Evaluation of Statistical Methods for Articles Published in《Chinese Circulation Journal》 / 中国循环杂志

ABSTRACT

Objective: To evaluate the applications and problems of statistical methods for articles published in《Chinese Circulation Journal》. Methods: We reviewed the statistical methods for original articles published in《Chinese Circulation Journal》in 2010, 2011 and 2015; general status with problems was analyzed and the situations were compared among different years. Results: A total of 404 articles were enrolled, statistical hypothesis testing was performed in 366 of them, 32 had descriptive statistics only and 6 articles were meta-analysis. The most frequently used statistical methods were student'st-test (57.4%), followed by chi-squared analysis (56.7%), analysis of variance (35.2%) and regression analysis (23.0%). The main statistical problems in Methods included in described method un-matched to really used method and insufifcient or wrong description. The proportion of insufifcient or wrong description in 2010 and 2011 was 36.6% (71/194), while in 2015 was 26.2% (55/210),P=0.02. The major wrong application of statistical approach was ignoring design with improper method and using unsatisifed precondition for speciifc analysis. The proportion of improper statistical method application in 2010 and 2011 was 19.1% (33/173), in 2015 was 19.7% (38/193),P=0.88. Conclusion: The quality of statistics was improved in 2015 than 2010 and 2011 for articles published in《Chinese Circulation Journal》; while there were still problems at certain degree, enhanced statistical review should be conducted in medical journals.