Personalization of Large Foundation Models for Health Interventions

TL;DR

Large foundation models (LFMs) enable broad, data-driven healthcare insights but face personalization and external validity challenges. The authors propose a hybrid framework that uses population-trained LFMs for rapid hypothesis generation and N-of-1 trials for within-person causal validation, with Bayesian updating to form a personalized digital twin while preserving privacy. They identify four structural tensions—privacy-performance, scale-specificity, personalization-external validity, and automation-empathy—and argue that selective, interpretable, and patient-centered experimentation can reconcile these tensions. An illustrative chronic migraine case demonstrates the practical workflow, emphasizing safety, regulatory, and equity considerations and suggesting broad applicability beyond healthcare to any domain needing personalized AI systems.

Abstract

Large foundation models (LFMs) transform healthcare AI in prevention, diagnostics, and treatment. However, whether LFMs can provide truly personalized treatment recommendations remains an open question. Recent research has revealed multiple challenges for personalization, including the fundamental generalizability paradox: models achieving high accuracy in one clinical study perform at chance level in others, demonstrating that personalization and external validity exist in tension. This exemplifies broader contradictions in AI-driven healthcare: the privacy-performance paradox, scale-specificity paradox, and the automation-empathy paradox. As another challenge, the degree of causal understanding required for personalized recommendations, as opposed to mere predictive capacities of LFMs, remains an open question. N-of-1 trials -- crossover self-experiments and the gold standard for individual causal inference in personalized medicine -- resolve these tensions by providing within-person causal evidence while preserving privacy through local experimentation. Despite their impressive capabilities, this paper argues that LFMs cannot replace N-of-1 trials. We argue that LFMs and N-of-1 trials are complementary: LFMs excel at rapid hypothesis generation from population patterns using multimodal data, while N-of-1 trials excel at causal validation for a given individual. We propose a hybrid framework that combines the strengths of both to enable personalization and navigate the identified paradoxes: LFMs generate ranked intervention candidates with uncertainty estimates, which trigger subsequent N-of-1 trials. Clarifying the boundary between prediction and causation and explicitly addressing the paradoxical tensions are essential for responsible AI integration in personalized medicine.

Figures (2)

• Figure 1: Illustration of an N-of-1 trial design with alternating interventions (A and B) following a baseline period to evaluate individual treatment responses over time.
• Figure 2: Visual illustration of the hybrid framework combining population-level LFMs with individual-level N-of-1 trials for evidence-based personalization of health interventions.