Content area
Full Text
The challenges encountered in conducting a systematic review are numerous, but few are more vexing than trying to sort out the bewildering array of disaggregation (when results from a multicentre trial are presented in several publications) as well as possible, probable, and certain redundant reporting of trials, in a scholarly attempt to prevent double-counting of data. This situation, in which one has to basically trace the genealogy of trials to determine the relations between trial reports, is not only intolerably time-consuming but also unjustifiable.
Multiple renditions of the same information is selfserving, wasteful, abuses the volunteer time of peer reviewers, and can be profoundly misleading; it brings into question the integrity of medical research. This is a pernicious trend in medical publication that needs to be addressed vigorously by the medical community.
In the mid-1980s, Huth1 described multiple publications of a single study as salami science, and asserted that this was damaging because of its economic implications for publishers, readers, libraries, and indexes. Kassirer and Angell2 have reiterated the position, generally accepted by medical journals, that reporting on the same trial more than once is unacceptable, except in very specific circumstances. One exception was large multicentre trials with numerous endpoints. They did not specify the limits of acceptability when such a trial had few a-priori specified primary endpoints, which is a more common situation. As a case study, we offer the following example of both redundancy and disaggregation in the reporting of two multicentre trials that had two primary endpoints.
Risperidone is a relatively new antipsychotic heralded as an important milestone in the treatment of schizophrenia. We did a literature search for primary research on this drug and identified 20 articles and several unpublished reports describing randomised, double-blind controlled trials. We believe that only two major and seven smaller trials were done. We discovered obvious redundancy in the results of a single-centre trial being published twice.3,4 However, identifying the disaggregation in the multicentre trials-which we called their genealogies-was far from obvious, because of the chronology of publications, changing authorship, lack of transparency in reporting, and the frequent citation of abstracts and unpublished reports. The trial design of all 20 studies was simple: various doses of risperidone were compared with another antipsychotic drug or placebo in...