Reporting of randomisation in randomised-controlled trials on traditional Chinese medicine: A comparison of three Chinese medicine journals




Poster session 2 Thursday: Evidence synthesis - methods / improving conduct and reporting


Thursday 14 September 2017 - 12:30 to 14:00


All authors in correct order:

Han X1, Li X2, Zhang Y2, Liu J2
1 School of Humanities, Beijing University of Chinese Medicine, China
2 Centre for Evidence-Based Chinese Medicine, Beijing University of Chinese Medicine, China
Presenting author and contact person

Presenting author:

Xueyan Han

Contact person:

Abstract text
Background: By March 2017, the Cochrane CENTRAL has incorporated 19 918 citations and abstracts of randomised trials (RCTs) on traditional Chinese medicine (TCM) identified and translated from Chinese journals. However, many empirical studies have presented methodological problems in the published RCTs on TCM, among them, the reporting of randomisation is essential.

Objectives: To explore the trend of randomisation reporting in RCTs on TCM and its association with the inclusion status of the journal in the Core Journals Guide of China (which is similar to the Science Citation Index of China) amongst other factors.

Methods: Journal of Beijing University of Traditional Chinese Medicine (core journal), Journal of Guangzhou University of Traditional Chinese Medicine (core journal after 2005), Journal of Fujian University of Traditional Chinese Medicine (non-core journal) in the year 2005, 2008, 2011, and 2014 were chosen to identify RCTs on TCM. The inclusion statuses of the journals in the Guide and other factors associated with the reporting quality were analysed by logistic regression.

Results: 313 randomised trials were identified, of which 151 (48.24%) reported the sequence generation methods, improved over the 4 years of observation (31.51%, 44.59%, 49.28% and 62.89%); 10 (3.19%) reported details on allocation concealment (table 1). Year of publication (before or after 2010) (OR 2.08, 95% CI 1.32-3.30) and the inclusion in the Guide (OR 1.60, 95% CI 1.01-2.53) were independently associated with the reporting of sequence generation. Funding levels (funding of national level reported or not), number of authors and intervention types had no statistically significant association with the reporting of randomisation.

Conclusions: The reporting of sequence generation shows improvement but the reporting on allocation concealment is alarmingly low, indicating the comprehensive methodological training is still warranted. Meanwhile, RCTs from the journals listed in the Core Journal Guide have a slightly higher reporting quality, indicating the possibility that the overall reporting quality can be improved by utilising a more rigorous vetting system.