arxiv: 2604.12019 · v3 · submitted 2026-04-13 · 💻 cs.AI · cs.HC

A Framework for Longitudinal Health AI Agents

Georgianna Lin , Rencong Jiang , No\'emie Elhadad , Xuhai "Orson" Xu This is my paper

Pith reviewed 2026-05-10 15:53 UTC · model grok-4.3

classification 💻 cs.AI cs.HC

keywords longitudinal health AIAI agentsmulti-layer frameworkadaptationcoherencecontinuityagencyhealth informatics

0 comments

The pith

A multi-layer framework operationalizes adaptation, coherence, continuity, and agency for AI agents in longitudinal health interactions.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper draws on clinical and personal health informatics to define how AI agents should handle health tasks like symptom management and behavior change across repeated interactions instead of isolated queries. It proposes a multi-layer framework and agent architecture that puts adaptation to changing goals, coherent reasoning across sessions, continuity of user context, and agency in decision support into practice. Current AI health tools often lack follow-up and sustained alignment with individual goals, which limits both effectiveness and safety. Representative use cases illustrate how such agents could maintain engagement, adjust to evolving needs, and enable safer personalized support over time. The work supplies concrete guidance for building systems that track health trajectories beyond single exchanges.

Core claim

We draw on established clinical and personal health informatics frameworks to define what it would mean to orchestrate longitudinal health interactions with AI agents. We propose a multi-layer framework and corresponding agent architecture that operationalizes adaptation, coherence, continuity, and agency across repeated interactions. Through representative use cases, we demonstrate how longitudinal agents can maintain meaningful engagement, adapt to evolving goals, and support safe, personalized decision-making over time.

What carries the argument

The multi-layer framework and corresponding agent architecture that operationalizes adaptation, coherence, continuity, and agency across repeated interactions.

Load-bearing premise

Synthesizing established clinical and personal health informatics frameworks into this multi-layer architecture will enable AI agents to effectively support longitudinal health interactions in practice.

What would settle it

A controlled comparison of agents built with the proposed multi-layer architecture against baseline single-session agents, measuring user-reported coherence, goal alignment, and continuity across at least five simulated health sessions.

Figures

Figures reproduced from arXiv: 2604.12019 by Georgianna Lin, No\'emie Elhadad, Rencong Jiang, Xuhai "Orson" Xu.

**Figure 1.** Figure 1: A four-part framework integrating coherence, continuity, adaptation, and agency to support sustained health engagement over time. Adaptation refines knowledge through responsiveness, personalization, and reflexivity, while coherence builds stable interpretive foundations (history, organization, relationship, persistence). Continuity sustains long-term goals via follow-up, alignment, and accountability, and… view at source ↗

**Figure 2.** Figure 2: A potential timeline of a longitudinal health agent supporting an individual with endometriosis. The agent maintains coherence by integrating past symptoms and interventions, ensures continuity by actively tracking strategies and follow-ups, enables adaptation by recalibrating guidance as symptoms and context change, and fosters agency by scaffolding patient reflection and self-management. Post-discharge f… view at source ↗

read the original abstract

Although artificial intelligence (AI) agents are increasingly proposed to support potentially longitudinal health tasks, such as symptom management, behavior change, and patient support, most current implementations fall short of facilitating user intent and fostering accountability. This contrasts with prior work on supporting longitudinal needs, both within and beyond clinical settings, where follow-up, coherent reasoning, and sustained alignment with individuals' goals are critical for both effectiveness and safety. In this paper, we draw on established clinical and personal health informatics frameworks to define what it would mean to orchestrate longitudinal health interactions with AI agents. We propose a multi-layer framework and corresponding agent architecture that operationalizes adaptation, coherence, continuity, and agency across repeated interactions. Through representative use cases, we demonstrate how longitudinal agents can maintain meaningful engagement, adapt to evolving goals, and support safe, personalized decision-making over time. Our findings underscore both the promise and the complexity of designing systems capable of supporting health trajectories beyond isolated interactions, and we offer guidance for future research and development in multi-session, user-centered health AI.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

