The paper introduces EpiCoach, an agentic system that treats a single athlete as the subject of accountable N‑of‑1 science. It uses a mechanistic digital twin built from wearable data to make probabilistic predictions about future training sessions, and an evidence‑gated hypothesis registry records each claim and its outcomes. Over eighteen months, the system’s predictions were evaluated both retrospectively and prospectively, showing competitive performance and demonstrating a framework for auditable, falsifiable personalization in digital health.
arXiv:2607. 21403v1 Announce Type: new Abstract: Mobile-health interventions increasingly use online learning and decision making algorithms to personalize when to nudge users toward healthier behavior, but a poorly designed algorithm can burden and disengage participants.
By Ziping Xu, Yuyi Chang, Chenshun Ni, Nithin Sugavanam, Asim H. Gazi, Pedja Klasnja, Emre Ertin, Susan A. Murphy
WearableQA is a new benchmark that tests AI systems on health reasoning using real-world wearable data from 200 users, each with up to 500 days of daily measurements. It contains 4,084 ten‑option multiple‑choice questions derived from wearable time series, blood biomarkers, and demographics, and is organized into 16 question types that distinguish data‑driven computation from physiological interpretation and single‑signal from cross‑signal reasoning. Evaluation of 14 large language models shows wide performance gaps, indicating that the benchmark remains challenging and useful for diagnosing model capabilities.
By Ji Soo Lee, Xilun Chen, Pierce Chuang, Ashish Shenoy, Jason Wei, Dohwan Ko, Hyunwoo J. Kim, Benoit Corda
arXiv:2607. 25340v1 Announce Type: new Abstract: The same episode of atrial fibrillation is a minor finding in a healthy adult and grounds for anticoagulation in an elderly patient with hypertension: identical signal, opposite decision.
By Sukju Oh, Moo-Yong Rhee, Jae-Sik Jang, Sukkyu Sun
The paper introduces an agent architecture that uses tools and memory to selectively correct wrist photoplethysmography (PPG) heart‑rate estimates. By generating candidate corrections from pulse, motion, and historical data, and evaluating their predicted benefits and harms, the system decides whether to retain, correct, or reject each estimate. Across four public datasets, the approach reduced mean absolute error by up to 44.5% and demonstrated real‑time deployment on a wearable‑smartphone platform.
By Xu Wang, Shuhan Yi, Zhen Cheng
The paper introduces a compact patient world model that forecasts digital health campaign outcomes by maintaining a latent state per patient and learning exposure‑conditioned dynamics. Evaluated on a large US campaign dataset, the model predicts new‑to‑brand prescription volume with low relative error (2.9% at week‑4 cutoff) compared to much higher errors from baseline classifiers. The study also shows that dense next‑exposure supervision is crucial for accurate forecasts when conversions are rare and highlights limitations in interpreting exposure‑conditioned rollouts causally.
By Yunlong Wang
Wrist photoplethysmography (PPG) heart rate estimates are vulnerable to motion artifacts, and corrections that reduce mean error can worsen individual windows. We propose an agent architecture with to...
The study examines how temporal resampling (binning) of irregular clinical data affects the safety and performance of offline reinforcement learning (ORL) models in healthcare. By analyzing MIMIC‑IV insulin infusion records, the authors show that binning can reverse causal order and obscure critical events. In simulated trials with virtual type‑1 diabetes patients, ORL models trained on binned data performed up to 60% worse than those trained on unprocessed data, yet retrospective evaluation on resampled data overestimated their returns.
By Thomas Frost, Hrisheekesh Vaidya, Steve Harris
BALMS is a benchmark for evaluating large language model (LLM) agents that analyze longitudinal wearable data to predict mental‑health wellbeing scores and generate evidence‑grounded rationales. It covers three real‑world datasets, two task families (score prediction and rationale generation), and tests five LLM backbones across open‑ and closed‑source paradigms. The study finds that zero‑shot agents rarely beat a simple mean baseline, and while chain‑of‑thought prompting helps reasoning, it does not ensure temporal grounding or numerical accuracy.
By Yu Yvonne Wu, Arvind Pillai, Yuliang Chen, Yuwei Zhang, Sudarshan Regmi, Tess Z. Griffin, Michael V. Heinz, Lisa A. Marsch, Nicholas C. Jacobson, Andrew Campbell
arXiv:2606. 14604v1 Announce Type: cross Abstract: Wearable devices and smartphones generate rich behavioural time series that can support proactive health interventions, yet systematic comparisons of modern forecasting architectures for these data are lacking.
By Pavlos Nicolaou, Kleanthis Malialis, Artemis Kontou, Panayiotis Kolios
arXiv:2607. 13940v1 Announce Type: new Abstract: Personal health management unfolds over repeated encounters, yet most health AI systems treat each request in isolation.
By Haoran Li, Jiebi Deng, Tong Jin, Jinghong Han, Yuxin Wang, Zexin Wang, Qingyi Si, Weikang Gong, Xiahai Zhuang, Jia You, Wei Cheng, Jianfeng Feng, Hongcheng Guo
The paper evaluates six onboarding strategies for federated wearable models on five datasets using a leakage‑controlled protocol that fixes source checkpoints and separates calibration from evaluation. Results show that while average accuracy is high, person‑level performance can drop significantly, with some methods causing negative transfer for certain users. The study highlights that mean accuracy alone is insufficient and provides an auditable benchmark and failure map for future development.
By Rahil Aftab, Vineet Kumar Rakesh, Soumya Mazumdar, Tapas Samanta