Reporting of Randomised Controlled Trials in Hodgkin’s Lymphoma from 1966 to 2002. A Pre- and after CONSORT Evaluation.

Introduction/Background: Health professionals base treatment decisions primarily on scientific findings in the medical literature. Randomised controlled trials (RCTs) are considered the best tool to evaluate the applicability of clinical research but there is overwhelming evidence to indicate that the quality of reporting RCTs is less than optimal. The Consolidated Standards for Reporting Trials (CONSORT) statement was developed to help to improve the quality of reports of RCTs and published in 1996.

Methods and Material: Retrospective survey of reports of RCTs in Hodgkin’s lymphoma published in English and German between 1966 and 2002. Evaluation tool was a 14-item questionnaire based on the validated CONSORT checklist. Statistical analysis was performed by chi-square method and Fisher’s exact test where appropriate.

Results: The analysis of 243 published full-text articles in 59 journals from early 1966 to the end of 2002 indicates an improvement in the quality of reporting RCTs after CONSORT but this enrichment is not consistent. A highly significant enhancement post CONSORT was observed in 5 out of 14 reporting items (i.e. 35.7%; description of statistics methods, reporting of power calculation, ITT analysis, primary outcomes and 95% confidence interval (P=<.0005 – .002)). A sixth reporting item (adverse events) achieved borderline significance (P=.055). Intention-to-treat was explicitly mentioned in 11.9% (29/243), summary results for each comparative group was 98.4% (239/243) and precision of estimated effect size (95% confidence interval) were provided in 21.8% (53/243) respectively. Adverse events were reported in 77% (187/243), and conflict of interest was declared in 4.5% (11/243). Deviations in the reported versus performed randomised allocation in a subset of reports specifically mentioning ITT analysis (n=29) ranged from 2.5%–15%.

Conclusions: Despite years of educational efforts, reporting of randomised clinical trials for Hodgkin’s lymphoma in the scientific literature remains unsatisfactory and requires further concerted action by publishers, journal editors, learned societies and medical schools.