Welcome to Chinese Journal of Clinical Pharmacology and Therapeutics,Today is Chinese

Chinese Journal of Clinical Pharmacology and Therapeutics ›› 2019, Vol. 24 ›› Issue (10): 1155-1160.doi: 10.12092/j.issn.1009-2501.2019.10.011

Previous Articles     Next Articles

Blinded analysis of double-blind randomized controlled trial for unequal integer ratio grouping

LIU Yaqi 1,2, LU Mengjie 2, LIU Yuxiu 1,2,3, CHEN Yu 3, YUAN Yangdan 3, ZHOU Chongchong 1,2, LIU Man 1,2   

  1. 1 Department of Biostatistics, School of Public Health, Nanjing Medical University, Nanjing 211166, Jiangsu, China; 2 Department of Medical Statistics, Jinling Hospital, Nanjing 210002, Jiangsu, China; 3 Department of Biostatistics, School of Public Health, Southern Medical University, Guangzhou 510515, Guangdong, China
  • Received:2019-04-19 Revised:2019-05-09 Online:2019-10-26 Published:2019-10-28

Abstract:

AIM: The blinded data analysis under the conventional equal ratio grouping is easy to be realized in the double-blind randomized controlled trial.However, there is no effective blinded analysis method for the trial of unequal integer ratio grouping.To develop a blinded analysis under the unequal integer ratio grouping to avoid the bias during data analysis due to knowledge of grouping information in the double-blind randomized controlled clinical trials. METHODS: In the design phase, according to the sum of unequal integer ratios, the samples were divided into several equal split groups.The random allocation sequence was generated by means of the new split groups.The drug labeling was conducted by a third party independent of the study based on the split groups.In the statistical analysis phase, all possible combinations of the split groups were analyzed and reported.RESULTS:The implementation of blind analysis was illustrated with a clinical trial in which two groups were grouped in 2∶1 ratio. In this method, only the codes of each subgroup were known after the first level unblinding, and the corresponding real group could not be identified, so it could keep blinded to treatment allocation in the statistical analysis.The final reporting of statistical analysis could be determined after the second level unblinding.CONCLUSION: The method we proposed is simple in principle, easy to implement and in line with the regulatory requirements.In spite of increasing workload during the random allocation sequence generating, drugs labeling,and statistical analysis,it could maintain the blindness of data analysis under the unequal integer ratio grouping and effectively avoid the possible analysis bias caused by unblinding.It has unique practical value in ensuring objectivity and science of clinical trial analysis.

Key words: double-blind randomized controlled trial, integer ratio grouping, blinded analysis

CLC Number: