Bridging clinical investigators and statisticians: publication status and problems of Statistical Analysis Plan (SAP)

ID: 

178

Session: 

Poster session 1

Date: 

Monday 24 October 2016 - 10:30 to 11:00

Location: 

All authors in correct order:

Zhang Y1, Li X1, Fei Y1, Wang S1, Liu J1
1 Beijing University of Chinese Medicine, China
Presenting author and contact person

Presenting author:

Xun Li

Contact person:

Abstract text
Background: The publication of study protocols has been increasingly accepted as a means for improving transparency and quality research. However, both commercial- and investigator-initiated trials are prone to inappropriate analysis and interpretation. The main reason for this may arise from ignorance of statistical methods. Furthermore, post hoc analyses which were not prespecified in the protocol involve laborious statistics and have the potential risk of misreporting and misleading in the conclusions.

Objectives: We aimed to learn the profile of the publication status of the Statistical Analysis Plan (SAP) and explore the existing problems.

Methods: A systematic literature search of PubMed was conducted from the date of inception onwards. Research papers that particularly stated the statistical analysis plans of clinical trials were included this review.

Results: A total of 66 articles was identified. After reading the titles and abstracts, 29 articles from 10 different countries were kept and others were excluded because they were reviews, methodology papers, or their objectives differed from this review. Two of the trials were conducted in China. Ten of the trials reported the SAP by British investigators. Twelve (41%) trials were ongoing until 10 March 2015. Six trials declared the interim analyses. Only four trials were not supported by foundations.

Conclusions: 'Data dredging' is an important factor for misleading conclusions. But SAPs usually did not get the consensus between clinical investigators and statisticians. In the meantime, the publications of SAPs were not given enough attention to maximum the transparency of study.