Optimizing Clinical Trial Protocols Using EHR-Derived Heterogeneous Treatment Effects

2026-07-18

AI Summary

Traditional randomized controlled trials estimate only the average treatment effect, often overlooking clinically meaningful heterogeneity in treatment response. This study leverages Mayo Clinic’s cloud-based electronic health records to emulate the DAPA-HF trial, integrating a Meta-S learner with a Cox proportional hazards model to estimate heterogeneous treatment effects. Innovatively, a decision tree–based thresholding approach is introduced for patient stratification. Although the overall emulated cohort showed no significant treatment benefit (HR = 1.681, p = 0.1507), the method successfully identified a subgroup with substantial benefit (HR = 0.203, p = 0.0002) and another at markedly elevated risk of harm (HR = 6.680, p < 0.0001). This work represents the first demonstration in real-world data of bidirectional effect discovery—simultaneously revealing beneficial and harmful subgroups—driven by heterogeneous treatment effects, offering a novel paradigm for optimizing clinical trial design and enabling precision medicine.

Abstract

Traditional randomized trials often obscure clinically meaningful heterogeneity in treatment response by focusing on average effects. Leveraging real-world data to emulate clinical trials and estimate heterogeneous treatment effects (HTEs) offers a promising path toward more precise and efficient trial design. In this study, we emulate the DAPA-HF trial using electronic health records from the Mayo Clinic Cloud (MCC) to investigate whether HTE-guided stratification can identify patient subgroups with distinct treatment responses to dapagliflozin versus placebo in patients with heart failure with reduced ejection fraction. All-cause mortality was evaluated using Cox proportional hazards models, with HTEs estimated using a Meta-S learner and subgroups defined using a decision tree-based thresholding approach. In the overall cohort of the emulation, no significant treatment difference was observed (HR, 1.681; 95% CI, 0.828-3.413; p = 0.1507). However, compared with the overall emulated cohort, in which dapagliflozin showed no statistically significant survival benefit, HTE-driven stratification identified subgroups with significant and directionally distinct treatment effects. The beneficial (low-HTE) subgroup showed a significant survival benefit from dapagliflozin (HR = 0.203, 95% CI, 0.087-0.476, p = 0.0002), whereas the harmful (high-HTE) subgroup showed a significant harmful association with markedly increased mortality risk (HR = 6.680, 95% CI, 2.759-16.171, p < 0.0001). These findings indicate that HTE-guided stratification can uncover clinically meaningful beneficial and harmful treatment-effect patterns that are masked in the full-cohort emulation.

Problem

• heterogeneous treatment effects
• clinical trial optimization
• treatment response heterogeneity
• patient stratification
• real-world data

Innovation

• heterogeneous treatment effects
• clinical trial emulation
• real-world data
• Meta-S learner
• treatment-effect stratification