临床意义与统计学意义结合的临床试验结果分类方法及评价研究

目的:统计学意义在临床试验结果解读中发挥重要作用，但在获得有统计学意义结果时，临床意义的评估常被忽略，本研究尝试提出一种简单且明确的试验结果新分类方法，重点针对已获得统计学阳性的试验，评估其结果是否具有临床意义。方法:以2019年美国心脏病学会（ACC）和欧洲心脏病学会（ESC）大会公布结果的临床试验为研究对象，通过荟...

Integration of clinical significance and statistical significance on clinical study results categorization: a Meta-epidemiology study

Objective Statistical significance plays an important role in the interpretation of clinical trial results. However, on the basis of obtaining statistical significance, the assessment of clinical significance is often neglected. This study attempted to propose a simple and unambiguous new classification method for study results, focusing on studies with statistical positive findings to evaluate whether the results have clinical significance. Methods Our study subjects were the clinical studies in 2019 ACC and ESC annual meetings. Meta-epidemiology methods were used to extract the characteristic variable from each study. The primary evaluation indicators included target effect-size and observed effect-size. Based on the difference between the two indicators, the studies that had statistical significance were subdivided to identify studies with possible insufficient clinical significance; Furthermore, the theoretical threshold based on power analysis was proposed, which was used as the basis for the interpretation of study results. Results There were 12 clinical studies included in the final analysis. All of them were published on top journals. Those studies had relative high quality on both study design and reporting. The correlation coefficient between the observed and target effect-size was 0.892. Among the 7 studies with statistical significance, two of them were classified as insufficient clinical significance. The counts was 1 (1/3) and 1 (1/4) for the studies reported in ACC and ESC respectively. Conclusions The achievement of clinical significance is critical even in the study with positive results. This paper proposes a new classification standard that combines clinical significance with statistical significance and further suggests a method to evaluate the reliability of clinical study results in order to assist researchers in identifying potential risks caused by insufficient clinical significance, and provide some reference and help for the reasonable interpretation of clinical study results.