Skip to main content
QUICK REVIEW

[论文解读] A Meta-Learning Method for Estimation of Causal Excursion Effects to Assess Time-Varying Moderation

Jieru Shi, Walter Dempsey|arXiv (Cornell University)|Jun 28, 2023
Advanced Causal Inference TechniquesMathematics被引用 3
一句话总结

本文提出了一种元学习框架,采用去偏/正交机器学习方法,以估计微随机试验(MRTs)中的因果激发效应,实现对时变治疗调节的高效且稳健的评估。通过利用监督学习估计干扰参数,并确保双重稳健性,该方法相比传统的加权最小二乘法在偏差减少和效率提升方面表现更优。

ABSTRACT

Advances in wearable technologies and health interventions delivered by smartphones have greatly increased the accessibility of mobile health (mHealth) interventions. Micro-randomized trials (MRTs) are designed to assess the effectiveness of the mHealth intervention and introduce a novel class of causal estimands called "causal excursion effects." These estimands enable the evaluation of how intervention effects change over time and are influenced by individual characteristics or context. Existing methods for analyzing causal excursion effects assume known randomization probabilities, complete observations, and a linear nuisance function with prespecified features of the high dimensional observed history. However, in complex mobile systems, these assumptions often fall short: randomization probabilities can be uncertain, observations may be incomplete, and the granularity of mHealth data makes linear modeling difficult. To address this issue, we propose a flexible and doubly robust inferential procedure, called "DR-WCLS," for estimating causal excursion effects from a meta-learner perspective. We present the bidirectional asymptotic properties of the proposed estimators and compare them with existing methods both theoretically and through extensive simulations. The results show a consistent and more efficient estimate, even with missing observations or uncertain treatment randomization probabilities. Finally, the practical utility of the proposed methods is demonstrated by analyzing data from a multiinstitution cohort of first-year medical residents in the United States (NeCamp et al., 2020).

研究动机与目标

  • 为解决由于高维历史数据中预设的手动设计特征导致的因果激发效应估计中的模型误设问题。
  • 通过用机器学习替代参数化工作模型来估计干扰参数,以提升估计效率和稳健性。
  • 基于元学习原则和去偏估计,开发一种双重稳健、渐近正态的因果激发效应估计器。
  • 实现在移动健康干预中对时变治疗调节进行数据驱动、灵活建模。

提出的方法

  • 该方法采用元学习框架,将干扰参数(如结果和治疗倾向性模型)的估计视为监督学习问题。
  • 使用Neyman正交 estimating equations 和交叉拟合,以减少机器学习算法带来的正则化偏差。
  • 核心估计器基于加权中心化最小二乘法(WCLS)准则,干扰成分通过灵活的机器学习模型(如随机森林、BART、神经网络)进行估计。
  • 该方法确保双重稳健性:若结果回归模型或倾向性得分模型中任一正确设定,最终的因果效应估计器即具一致性。
  • 通过去偏/正交估计方法,确保在高维、时变协变量下具有渐近正态性和有效的推断。
  • 通过适当的建模调整,将该框架扩展至处理缺失数据、滞后效应和二值结果。
Figure 1: Causal effects estimates with confidence intervals of R-WCLS ( left ) and DR-WCLS ( middle ), and their relative efficiency in comparisons with WCLS ( right ).
Figure 1: Causal effects estimates with confidence intervals of R-WCLS ( left ) and DR-WCLS ( middle ), and their relative efficiency in comparisons with WCLS ( right ).

实验结果

研究问题

  • RQ1机器学习能否在不引入偏差的情况下有效估计因果激发效应中的干扰参数?
  • RQ2与使用预设特征的传统WCLS相比,所提出的元学习方法在估计效率方面有何提升?
  • RQ3当结果模型或治疗模型存在模型误设时,该方法是否仍能实现双重稳健性?
  • RQ4使用灵活的机器学习模型对时变因果效应的置信区间精度和覆盖概率有何影响?
  • RQ5在具有复杂时变混杂因素的真实mHealth数据中,该方法表现如何?

主要发现

  • 所提出的双重稳健WCLS估计器在模拟和真实数据中均显著优于标准WCLS,置信区间更窄,效率提升明显。
  • 当其中一个干扰模型被误设时,DR-WCLS方法仍表现出更优的覆盖概率和更低的偏差,验证了双重稳健性。
  • 在第一年住院医师的案例研究中,移动提示在早期周次对步数具有积极因果效应,但随时间推移效应减弱——表明存在习惯化现象。
  • 该方法成功识别出时变调节效应,表明干预效果取决于先前行为、情绪和情境因素。
  • 模拟结果验证了估计器的渐近正态性,并在高维、时变协变量下证实了理论性质的成立。
  • R-WCLS和DR-WCLS方法在精度和统计功效方面均优于标准WCLS,尤其在存在复杂非线性关系的情境下表现更优。
Figure 2: The decision tree used to generate $g(H_{t})$ , where $\{X_{1,t},\dots,X_{4,t}\}\subset H_{t}$
Figure 2: The decision tree used to generate $g(H_{t})$ , where $\{X_{1,t},\dots,X_{4,t}\}\subset H_{t}$

更好的研究,从现在开始

从阅读论文到最终审阅,大幅缩短您的研究时间。

无需绑定信用卡

本解读由 AI 生成,并经人工编辑审核。