Zhong Xiaobo, Cheng Bin, Wang Xinru, Cheung Ying Kuen
Department of Population Health Science and Policy, Icahn School of Medicine at Mount Sinai, New York, NY, USA.
Department of Biostatistics, Columbia University, New York, NY, USA.
PeerJ. 2021 Jan 11;9:e10559. doi: 10.7717/peerj.10559. eCollection 2021.
This article introduces an R package, SMARTAR (Sequential Multiple Assignment Randomized Trial with Adaptive Randomization), by which clinical investigators can design and analyze a sequential multiple assignment randomized trial (SMART) for comparing adaptive treatment strategies. Adaptive treatment strategies are commonly used in clinical practice to personalize healthcare in chronic disorder management. SMART is an efficient clinical design for selecting the best adaptive treatment strategy from a family of candidates. Although some R packages can help in adaptive treatment strategies research, they mainly focus on secondary data analysis for observational studies, instead of clinical trials. SMARTAR is the first R package provides functions that can support clinical investigators and data analysts at every step of the statistical work pipeline in clinical trial practice. In this article, we demonstrate how to use this package, using a real data example.
本文介绍了一个R包SMARTAR（具有自适应随机化的序贯多分配随机试验），临床研究人员可以通过它来设计和分析序贯多分配随机试验（SMART），以比较自适应治疗策略。自适应治疗策略在临床实践中常用于慢性疾病管理中的个性化医疗。SMART是一种从一系列候选策略中选择最佳自适应治疗策略的有效临床设计。虽然一些R包有助于自适应治疗策略研究，但它们主要侧重于观察性研究的二次数据分析，而非临床试验。SMARTAR是首个提供能在临床试验实践中支持临床研究人员和数据分析师完成统计工作流程各步骤功能的R包。在本文中，我们将通过一个实际数据示例演示如何使用这个包。
