Lam Simon W, Bauer Seth R, Yang Wei, Miano Todd A
1 Cleveland Clinic, Cleveland, OH, USA.
2 Perelman School of Medicine at the University of Pennsylvania, Philadelphia, PA, USA.
Ann Pharmacother. 2017 May;51(5):429-438. doi: 10.1177/1060028016686356. Epub 2017 Jan 7.
Proficiency in research design and statistical analysis is crucial to the success of a clinical pharmacist. However, new pharmacy graduates and residents may not have received adequate training and education in these areas. During the authors' tenure as clinical pharmacists, several statistical "myths" were consistently maintained by residents and new clinical practitioners. The purpose of this narrative review is to discuss and dispel several of these statistical fallacies. The myths discussed involve 3 common areas of consideration when evaluating any clinical study: assessing the risk of bias from confounding (propensity score analysis and multivariable modeling), interpretation of the main study findings ( P values and hypothesis testing), and secondary evaluations (subgroup analyses). Literature examples are used to illustrate each of the topics. The authors hope that the discussion will augment each pharmacist's knowledge of medical literature interpretation leading to improvements in patient care, education of future residents, and personal research endeavors.
精通研究设计和统计分析对于临床药师的成功至关重要。然而，新毕业的药学专业学生和住院医师可能在这些领域没有接受过充分的培训和教育。在作者担任临床药师期间，住院医师和新的临床从业者一直存在几个统计学“误区”。本叙述性综述的目的是讨论并消除其中的一些统计谬误。所讨论的误区涉及评估任何临床研究时常见的三个考量领域：评估混杂因素导致的偏倚风险（倾向得分分析和多变量建模）、主要研究结果的解读（P值和假设检验）以及二次评估（亚组分析）。文中使用文献实例来说明每个主题。作者希望该讨论能增进每位药师对医学文献解读的知识，从而改善患者护理、对未来住院医师的教育以及个人研究工作。
