Outcome risk model development for heterogeneity of treatment effect analyses: a comparison of non-parametric machine learning methods and semi-parametric statistical methods

• Published in: BMC medical research methodology Vol. 24; no. 1; pp. 158 - 9
• Main Authors: Xu, Edward, Vanghelof, Joseph, Wang, Yiyang, Patel, Anisha, Furst, Jacob, Raicu, Daniela Stan, Neumann, Johannes Tobias, Wolfe, Rory, Gao, Caroline X., McNeil, John J., Shah, Raj C., Tchoua, Roselyne
• Format: Journal Article
• Language: English
• Published: England BioMed Central Ltd 23.07.2024; BioMed Central; BMC
• Subjects: Accuracy; Aged; Aspirin; Aspirin - therapeutic use; Care and treatment; Clinical outcomes; Clinical trials; Comparative analysis; Datasets; Decision tree; Decision Trees; Dementia; Diagnosis; Disability-free longevity; Dosage and administration; Female; Health risk assessment; Heterogeneity of treatment effect; Humans; Machine Learning; Male; Methods; Models, Statistical; Mortality; Outcome Assessment, Health Care - methods; Outcome Assessment, Health Care - statistics & numerical data; Outcome risk modelling; Proportional Hazards Models; Random forest; Randomized Controlled Trials as Topic - methods; Randomized Controlled Trials as Topic - statistics & numerical data; Risk assessment; Risk Assessment - methods; Risk Assessment - statistics & numerical data; Treatment Outcome; United States; United States; Australia; United States > US; Aspirin; Disability-free longevity; Outcome risk modelling; Clinical trial; Random forest; Decision tree; Heterogeneity of treatment effect

In randomized clinical trials, treatment effects may vary, and this possibility is referred to as heterogeneity of treatment effect (HTE). One way to quantify HTE is to partition participants into subgroups based on individual's risk of experiencing an outcome, then measuring treatment effect b...