The literature of anaesthesia: what are we learning?

In an effort to identify the types of articles published in anaesthesia literature, a stratified random sample of articles published in North America between 1977 and 1986 was analyzed (N = 571). Human studies constituted 63 per cent of the total, with case reports and case series constituting over half. Study designs classes as descriptive in nature were remarkably rare in the anaesthesia literature, with prevalence and case-control studies constituting 0.8 and 3.3 per cent of the total respectively. Cohort studies (7.8 per cent), non-randomized intervention studies (12.8 per cent), and randomized controlled trials (17.8 per cent) were more numerous, but many suffered major contamination of experimental design. Frequently identified concerns in assessing the applicability of a given study to general anaesthetic practice were a bias induced by selection of the study subjects, application of the results from tertiary care hospitals to community hospitals, and contamination of the study protocol. These factors were identified as present in the majority of articles. The results suggest that growth of the specialty of anaesthesia is constrained by the narrow spectrum of study designs, as well as major problems affecting generalizability of the published results.