Scalable and Distributed Individualized Treatment Rules for Massive Datasets

2025-11-08

AI Summary

Addressing the challenge of balancing privacy preservation and statistical efficiency in constructing individualized treatment rules (ITRs) from multi-center medical data, this paper proposes a scalable distributed learning framework. Methodologically, we design a convex, smooth loss function via convolutional smoothing and integrate it with a weighted support vector machine optimized via coordinate gradient descent; collaboration across centers relies solely on shared summary statistics—eliminating raw data transmission. The algorithm enjoys linear convergence guarantees with a fixed number of communication rounds. Our key innovation is the first application of convolutional smoothing to distributed ITR estimation, which eliminates local estimation bias while preserving privacy. Experiments on multi-center ICU sepsis data demonstrate substantial improvements in decision accuracy and out-of-sample generalizability, alongside computational efficiency and clinical applicability.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Search and Optimization: Distributed Search
• Constraint Satisfaction and Optimization: Distributed CSP/Optimization

Application Category

• Security and Privacy: Large-scale security measurements
• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems
• Responsible Web: Data and user privacy-enhancing technologies for the Web

Abstract

Synthesizing information from multiple data sources is crucial for constructing accurate individualized treatment rules (ITRs). However, privacy concerns often present significant barriers to the integrative analysis of such multi-source data. Classical meta-learning, which averages local estimates to derive the final ITR, is frequently suboptimal due to biases in these local estimates. To address these challenges, we propose a convolution-smoothed weighted support vector machine for learning the optimal ITR. The accompanying loss function is both convex and smooth, which allows us to develop an efficient multi-round distributed learning procedure for ITRs. Such distributed learning ensures optimal statistical performance with a fixed number of communication rounds, thereby minimizing coordination costs across data centers while preserving data privacy. Our method avoids pooling subject-level raw data and instead requires only sharing summary statistics. Additionally, we develop an efficient coordinate gradient descent algorithm, which guarantees at least linear convergence for the resulting optimization problem. Extensive simulations and an application to sepsis treatment across multiple intensive care units validate the effectiveness of the proposed method.

Problem

• Addressing privacy barriers in multi-source data analysis for treatment rules
• Overcoming suboptimal performance of classical meta-learning for ITRs
• Developing distributed learning methods that preserve data privacy

Innovation

• Convolution-smoothed weighted SVM for optimal ITRs
• Multi-round distributed learning with summary statistics
• Coordinate gradient descent ensuring linear convergence