PreprintCatalog paper date: 2026Added to atlas: Nov 5, 2026Source
Activity counts measure attention, not model quality. Missing counts are unknown. Publication status follows the linked source; a listed venue alone does not confirm peer review.
FITTER is a multi-agent LLM framework that writes guideline-grounded FITT-VP exercise prescriptions from linked EHR and Fitbit data. Deterministic audits correct rule-checkable safety violations. Evaluated on 400 All of Us participants (300 cancer, 100 stroke) against five LLM comparators.
Parameters
9B
Signals
HR, Activity, EHR, Text
Form factor
Wrist
Year
2026
Architecture
Multi-agent LLM pipeline on an open-weight Qwen3.5-9B backbone with deterministic guideline/safety auditing; outputs FITT-VP (frequency, intensity, time, type, volume, progression) prescriptions
Pretraining data
No pretraining; prompting/agentic pipeline over All of Us EHR + Fitbit data
License
arXiv perpetual non-exclusive license; code repository not released
Translating population-level exercise guidance into individualized prescriptions requires integrating clinical and wearable data, including diagnoses, medications, functional limitations, and baseline activity, with condition-specific recommendations, a process difficult to scale beyond specialist rehabilitation settings. We developed FITTER, a multi-agent framework that integrates these data with condition-specific guidance to generate personalized FITT-VP prescriptions and applies deterministic safety audit and repair to correct rule-checkable violations. We evaluated FITTER using data from 400 All of Us participants (300 cancer, 100 stroke) against five LLM comparators through computational evaluation, ablation, and structured clinician review. FITTER achieved 98.5% safety-constraint compliance versus 75.8-85.8% for comparators and showed greater quantitative dosing precision and guideline alignment. Cancer-cohort ablation identified safety audit and repair as the largest contributor to safety-constraint compliance. Clinicians reviewed 80 prescriptions across safety, personalization, instruction concreteness, anticipated effectiveness, and clinical acceptability. FITTER provides a clinician-facing approach integrating clinical and wearable data, guidance, and explicit safety constraints.
Verbatim fromarXiv· arXiv metadata (abstracts are released under CC0)
How it works
Pretraining data
No pretraining; prompting/agentic pipeline over All of Us EHR + Fitbit data
Architecture
Multi-agent LLM pipeline on an open-weight Qwen3.5-9B backbone with deterministic guideline/safety auditing; outputs FITT-VP (frequency, intensity, time, type, volume, progression) prescriptions
Inputs
HR, Activity, EHR, Text
Approach
LLM agent · multi agent · exercise prescription · FITT VP · Fitbit · All of Us
Cite the publication recorded for this model. Check it against the publisher before submitting.
Li, X., Lin, Y., Wu, Y., Wang, F., Wang, M., Gary, R. A., Kariuki, J., Smith, J., Billinger, S. A., Xiao, C., Hu, X., & Yan, R. (2026). A Guideline-Grounded Multi-Agent Framework for Personalized Exercise Prescription from Clinical and Wearable Data (FITTER; arXiv 2611.03983). arXiv 2611.03983v1. https://arxiv.org/abs/2611.03983
Li, Xintong, Yufen Lin, Yuhua Wu, Fengran Wang, Minxiao Wang, Rebecca A. Gary, and Jacob Kariuki, et al. 2026. "A Guideline-Grounded Multi-Agent Framework for Personalized Exercise Prescription from Clinical and Wearable Data (FITTER; arXiv 2611.03983)." arXiv 2611.03983v1. https://arxiv.org/abs/2611.03983.
Li, X., Lin, Y., Wu, Y., Wang, F., Wang, M., Gary, R. A. and Kariuki, J. et al. (2026) 'A Guideline-Grounded Multi-Agent Framework for Personalized Exercise Prescription from Clinical and Wearable Data (FITTER; arXiv 2611.03983)', arXiv 2611.03983v1. Available at: https://arxiv.org/abs/2611.03983
Li, Xintong, et al. "A Guideline-Grounded Multi-Agent Framework for Personalized Exercise Prescription from Clinical and Wearable Data (FITTER; arXiv 2611.03983)." arXiv 2611.03983v1, 2026. https://arxiv.org/abs/2611.03983.
X. Li et al., "A Guideline-Grounded Multi-Agent Framework for Personalized Exercise Prescription from Clinical and Wearable Data (FITTER; arXiv 2611.03983)," arXiv 2611.03983v1, 2026, [Online]. Available: https://arxiv.org/abs/2611.03983.
@misc{li2026guidelinegrounded,
title = {{A Guideline-Grounded Multi-Agent Framework for Personalized Exercise Prescription from Clinical and Wearable Data (FITTER; arXiv 2611.03983)}},
author = {Li, Xintong and Lin, Yufen and Wu, Yuhua and Wang, Fengran and Wang, Minxiao and Gary, Rebecca A. and Kariuki, Jacob and Smith, James and Billinger, Sandra A. and Xiao, Canhua and Hu, Xiao and Yan, Runze},
year = {2026},
howpublished = {arXiv 2611.03983v1},
eprint = {2611.03983},
archivePrefix = {arXiv}
}
History
Added to the ChatPPG research hub
Newsletter
Get new PPG datasets, papers and models
Get alerts when FITTER is benchmarked on new datasets, plus the weekly digest.
New datasets, papers and models
Practical guides from the hub
Unsubscribe anytime
The waveform is an illustration of a PPG pulse, not a recording.
