This study was carried out to determine if the content of structured abstracts conforms with recommendations of the Ad Hoc Working Group for the critical appraisal of the medical literature as adopted by the Annals of Internal Medicine. The study design was a survey. All articles published in Annals of Internal Medicine in 1991, excluding editorials, case-reports, literature reviews, decision analysis, studies in medical education, descriptive studies of clinical and basic phenomena, and papers lacking a structured abstract, were studied. Of a total of 150 articles, 20 were excluded. The abstract and text of each article were assessed for the presence of the following items; patient selection criteria, statements concerning extrapolation of findings, need for further study, and whether or not the information should be used now. Number of refusers, drop outs and reason(s) for drop outs were assessed for intervention and prospective cohort studies only. Deficiencies of assessed items were noted in both abstracts and texts. For abstracts, patient selection criteria, numbers of refusers, number of drop outs and reason(s) for drop outs were reported in 44.6% (58/130), 3.1% (4/130), 16.9% (14/83) and 2.4% (2/83) respectively. These items were reported more frequently in the texts 87.7% (114/130), 9.2% (12/130), 60.2% (50/83) and 37.3% (31/83) respectively (p < 0.05). Statements concerning extrapolation of findings, need for further study and use of information now were also more frequent in texts than abstracts (p < 0.0001). A large number of structured abstracts published in the Annals of Internal Medicine in 1991, lack information recommended by the Ad Hoc Working Group. Our findings should not be extrapolated to other journals requiring structured abstracts.