Salvato in:
Dettagli Bibliografici
Autori principali: Singh, Prabhjot, Gupta, Abhishek, Betz, Chris, Flansburg, Abe, Ives, Brett, Lama, Sudeep, Son, Jung Hoon
Natura: Preprint
Pubblicazione: 2026
Soggetti:
Accesso online:https://arxiv.org/abs/2605.09818
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866917479575453696
author Singh, Prabhjot
Gupta, Abhishek
Betz, Chris
Flansburg, Abe
Ives, Brett
Lama, Sudeep
Son, Jung Hoon
author_facet Singh, Prabhjot
Gupta, Abhishek
Betz, Chris
Flansburg, Abe
Ives, Brett
Lama, Sudeep
Son, Jung Hoon
contents Reinforcement learning (RL) in healthcare has had mixed results, with reward sparsity, unreliable off-policy evaluation, and deployment-simulation gap as recurring failure modes. We argue that chronic disease management is structurally a more tractable RL setting than the acute-care problems the field has primarily studied, but only if the problem is formalized to exploit chronic care's properties. We propose such a formalization. The agent's objective is to compress time-to-control (TTC) under a tiered reward calibrated to the CMS ACCESS Model. Two quantities from our companion preference-learning paper [Singh et al. 2026] enter as load-bearing structural elements: the execution intensity εbounds action availability under a constrained Markov Decision Process, and the clinician capability κweights offline-data transitions during RL training. Together they couple preference learning and RL into a two-loop architecture. We present simulation results on synthetic state machines for hypertension and type 2 diabetes. Capability-weighted offline RL outperforms uniform-weighted offline RL and the behavior policy by 15 percentage points on T2D TTC; the uniform-weighted formulation (the standard in existing healthcare RL) underperforms even the heterogeneous behavior policy. \Epsilon-aware policies generalize across deployment regimes while ε-naive policies do not.
format Preprint
id arxiv_https___arxiv_org_abs_2605_09818
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Learning to Compress Time-to-Control: A Reinforcement Learning Framework for Chronic Disease Management
Singh, Prabhjot
Gupta, Abhishek
Betz, Chris
Flansburg, Abe
Ives, Brett
Lama, Sudeep
Son, Jung Hoon
Machine Learning
Reinforcement learning (RL) in healthcare has had mixed results, with reward sparsity, unreliable off-policy evaluation, and deployment-simulation gap as recurring failure modes. We argue that chronic disease management is structurally a more tractable RL setting than the acute-care problems the field has primarily studied, but only if the problem is formalized to exploit chronic care's properties. We propose such a formalization. The agent's objective is to compress time-to-control (TTC) under a tiered reward calibrated to the CMS ACCESS Model. Two quantities from our companion preference-learning paper [Singh et al. 2026] enter as load-bearing structural elements: the execution intensity εbounds action availability under a constrained Markov Decision Process, and the clinician capability κweights offline-data transitions during RL training. Together they couple preference learning and RL into a two-loop architecture. We present simulation results on synthetic state machines for hypertension and type 2 diabetes. Capability-weighted offline RL outperforms uniform-weighted offline RL and the behavior policy by 15 percentage points on T2D TTC; the uniform-weighted formulation (the standard in existing healthcare RL) underperforms even the heterogeneous behavior policy. \Epsilon-aware policies generalize across deployment regimes while ε-naive policies do not.
title Learning to Compress Time-to-Control: A Reinforcement Learning Framework for Chronic Disease Management
topic Machine Learning
url https://arxiv.org/abs/2605.09818