Journal of Anaesthesia). The citation error rate of anesthesia journals is similar to that reported in other specialties, where error rates ranging from 38% to 54% have been documented.