Causal machine learning for predicting treatment outcomes (2024-04-01T00:00:00.000000Z)

TL;DR

The benefits of causal ML (relative to traditional statistical or ML approaches) are discussed and the key components and steps are outlined and recommended, offering practical guidance for appropriate clinical use.

Abstract

Causal machine learning (ML) offers flexible, data-driven methods for predicting treatment outcomes including efficacy and toxicity, thereby supporting the assessment and safety of drugs. A key benefit of causal ML is that it allows for estimating individualized treatment effects, so that clinical decision-making can be personalized to individual patient profiles. Causal ML can be used in combination with both clinical trial data and real-world data, such as clinical registries and electronic health records, but caution is needed to avoid biased or incorrect predictions. In this Perspective, we discuss the benefits of causal ML (relative to traditional statistical or ML approaches) and outline the key components and steps. Finally, we provide recommendations for the reliable use of causal ML and effective translation into the clinic. Causal machine learning methods could be used to predict treatment outcomes for subgroups and even individual patients; this Perspective outlines the potential benefits and limitations of the approach, offering practical guidance for appropriate clinical use.

Authors

Stefan Feuerriegel

1 papers

Dennis Frauen

1 papers

Valentyn Melnychuk

2 papers

TL;DR

Abstract

Authors

References102 items

Bayesian Neural Controlled Differential Equations for Treatment Effect Estimation

Estimated Average Treatment Effect of Psychiatric Hospitalization in Patients With Suicidal Behaviors: A Precision Treatment Analysis.

Clinical AI tools must convey predictive uncertainty for each individual patient

Conformal Meta-learners for Predictive Inference of Individual Treatment Effects

Comparative effectiveness of second line glucose lowering drug treatments using real world data: emulation of a target trial

Comparative effectiveness of SGLT2 inhibitors, GLP-1 receptor agonists, DPP-4 inhibitors, and sulfonylureas on risk of major adverse cardiovascular events: emulation of a randomised target trial using electronic health records.

Accounting For Informative Sampling When Learning to Forecast Treatment Outcomes Over Time

Reliable Off-Policy Learning for Dosage Combinations

Sharp Bounds for Generalized Causal Sensitivity Analysis

A causal roadmap for generating high-quality real-world evidence

B-Learner: Quasi-Oracle Bounds on Heterogeneous Causal Effects Under Hidden Confounding

Prediction Under Interventions: Evaluation of Counterfactual Performance Using Longitudinal Observational Data

In Search of Insights, Not Magic Bullets: Towards Demystification of the Model Selection Dilemma in Heterogeneous Treatment Effect Estimation

Emulating a target trial using primary care electronic health records: SGLT-2i medications and Hemoglobin A1c.

Normalizing Flows for Interventional Density Estimation

Estimating treatment effects for time-to-treatment antibiotic stewardship in sepsis

Estimating individual treatment effects under unobserved confounding using binary instruments

Machine Learning for Predicting the Risk of Transition from Prediabetes to Diabetes

Causal Machine Learning: A Survey and Open Problems

Continuous-Time Modeling of Counterfactual Outcomes Using Neural Controlled Differential Equations

Protocol for the development of a reporting guideline for causal and counterfactual prediction models in biomedicine

Multi-disciplinary fairness considerations in machine learning for clinical trials

Causal Transformer for Estimating Counterfactual Outcomes

Estimating average causal effects from patient trajectories

Combining Observational and Randomized Data for Estimating Heterogeneous Treatment Effects

Generalizing Off-Policy Learning under Sample Selection Bias

Sensitivity analysis of individual treatment effects: A robust conformal inference approach

DoWhy: Addressing Challenges in Expressing and Validating Causal Assumptions

Mitigating bias in machine learning for medicine

When Effects Cannot be Estimated: Redefining Estimands to Understand the Effects of Naloxone Access Laws.

DoubleML - An Object-Oriented Implementation of Double Machine Learning in Python

VCNet and Functional Targeted Regularization For Learning Causal Effects of Continuous Treatments

Sharp Sensitivity Analysis for Inverse Propensity Weighting via Quantile Balancing

Nonparametric Estimation of Heterogeneous Treatment Effects: From Theory to Learning Algorithms

Causal Inference Methods for Combining Randomized Trials and Observational Studies: A Review.

Machine learning prediction in cardiovascular diseases: a meta-analysis

A Crash Course in Good and Bad Controls

Minimum information about clinical artificial intelligence modeling: the MI-CLAIM checklist

Estimating Individual Treatment Effects with Time-Varying Confounders

Early detection of type 2 diabetes mellitus using machine learning-based prediction models

Identifying Causal Effect Inference Failure with Uncertainty-Aware Models

Towards optimal doubly robust estimation of heterogeneous causal effects

Prediction meets causal inference: the role of treatment in clinical prediction models

G-Net: A Deep Learning Approach to G-computation for Counterfactual Outcome Prediction Under Dynamic Treatment Regimes

Identification of the Average Treatment Effect When SUTVA Is Violated

Estimates of Overall Survival in Patients With Cancer Receiving Different Treatment Regimens

Estimating the Effects of Continuous-valued Interventions using Generative Adversarial Networks

Estimating Counterfactual Treatment Outcomes over Time Through Adversarially Balanced Representations

Prediction of progression from pre‐diabetes to diabetes: Development and validation of a machine learning model

A Framework for Separating Individual-Level Treatment Effects From Spillover Effects

Adapting Neural Networks for the Estimation of Treatment Effects

Machine Learning Estimation of Heterogeneous Treatment Effects with Instruments

Cardiovascular disease risk prediction using automated machine learning: A prospective study of 423,604 UK Biobank participants

Learning Counterfactual Representations for Estimating Individual Dose-Response Curves

Real-world data: towards achieving the achievable in cancer care

Orthogonal Statistical Learning

A guide to deep learning in healthcare

Removing Hidden Confounding by Experimental Grounding

Interval Estimation of Individual-Level Causal Effects Under Unobserved Confounding

Quasi-oracle estimation of heterogeneous treatment effects

Metalearners for estimating heterogeneous treatment effects using machine learning

Double/Debiased Machine Learning for Treatment and Structural Parameters

Bayesian Inference of Individualized Treatment Effects using Multi-task Gaussian Processes

Reliable Decision Support using Counterfactual Models

Real-World Evidence - What Is It and What Can It Tell Us?

Generalized random forests

Estimating individual treatment effect: generalization bounds and algorithms

Super-Learning of an Optimal Dynamic Treatment Rule

Using Big Data to Emulate a Target Trial When a Randomized Trial Is Not Available.

Racial and Ethnic Disparities in Health Care Access and Utilization Under the Affordable Care Act

Estimation and Inference of Heterogeneous Treatment Effects using Random Forests

Non‐parametric methods for doubly robust estimation of continuous treatment effects

Recursive partitioning for heterogeneous causal effects

Modern radiation therapy for Hodgkin lymphoma: field and dose guidelines from the international lymphoma radiation oncology group (ILROG).

Causal Models and Learning from Data: Integrating Causal Modeling and Statistical Estimation

Targeted Data Adaptive Estimation of the Causal Dose–Response Curve