A clean synthesis of existing health informatics ideas into a multi-layer agent architecture, but it stops at the blueprint stage with no implementation details or tests.

read the letter

The paper's main contribution is a multi-layer framework that tries to turn one-off health AI interactions into sustained ones by layering in adaptation, coherence, continuity, and agency. It pulls from established clinical and personal health informatics work and shows the idea through a few representative use cases like symptom management and behavior change. That synthesis is useful as a high-level map for why current agents often fail at follow-up and goal alignment over time. The use cases make the intent concrete without overclaiming results. The architecture itself is not a radical departure, but organizing the properties this way gives people a shared vocabulary to build on. The central weakness is that the proposal remains descriptive. There are no formal definitions of the layers, no memory or reasoning mechanisms spelled out, no algorithms, and no evaluation criteria or pilot data. Without those, it is impossible to judge whether the structure actually produces the claimed properties in practice or just restates the problem in new terms. The paper acknowledges the complexity but does not move past it. This is the kind of framing paper that belongs in a workshop or a journal that publishes conceptual work in health informatics. Readers already working on longitudinal agents or digital health systems will find the structure helpful for organizing their own thinking. It does not contain enough technical substance or evidence to stand as a finished result. I would send it out for peer review rather than desk reject, with the expectation that referees will push for at least one concrete implementation sketch or evaluation plan.

Referee Report

2 major / 1 minor

Summary. The paper proposes a multi-layer framework and corresponding agent architecture for longitudinal health AI agents. Drawing on established clinical and personal health informatics frameworks, it defines and aims to operationalize adaptation, coherence, continuity, and agency across repeated interactions for tasks such as symptom management, behavior change, and patient support. The proposal is illustrated through representative use cases showing how such agents could maintain engagement, adapt to evolving goals, and support safe personalized decisions over time.

Significance. If the framework can be further specified and implemented, it could provide valuable guidance for developing health AI systems that address the limitations of single-session agents, potentially improving long-term effectiveness and safety in chronic care and behavior support. The synthesis of prior frameworks is a constructive step, but the conceptual nature without implementation details or metrics limits demonstrated impact.

major comments (2)

[multi-layer framework and agent architecture] The central claim that the multi-layer framework and agent architecture operationalize coherence and continuity rests on high-level descriptions without formal definitions, memory models, state-transition mechanisms, or algorithms for cross-session reasoning (see the sections on the proposed framework and agent architecture). This makes it impossible to verify that the synthesis produces the claimed properties rather than remaining descriptive.
[representative use cases] The representative use cases are presented as demonstrations of adaptation and agency, yet they contain no quantitative evaluation criteria, success metrics (e.g., coherence scores over simulated sessions), or comparisons to non-longitudinal baselines, which are load-bearing for substantiating that the architecture achieves its goals in practice.

minor comments (1)

Clarify the precise boundaries and interactions between the proposed layers to prevent potential overlap in responsibilities for adaptation versus continuity.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for their detailed and constructive feedback on our manuscript. We agree that the current version is primarily conceptual and will make revisions to clarify the scope, enhance the descriptions of the framework, and discuss evaluation approaches. Our responses to the major comments are as follows.

read point-by-point responses

Referee: The central claim that the multi-layer framework and agent architecture operationalize coherence and continuity rests on high-level descriptions without formal definitions, memory models, state-transition mechanisms, or algorithms for cross-session reasoning (see the sections on the proposed framework and agent architecture). This makes it impossible to verify that the synthesis produces the claimed properties rather than remaining descriptive.

Authors: We recognize that the framework is presented at a high level of abstraction, consistent with its role as a conceptual synthesis of prior clinical and health informatics frameworks. The multi-layer architecture is designed to provide a structure for operationalizing the key properties through layered components that handle adaptation, coherence, continuity, and agency. To address this, we will revise the manuscript to include more explicit descriptions of the layers, including example mechanisms for memory and state management across sessions, without claiming full algorithmic specifications. This will help readers better understand how the properties are intended to be achieved. revision: partial
Referee: The representative use cases are presented as demonstrations of adaptation and agency, yet they contain no quantitative evaluation criteria, success metrics (e.g., coherence scores over simulated sessions), or comparisons to non-longitudinal baselines, which are load-bearing for substantiating that the architecture achieves its goals in practice.

