Holmes Tyson H, Zulman Donna M, Kushida Clete A
*Stanford University Human Immune Monitoring Center, Institute for Immunity Transplantation and Infection, Stanford University School of Medicine, Stanford †Center for Innovation to Implementation, VA Palo Alto Health Care System, Menlo Park ‡Department of Medicine, Division of General Medical Disciplines §Stanford Sleep Medicine Center and Department of Psychiatry and Behavioral Sciences, Stanford University School of Medicine, Stanford, CA.
Med Care. 2017 Dec;55(12):e120-e130. doi: 10.1097/MLR.0000000000000464. Epub 2016 Jan 13.
Variable adherence to assigned conditions is common in randomized clinical trials.
A generalized modeling framework under longitudinal data structures is proposed for regression estimation of the causal effect of variable adherence on outcome, with emphasis upon adjustment for unobserved confounders.
A nonlinear, nonparametric random-coefficients modeling approach is described. Estimates of local average treatment effects among compliers can be obtained simultaneously for all assigned conditions to which participants are randomly assigned within the trial. Two techniques are combined to address time-varying and time-invariant unobserved confounding-residual inclusion and nonparametric random-coefficients modeling. Together these yield a compound, 2-stage residual inclusion, instrumental variables model.
The proposed method is illustrated through a set of simulation studies to examine small-sample bias and in application to neurocognitive outcome data from a large, multicenter, randomized clinical trial in sleep medicine for continuous positive airway pressure treatment of obstructive sleep apnea.
Results of simulation studies indicate that, relative to a standard comparator, the proposed estimator reduces bias in estimates of the causal effect of variable adherence. Bias reductions were greatest at higher levels of residual variance and when confounders were time varying.
The proposed modeling framework is flexible in the distributions of outcomes that can be modeled, applicable to repeated measures longitudinal structures, and provides effective reduction of bias due to unobserved confounders.
在随机临床试验中，对指定条件的依从性存在差异是常见现象。
提出一种纵向数据结构下的广义建模框架，用于对可变依从性对结局的因果效应进行回归估计，重点是对未观察到的混杂因素进行调整。
描述了一种非线性、非参数随机系数建模方法。对于试验中参与者被随机分配的所有指定条件，可以同时获得依从者中局部平均治疗效应的估计值。两种技术相结合以解决随时间变化和不随时间变化的未观察到的混杂因素——残差纳入和非参数随机系数建模。这两者共同产生了一个复合的两阶段残差纳入工具变量模型。
通过一组模拟研究来说明所提出的方法，以检验小样本偏差，并将其应用于一项大型多中心睡眠医学随机临床试验的神经认知结局数据，该试验用于持续气道正压通气治疗阻塞性睡眠呼吸暂停。
模拟研究结果表明，相对于标准比较器，所提出的估计器减少了可变依从性因果效应估计中的偏差。在残差方差较高以及混杂因素随时间变化时，偏差减少最为显著。
所提出的建模框架在可建模的结局分布方面具有灵活性，适用于重复测量的纵向结构，并有效减少了未观察到的混杂因素导致的偏差。
