Optimal individualized decision rules from a multi-arm trial: A comparison of methods and an application to tailoring inter-donation intervals among blood donors in the UK.

Yuejia Xu,Brian Dm Tom,Michael J Sweeting,Angela M Wood,David J Roberts

doi:10.1177/0962280220920669

Abstract

There is a growing interest in precision medicine where individual heterogeneity is incorporated into decision-making and treatments are tailored to individuals to provide better healthcare. One important aspect of precision medicine is the estimation of the optimal individualized treatment rule (ITR) that optimizes the expected outcome. Most methods developed for this purpose are restricted to the setting with two treatments, while clinical studies with more than two treatments are common in practice. In this work, we summarize methods to estimate the optimal ITR in the multi-arm setting and compare their performance in large-scale clinical trials via simulation studies. We then illustrate their utilities with a case study using the data from the INTERVAL trial, which randomly assigned over 20,000 male blood donors from England to one of the three inter-donation intervals (12-week, 10-week, and eight-week) over two years. We estimate the optimal individualized donation strategies under three different objectives. Our findings are fairly consistent across five different approaches that are applied: when we target the maximization of the total units of blood collected, almost all donors are assigned to the eight-week inter-donation interval, whereas if we aim at minimizing the low hemoglobin deferral rates, almost all donors are assigned to donate every 12 weeks. However, when the goal is to maximize the utility score that “discounts” the total units of blood collected by the incidences of low hemoglobin deferrals, we observe some heterogeneity in the optimal inter-donation interval across donors and the optimal donor assignment strategy is highly dependent on the trade-off parameter in the utility function.

Highlights

Precision medicine is a rapidly expanding field in the new era of healthcare with major advancements having been made in technologies for collecting patient-level data and better characterizing each individual patient
We describe several methodological options that can be used to identify the optimal individualized treatment rule (ITR) in clinical trials with more than two treatment arms and are computationally feasible for large-scale trials, including l1-penalized least squares,[7] adaptive contrast weighted learning,[8] direct learning,[9] and a Bayesian approach that is based on Bayesian additive regression trees.[10]
We review a selection of approaches that can be used to estimate the optimal ITR in multi-arm trials and scale well for large datasets similar in size to the INTERVAL trial, including l1-penalized least squares,[7] adaptive contrast weighted learning,[8] direct learning,[9] and a Bayesian approach that is based on Bayesian additive regression trees.[10]

Summary

Introduction

Precision medicine is a rapidly expanding field in the new era of healthcare with major advancements having been made in technologies for collecting patient-level data and better characterizing each individual patient. The goal of Statistical Methods in Medical Research 29(11). Precision medicine is to improve patient outcomes by tailoring treatment selection based on observed patient characteristics, for example, demographic information, clinical and laboratory measurements, medical history, and genetic data. The treatment that is regarded as the best for a patient with one set of characteristics might not be the best for another, and the traditional “one-size-fits-all” approach does not lead to optimal clinical outcomes in many cases. In light of this, individualized, evidence-based clinical decisionmaking strategies that account for such heterogeneity are considered more desirable and are gaining much popularity in medical research. One important aspect of precision medicine is the estimation of the optimal individualized treatment rule (ITR) by mapping patient information onto the set of treatment options. We refer readers to Lipkovich et al.[1] and Kosorok and Laber[2] for a comprehensive review of existing methods

Objectives

Methods

Results

Conclusion