Background: Multiple publication is a type of reporting bias where a study is published more than once. There is evidence that multiple publication introduces bias in systematic reviews and meta-analyses of pharmacological interventions. The frequency and characteristics of multiple publications of trials that assess the effectiveness of complementary and alternative medicine (CAM) interventions, such as meditation practices, are unknown.
Objectives: To investigate the frequency of multiple publication of trials assessing the effects of meditation practices and to describe the characteristics of these reports.
Methods: A systematic search was performed for full published reports of randomized controlled trials (RCTs) and controlled clinical trials (CCTs) on the effects of meditation practices for a variety of health-related outcomes in adult populations, with sample sizes greater than 10 participants. Fifty-two electronic databases were searched with restriction to the English literature (date of the last search September 2005). Additional reports were identified from reference tracking and hand searches of relevant journals. Trials reported as conference abstracts, thesis or dissertations, or unpublished trials were not included in the study. The existence of multiple publication was assessed by cross checking authors' names, interventions, comparison groups, sample size, sample composition, and results. We classified a published article as a stand-alone publication when the trial has been published just once whereas articles based on data from two or more submitted studies were classified as trials with multiple publications.
Results: From 477 trials on the effects of a variety of meditation practices, 392 were included in the review. A total of 54 (13.8%) multiple publications originated from 37 trials of meditation practices (15 trials on tai chi, 13 on yoga and transcendental meditation each, 4 on mindfulness interventions, qi gong, and the relaxation response each, and 1 on a mantra meditation practice not further classified). Of the 37 trials, 29 were published twice, 5 trials three times, 1 trial four times, 1 trial five times, and 1 trial six times. The median year of publication of the multiple publications was 1996 (interquartile range (IQR):1984, 2003). The median time elapsed from the first publication to a multiple publication was 2 years (IQR: 0.7 to 3.25 years). The majority of the multiple publications originated in the USA (67%). Of the 54 multiple publications, 46 had the same first author as the original publication, but discrepancies in the numbers of authors in identical or virtually identical papers were identified in 36 trials. Eight multiple publications reordered authors without substantial change in content from the original publication. The 54 multiple publications were published in 41 different journals with 8 multiple publications published in the same journal as the original publication. Four of the 37 trials with multiple publications were multicenter studies, 2 of which were published six times as single center studies. The remaining 33 trials were single centre studies, two of which also produced 6 multiple publications. Of the 91 reports (37 original publications and 54 multiple publications), 56% declared funding. Discrepancies in the reporting of funding when compared with the original publication were identified in 15% of the multiple publications.
Conclusions: The frequency of multiple publications in a cohort of trials on the effectiveness of meditation practices is similar to that reported in other areas of scientific inquiry. The impact of undetected multiple publication in systematic reviews and meta-analyses of the effectiveness of health care interventions is a methodological problem that should be further explored. Authors of future studies on meditation and other CAM approaches should adhere to guidelines for good publication practices and avoid multiple publications.