统计学分析极端重要

13.统计学分析极端重要

2012年,我们首次报道IL-27是诊断结核性胸腔积液良好的可溶性指标之后,我对此仍然念念不忘,并于2013年再次报告了另外一项重复该研究的结果。觉得还不过瘾,最近又利用朝阳医院和武汉协和医院两家单位从2013年初开始前瞻性收集的连续性胸水和血清标本(分别为154例和120例)第三次检测IL-27、腺苷脱氨酶和IFN-γ的浓度,希望通过更完善的科研设计来评价这些指标确切的诊断价值。

上述274例胸水和配对血清均为确诊病例的标本,所有未能明确病因的标本均被剔除在外,目的在于严格控制结果被混淆的可能性。目前,论文的初稿已经基本完成,希望能够发表在比较好的学术期刊上。

关于临床研究,没有哪一个环节是不重要的,因为任何一个细节都会严重影响结果的准确性和结论的可靠性。今天想强调一下统计学分析的重要性,实际上无论如何强调都不过分。下面的文字是我们新论文中的统计学部分初稿(未经修改),单单看一眼就该知道,花费在其中的心血是不容小觑的。

Statistical Analysis

The concentrations of IL-27 and ADA were normally distributed as determined by the Kolmogorov-Smirnov test and they are presented as means±SEMs,while IFN-γdata are presented as medians(25th to 75th percentiles)since they were not normally distributed.Differences between two or multiple groups were compared using Student’s t test,Mann–Whitney U test,one-way analysis of various(ANOVA),or Kruskal-Wallis ANOVA on ranks followed by Bonferroni’s test for multiple comparisons as appropriate.Comparisons of data in the pleural effusion and in the corresponding serum were made using paired t test or Wilcoxon signed-rank test.(https://www.daowen.com)

The receiver operanting characteristic curves were drawn and the areas under the curves(AUCs)were calculated in order to determine the diagnostic value of the concentrations of each biomarker in pleural effusion;8,9 and AUCs were compared using z-statistic with the Hanley and Mc Neil procedure.10 The optimum cutoff values were defined based on their highest diagnostic accuracy according to AUCs.The parameters of diagnostic accuracy are shown together with their 95%confidence intervals(CIs).The parameters of diagnostic accuracy obtained in the derivation cohort were compared with the validation cohort using Pearson chisquare test(or the exact Fisher test in 2×2 tables where the expected frequencies were lower than 5).All statistical analyses were performed with the use of SPSSand MedCalc softwares,and P<0.05 was considered to indicate statistical significance.

由于经常接触数据处理,我对于医学统计学的某些概念、理论和方法还算是有所了解。然而,我在这方面的知识远远不能满足工作的实际需要。过去八年以来,我每一年都会向同济医学院统计学教研室的尹平教授求教和求助若干次。可以这么说,我们很多比较重要的文章都融入了尹平教授的统计学贡献。

现代的临床研究越来越复杂,原先那些简单的卡方检验、t检验、方差分析等根本不再可能满足需要。有些庞大的研究工作离开统计学家的参与一开始就不能起步。譬如说,医生绝对没有能力设计出一项流行病学研究的方案,也不能保证技术路线的正确实施,得到数据之后更不会分层处理。反过来说也对,如果能够做好这些工作,其人肯定不是临床医生。没有人具有在两个不同领域的专业能力,除非满足于半生不熟。

很多时候,花时间阅读一些统计学专著是一种很有收获的美好事情,除了有趣,还能帮助人们从更多的角度看待同一批研究资料,有可能找到很多值得深入探讨的课题。(2016年8月11日)