🤖 AI Summary
This work addresses the challenge of effectively integrating population-level aggregate evidence—such as clinical trial statistics—with individual patient data to enhance the reliability of predictive models in clinical decision-making. The authors propose FRESH, a method that operates within an information-geometric framework to apply minimal KL-divergence perturbations to generative models trained on individual-level data, ensuring their outputs exactly match specified aggregate statistics of a target population. FRESH achieves the first unbiased, data-efficient post-training calibration with respect to external summary evidence while preserving the original distributional structure. The approach supports contextual comparisons between single-arm trials and standard-of-care benchmarks, facilitates clinical trial simulation, and enables comparative effectiveness analyses, substantially improving model applicability and generalization in real-world clinical settings.
📝 Abstract
This note introduces FRESH (Fusion of Recent Evidence and Subject Histories), a method for incorporating population-level summary results -- published clinical trials, registry summaries, prior natural-history studies, and peer-reviewed indirect comparisons -- into predictive models trained on patient-level data. This method provides a principled means of combining both patient-level and aggregate-level data types into a unified data-efficient model for clinical decision making. FRESH assumes access to a generative model trained on patient-level data sources (e.g. clinical trial or real-world data). The method produces patient-level predictions from a re-calibrated model that matches a set of specified aggregate statistics for a target population. This can be understood as a patient-level recapitulation of the aggregate source -- with the key property that the recalibration is a minimal perturbation of the original joint distribution in a specific information-geometric sense. The resulting samples can be analyzed directly or combined into a post-training procedure to update the original generative model. This approach enables several applications where rigorously incorporating patient-level data with summary information is valuable, including (i) contextualizing single-arm trial results with respect to recent standard-of-care, (ii) clinical-trial simulations for design and probability-of-technical-success estimation, and (iii) comparative-effectiveness analyses of on-market therapies.