Audit of reports of randomised clinical trials published in one journal over 45 yearsBMJ 1995; 311 doi: https://doi.org/10.1136/bmj.311.7010.918 (Published 07 October 1995) Cite this as: BMJ 1995;311:918
- Correspondence to: Dr van der Wijden.
- Accepted 7 July 1995
The randomised clinical trial is considered to be one of the most reliable and therefore most important methods of investigation in medicine. This is because the random allocation of patients to an intervention or control group and the ignorance of patient and observer of what treatment is being undergone is believed to exclude subjectivity and selection bias that could interfere with the results.
Published reports of randomised clinical trials began to appear in the late 1940s,1 since when many have appeared. Not all randomised clinical trials ever performed can be retrieved from the international literature.2 Since the late 1980s randomised clinical trials have been increasingly analysed in systematic reviews and meta-analyses.3 The first phase of data collection for a systematic review of randomised clinical trials involves identifying as high a proportion as possible of the potentially relevant trials. A call for help in establishing an international register of randomised clinical trials was made at the initiation of the Cochrane Collaboration towards the end of 1992 to establish a database of systemic, up to date reviews of randomised clinical trials of health care.4
The Nederlands Tijdschrift voor Geneeskunde (Dutch Journal of Medicine), a general medical scientific weekly, has contributed to the Cochrane Collaboration by identifying all articles published that might be eligible for inclusion in a systematic review of randomised clinical trials.
Material, methods, and results
We defined randomised clinical trial as “a study design in which patients are allocated at random to an intervention group undergoing a particular intervention whether for diagnostic, preventive, or therapeutic purposes, or to a control group.” At the editorial office all original papers obviously reporting trials which had been published in the Dutch Journal of Medicine between 1 May 1948 and 1 May 1993 were identified. This was done by reading all the original papers published in this period. Then we studied the method of randomisation.
The whole study took about 10 weeks. Of the total of 6820 original papers published in the studied period 89 papers (1.3%) described a trial. The method of randomisation was explained in a clear way in 38 of these (0.6% of the total). The year of publication of the 89 reports is reflected in the table.
Like the BMJ, the Nederlands Tijdschrift voor Geneeskunde published its first randomised clinical trial in 1948: “Een onderzoek naar de suppressieve werking van paludrine bij malaria tertiana” (A study of the suppressive action of paludrine in tertian malaria) by Dr S Klopper of Wormerveer, in collaboration with Miss D Slop, analyst, and Miss C Op ‘t Land.5 During the five years following publication of this first report nine randomised clinical trials were published, compared with 25 reports in the past five years.
The 89 trials found in our journal have been submitted for inclusion in the international register of randomised clinical trials. As suggested by Dickersin et al,2 the key word “randomised clinical trial” has been introduced at the journal (and included in the electronic records since 1986), so that retrieving trials will be easier in the future, improving the quality of reporting.
We thank Dr Iain Chalmers, director of the Cochrane Centre, Oxford, for his suggestions.
Conflict of interest None.