From the Department of Mathematics, Universidade da Coruña, A Coruña, Spain (AQ-d-R); Department of Physiotherapy, Medicine and Biomedical Sciences, Universidade da Coruña, A Coruña, Spain (BR-R, VR-G, JC-M, AM-R); and Department of Physical and Sports Education, Universidade da Coruña, A Coruña, Spain (PA-R).
Am J Phys Med Rehabil. 2019 Jun;98(6):516-520. doi: 10.1097/PHM.0000000000001124.
Bayesian techniques, as an alternative method of statistical analysis in rehabilitation studies, have some advantages such as handling small sample sizes, allowing incorporation of previous experience of the researchers or clinicians, being suitable for different kinds of studies, and managing highly complex models. These characteristics are important in rehabilitation research. In the present article, the Bayesian approach is displayed through three examples in previously analyzed data with traditional or frequentist methods. The studies used as examples have small sample sizes and show that the Bayesian procedures enhance the statistical information of the results. The Bayesian credibility interval includes the true value of the corresponding parameter diminishing uncertainty about the treatment effect. In addition, the Bayes factor value quantifies the evidence provided by the data in favor of the alternative hypothesis as opposed to the null hypothesis. Bayesian inference could be an interesting and adaptable alternative statistical method for physical medicine and rehabilitation applications.
贝叶斯技术作为康复研究中统计分析的一种替代方法，具有一些优势，例如处理小样本量、允许纳入研究人员或临床医生的先前经验、适用于不同类型的研究以及处理高度复杂的模型。这些特点在康复研究中很重要。在本文中，通过三个使用传统或频率方法分析过的数据的例子展示了贝叶斯方法。这些例子所使用的研究样本量较小，表明贝叶斯程序增强了结果的统计信息。贝叶斯可信区间包括相应参数的真实值，减少了对治疗效果的不确定性。此外，贝叶斯因子值量化了数据支持备择假设而不是零假设的证据。贝叶斯推断可能是物理医学和康复应用中一种有趣且适应性强的替代统计方法。
