The size of the gap between research evidence and clinical practise: questionnaire survey on treatment of patients with liver disease

Article type
Authors
Gluud L, Kürstein P, Kjellberg J, Olsen K, Willemann M, Søgaard J, Gluud C
Abstract
Background: Systematic reviews of randomized clinical trials generally provide the best research evidence for evaluating intervention effects. The extent, to which clinical practice concurs with the results of systematic reviews, may therefore reflect the propagation of evidence based medicine.

Objectives: The objective of this survey was to evaluate the agreement between clinical practice and evidence presented in systematic reviews. Methods: The design was a questionnaire survey on clinical practice and perceived effects of 36 interventions for hepato-biliary diseases that were evaluated in Cochrane reviews and were available in Denmark. The questionnaire was divided into 16 disease areas. Participants were asked about how often they used the interventions and whether they believed that the interventions reduced mortality or morbidity. Participating departments and doctors were selected by drawing of lots. Five university and 19 non-university hospital departments in Denmark participated. Questionnaires were mailed to 108 doctors of whom 54 were head of departments and 54 were specialist registrars or consultants. Sixty-six doctors replied and were included in our analyses. In the statistical analyses, the evaluated interventions were classified as significantly beneficial effects on mortality or morbidity (n=3), possible beneficial effects (n=19), or undetermined effects (n=14) based on the results of the Cochrane reviews. Regression analyses were used to evaluate potential predictors of frequent use.

Results: There was a considerable variation in clinical practice. For some interventions, clinical practice did not seem to concur with the evidence presented in Cochrane reviews. On average, 39% of respondents always used interventions with significant beneficial effects, 9% always used interventions with possible beneficial effects, and 2% always used interventions with undetermined effects. Sixty-nine percent of respondents believed that interventions with significant beneficial effects reduced mortality and 71% that they reduced morbidity. Forty percent of respondents believed that interventions with possible beneficial effects reduced mortality and 49% that they reduced morbidity. Thirteen percent of respondents believed that interventions with undetermined effects reduced mortality and 19% that they reduced morbidity. Perceived effects, employment at university department, and years in clinical practice were significantly associated with more frequent use of with significant beneficial effects.

Conclusions: Although interventions were used more often when supported by research evidence, there was a considerable gap between research evidence and clinical practice for some interventions. Additional studies are necessary to explore reasons for the demonstrated variation.