Authors: The use cases serve to illustrate potential applications of the framework in real-world health scenarios, such as symptom management and behavior change, rather than to provide empirical validation. As this is a framework paper, we do not include quantitative evaluations or simulations. We will revise to explicitly state that the use cases are illustrative and add a new subsection outlining possible metrics and evaluation strategies for future implementations of the architecture, such as longitudinal coherence tracking and comparisons to single-session agents. revision: partial

Circularity Check

0 steps flagged

No circularity: framework is external synthesis without reduction to inputs

full rationale

The paper's derivation consists of drawing on established external clinical and personal health informatics frameworks to define and propose a multi-layer agent architecture that operationalizes adaptation, coherence, continuity, and agency. This is presented as a definitional synthesis demonstrated via representative use cases, with no equations, fitted parameters, predictions, or self-referential derivations that reduce the claimed outputs to the inputs by construction. No load-bearing self-citations, uniqueness theorems, or ansatzes are invoked. The central claim remains an independent organizational proposal rather than a tautology.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 1 invented entities

The proposal rests primarily on domain assumptions from prior health informatics work and introduces the framework as a new organizational structure without independent empirical grounding.

axioms (1)

domain assumption Established clinical and personal health informatics frameworks provide a sufficient basis for defining requirements of longitudinal health interactions.
Invoked in the abstract when stating the framework draws on these to define orchestration of longitudinal interactions.

invented entities (1)

Multi-layer framework and agent architecture for longitudinal health AI no independent evidence
purpose: To operationalize adaptation, coherence, continuity, and agency in repeated health interactions.
This is the core proposed construct without external validation or falsifiable predictions provided in the abstract.

pith-pipeline@v0.9.0 · 5485 in / 1186 out tokens · 45398 ms · 2026-05-10T15:53:03.998406+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

71 extracted references · 71 canonical work pages · 1 internal anchor

[1]

Nedos, I.et al.Is artificial intelligence ready for emergency department triage? a retrospective evaluation of multiple large language models in 39,375 patients at a university emergency department.J. Clin. Medicine15, 1512, DOI: 10.3390/jcm15041512 (2026)

work page doi:10.3390/jcm15041512 2026
[2]

npj Digit

Gaber, F.et al.Evaluating large language model workflows in clinical decision support for triage and referral and diagnosis. npj Digit. Medicine8, 263, DOI: 10.1038/s41746-025-01684-1 (2025)

work page doi:10.1038/s41746-025-01684-1 2025
[3]

Xie, Q.et al.Medical foundation large language models for comprehensive text analysis and beyond.NPJ digital medicine 8, 141, DOI: 10.1038/s41746-025-01533-1 (2025)

work page doi:10.1038/s41746-025-01533-1 2025
[4]

& Talukdar, W

Biswas, A. & Talukdar, W. Intelligent clinical documentation: Harnessing generative ai for patient-centric clinical note generation.Int. J. Innov. Sci. Res. Technol.9, DOI: 10.38124/ijisrt/IJISRT24MAY1483 (2024)

work page doi:10.38124/ijisrt/ijisrt24may1483 2024
[5]

& Saikia, M

Maity, S. & Saikia, M. J. Large language models in healthcare and medical applications: a review.Bioengineering12, 631, DOI: 10.3390/bioengineering12060631 (2025)

work page doi:10.3390/bioengineering12060631 2025
[6]

& Margetis, K

Aydin, S., Karabacak, M., Vlachos, V . & Margetis, K. Large language models in patient education: a scoping review of applications in medicine.Front. medicine11, 1477898, DOI: 10.3389/fmed.2024.1477898 (2024)

work page doi:10.3389/fmed.2024.1477898 2024
[7]

& Kuo, C.-F

Lin, C. & Kuo, C.-F. Roles and potential of large language models in healthcare: a comprehensive review.Biomed. J.48, 100868, DOI: 10.1016/j.bj.2025.100868 (2025)

work page doi:10.1016/j.bj.2025.100868 2025
[8]

