在缺少结果数据的情况下,对异质治疗效果的因果机器学习
Matthew Pryce1, Karla Diaz-Ordaz2, Ruth H Keogh1
1Department of Medical Statistics, London School of Hygiene & Tropical Medicine, London WC1E 7HT, United Kingdom.
Biometrics
|July 31, 2025
概括
缺少结果数据可能会影响治疗效果估计. 新的方法,mDR-learner和mEP-learner,使用审查权重的逆概率来改善因果机器学习以估计条件平均治疗效应 (CATE).
科学领域:
- 因果推理的原因推理.
- 机器学习 机器学习
- 生物统计学 生物统计学
背景情况:
- 缺少结果数据是因果推理中的一个常见问题.
- 它可能导致子组的代表性不足和对异质治疗效果的偏见估计.
- 对于条件平均治疗效果 (CATE) 的现有的因果机器学习估计器可能对缺失的数据敏感.
研究的目的:
- 解决异质治疗效果估计中缺少结果数据的挑战.
- 为CATE提出新的无偏见的机器学习估计器,以解释缺失.
- 与现有方法相比,评估拟议估计器的性能.
主要方法:
- 提出了两个新的无偏差机器学习估计器:mDR-learner和mEP-learner.
- 在现有的DR-learner和EP-learner框架中整合审查权重的逆概率.
- 使用模拟数据和真实世界的临床试验 (GBSG2) 评估估计器.
主要成果:
- 建议的mDR-learner和mEP-learner估计器在合理的条件下被证明是Oracle有效的.
- 与现有的CATE估计器相比,模拟显示了新估计器的良好表现,包括使用标准缺失数据技术的估计器.
- 估计器有效地解决了由于缺少结果数据而导致的子组代表性不足问题.
结论:
- 在缺乏结果数据的情况下,mDR-learner和mEP-learner提供了可靠的方法来估计条件平均治疗效果.
- 这些新型估计器通过减轻由于代表性不足的子组造成的偏差来改进现有方法.
- 该研究为在现实应用中实施这些估计器提供了实际指导,例如分析乳腺癌治疗效果.
更多相关视频
12:18A Machine Learning Approach to Design an Efficient Selective Screening of Mild Cognitive Impairment
Published on: January 11, 2020
7.6K
06:55Inverse Probability of Treatment Weighting Propensity Score using the Military Health System Data Repository and National Death Index
Published on: January 8, 2020
14.6K
相关概念视频
01:23Mechanistic Models: Compartment Models in Individual and Population Analysis
87
Mechanistic models are utilized in individual analysis using single-source data, but imperfections arise due to data collection errors, preventing perfect prediction of observed data. The mathematical equation involves known values (Xi), observed concentrations (Ci), measurement errors (εi), model parameters (ϕj), and the related function (ƒi) for i number of values. Different least-squares metrics quantify differences between predicted and observed values. The ordinary least...
87
01:20Comparing the Survival Analysis of Two or More Groups
289
Survival analysis is a cornerstone of medical research, used to evaluate the time until an event of interest occurs, such as death, disease recurrence, or recovery. Unlike standard statistical methods, survival analysis is particularly adept at handling censored data—instances where the event has not occurred for some participants by the end of the study or remains unobserved. To address these unique challenges, specialized techniques like the Kaplan-Meier estimator, log-rank test, and...
289
01:09Censoring Survival Data
241
Survival analysis is a statistical method used to analyze time-to-event data, often employed in fields such as medicine, engineering, and social sciences. One of the key challenges in survival analysis is dealing with incomplete data, a phenomenon known as "censoring." Censoring occurs when the event of interest (such as death, relapse, or system failure) has not occurred for some individuals by the end of the study period or is otherwise unobservable, and it might have many different...
241
01:21Causality in Epidemiology
854
Causality or causation is a fundamental concept in epidemiology, vital for understanding the relationships between various factors and health outcomes. Despite its importance, there's no single, universally accepted definition of causality within the discipline. Drawing from a systematic review, causality in epidemiology encompasses several definitions, including production, necessary and sufficient, sufficient-component, counterfactual, and probabilistic models. Each has its strengths and...
854
01:16Crossover Experiments
3.0K
Crossover experiments, also called the repeated-measurements design, is a study design in which all experimental units are exposed to all treatments in different periods. Crossover experiments are generally used in psychology, the pharmaceutical industry, agriculture, and medicine.
Crossover designs are performed even with smaller sample sizes since the samples can act as their controls. These are better than simple randomized trials since patients are exposed to all the treatments.
Crossover designs are performed even with smaller sample sizes since the samples can act as their controls. These are better than simple randomized trials since patients are exposed to all the treatments.
3.0K
01:24Kaplan-Meier Approach
270
The Kaplan-Meier estimator is a non-parametric method used to estimate the survival function from time-to-event data. In medical research, it is frequently employed to measure the proportion of patients surviving for a certain period after treatment. This estimator is fundamental in analyzing time-to-event data, making it indispensable in clinical trials, epidemiological studies, and reliability engineering. By estimating survival probabilities, researchers can evaluate treatment effectiveness,...
270