Zhang Ting D, Scheinberg Mila, Hoffman Alexander, Mehta Labdhi, Hayes Caleb, Schick Samuel, Bernstein Marc, Shah Ashish
Orthopedic Surgery, University of Alabama at Birmingham, Birmingham, USA.
Orthopedics, University of Alabama at Birmingham, Birmingham, USA.
Cureus. 2024 Oct 30;16(10):e72709. doi: 10.7759/cureus.72709. eCollection 2024 Oct.
Statistical tests are important tools in research as they provide a systematic and objective approach to analyzing data, testing hypotheses, and drawing conclusions. We investigated the application of statistical tests used in articles published in the Journal of Bone and Joint Surgery (JBJS) over the last 30 years. We searched PubMed for JBJS articles published from 1993 to 2023 and randomly sampled 5% of articles from each year. Our inclusion criteria identified a selection of articles for review, and we collected and analyzed variables, including statements of significance, significance level, power calculations, usage of statistical tests and methods, power analysis, and orthopedic subspecialties. Articles were grouped into three 10-year periods, and then analysis of variance, Tukey's honestly significant difference, and chi-square tests were conducted to investigate the changes between decade intervals. Our 5% stratified random sample of JBJS publications yielded 593 unique articles, of which 559 were accessible as full-text articles. Common statistical tests included t-test, chi-squared, and ANOVA. The average number of statistical tests per paper increased from 1993-2002 (n = 0.64) to 2013-2023 (n = 1.25) (p value = 0.0009). Among the articles stating alpha levels (n = 156), 0.05 was most prevalent (94%), with 0.01 and 0.1 less commonly used (3.2% and 1.2%, respectively). Article types varied, with "commentary, editor, forum" being the most common (n = 124), followed by retrospective cohort studies (n = 105). The most frequent power goals were 80% and 90%. In studies reporting power calculations, 34% declared adequate power, 14% admitted insufficient power, and 52% did not disclose adequacy. Statistical test and publication trends varied greatly between orthopedic subspecialties. With the increasing usage of statistical tests in orthopedic research, it is increasingly important for readers to have a strong understanding of statistical analysis. This foundational knowledge will enable critical engagement with literature and aid in the growth of the orthopedic field.
统计检验是研究中的重要工具，因为它们为分析数据、检验假设和得出结论提供了一种系统且客观的方法。我们调查了过去30年发表在《骨与关节外科杂志》（JBJS）上的文章中所使用的统计检验的应用情况。我们在PubMed上搜索了1993年至2023年发表的JBJS文章，并从每年的文章中随机抽取5%。我们的纳入标准确定了一批用于综述的文章，并且我们收集并分析了各种变量，包括显著性声明、显著性水平、功效计算、统计检验和方法的使用、功效分析以及骨科亚专业。文章被分为三个10年期，然后进行方差分析、图基（Tukey）真实显著性差异检验和卡方检验，以研究十年间隔之间的变化。我们对JBJS出版物进行的5%分层随机抽样产生了593篇独特的文章，其中559篇可获取全文。常见的统计检验包括t检验、卡方检验和方差分析。每篇论文的统计检验平均数量从1993 - 2002年（n = 0.64）增加到2013 - 2023年（n = 1.25）（p值 = 0.0009）。在声明了α水平的文章中（n = 156），0.05最为常见（94%），0.01和0.1使用较少（分别为3.2%和1.2%）。文章类型各异，“评论、编辑、论坛”最为常见（n = 124），其次是回顾性队列研究（n = 105）。最常见的功效目标是80%和90%。在报告功效计算的研究中，34%声明功效充足，14%承认功效不足，52%未披露功效是否充足。骨科亚专业之间的统计检验和发表趋势差异很大。随着统计检验在骨科研究中的使用增加，读者深入理解统计分析变得越来越重要。这一基础知识将有助于对文献进行批判性参与，并促进骨科领域的发展。
