Value of flow diagrams in reports of randomized controlled trials

CONTEXT: Diagrams of the flow of participants through a clinical trial are recommended in the Consolidated Standards for Reporting of Trials (CONSORT) statement, but it is unclear whether such flow diagrams improve the quality of trial reports. OBJECTIVE: To examine the information contributed by flow diagrams and the completeness of reporting overall in reports of randomized controlled trials (RCTs) published in 5 general and internal medicine journals. DESIGN AND SETTING: Analysis of 270 reports of RCTs published in 1998 in the Annals of Internal Medicine (AIM; n = 19), BMJ (n = 42), JAMA (n = 45), The Lancet (n = 81), and The New England Journal of Medicine (NEJM; n = 83). MAIN OUTCOME MEASURES: Proportion of reports that included a flow diagram, information provided in flow diagrams, and completeness of reporting about flow of participants overall in flow diagrams or text. RESULTS: A total of 139 reports (51.5%) of RCTs included a flow diagram, but this varied widely among journals (AIM, 21.0%; BMJ, 38.1%; JAMA, 80.0%; The Lancet, 93.8%; and NEJM, 8.4%). Diagrams generally provided useful information, but only 73 (52.5%) included the number of participants who received allocated interventions and only 32 (23.0%) included the number of participants included in the analysis. In logistic regression analysis, overall completeness of reporting about flow of study participants was associated with publication of a flow diagram. CONCLUSIONS: Flow diagrams are associated with improved quality of reporting of randomized controlled trials. However, the structure of current flow diagrams is less than ideal. We propose a revised flow diagram that includes all important counts through the stages of parallel group trials.