InProceedings of the 2025 CHI Conference on Human Factors in Computing Systems, DOI: 10.1145/3706598.3713819 (2024)

Jörke, M.et al.GPTCoach: towards LLM-based physical activity coaching. InProceedings of the 2025 CHI Conference on Human Factors in Computing Systems, DOI: 10.1145/3706598.3713819 (2024)

work page doi:10.1145/3706598.3713819 2025
[9]

C.et al.Advancing health coaching: A comparative study of large language model and health coaches.Artif

Ong, Q. C.et al.Advancing health coaching: A comparative study of large language model and health coaches.Artif. Intell. Medicine157, 103004, DOI: 10.1016/j.artmed.2024.103004 (2024)

work page doi:10.1016/j.artmed.2024.103004 2024
[10]

nursing scholarship44, 136–144, DOI: 10.1111/j.1547-5069.2012.01444.x (2012)

Schulman-Green, D.et al.Processes of self-management in chronic illness.J. nursing scholarship44, 136–144, DOI: 10.1111/j.1547-5069.2012.01444.x (2012)

work page doi:10.1111/j.1547-5069.2012.01444.x 2012
[11]

F.et al.Conversational agents supporting self-management in people with a chronic disease: Systematic review.J

Peerbolte, T. F.et al.Conversational agents supporting self-management in people with a chronic disease: Systematic review.J. Med. Internet Res.27, e72309, DOI: 10.2196/72309 (2025)

work page doi:10.2196/72309 2025
[12]

M.et al.Using large language models for chronic disease management tasks: scoping review.JMIR Med

Serugunda, H. M.et al.Using large language models for chronic disease management tasks: scoping review.JMIR Med. Informatics13, e66905, DOI: 10.2196/66905 (2025)

work page doi:10.2196/66905 2025
[13]

& Yalçin, Ö

Shayaninasab, M., Zahoor, M. & Yalçin, Ö. N. Enhancing patient intake process in mental health consultations using rag-driven chatbot. In2024 12th International Conference on Affective Computing and Intelligent Interaction Workshops and Demos (ACIIW), 256–264, DOI: 10.1109/ACIIW63320.2024.00053 (IEEE, 2024)

work page doi:10.1109/aciiw63320.2024.00053 2024
[14]

Ayers, J. W.et al.Comparing physician and artificial intelligence chatbot responses to patient questions posted to a public social media forum.JAMA internal medicine183, 589–596, DOI: 10.1001/jamainternmed.2023.1838 (2023)

work page doi:10.1001/jamainternmed.2023.1838 2023
[15]

InProceedings of the 2025 CHI Conference on Human Factors in Computing Systems, 1–18, DOI: 10.1145/3706598.3713307 (2024)

Haag, D.et al.The last jitai? exploring large language models for issuing just-in-time adaptive interventions: fostering physical activity in a prospective cardiac rehabilitation setting. InProceedings of the 2025 CHI Conference on Human Factors in Computing Systems, 1–18, DOI: 10.1145/3706598.3713307 (2024)

work page doi:10.1145/3706598.3713307 2025
[16]

Artsi, Y .et al.Large language models in real-world clinical workflows: a systematic review of applications and implementation.Front. Digit. Heal.7, 1659134, DOI: 10.3389/fdgth.2025.1659134 (2025)

work page doi:10.3389/fdgth.2025.1659134 2025
[17]

& Sabeti, F

Farzan, M., Ebrahimi, H., Pourali, M. & Sabeti, F. Artificial intelligence-powered cognitive behavioral therapy chatbots, a systematic review.Iran. journal psychiatry20, 102–110, DOI: 10.18502/ijps.v20i1.17395 (2025)

work page doi:10.18502/ijps.v20i1.17395 2025
[18]

Wang, J.et al.Psychological counseling cannot be achieved overnight: Automated psychological counseling through multi-session conversations.arXiv preprint arXiv:2506.06626DOI: 10.48550/arXiv.2506.06626 (2025)

work page doi:10.48550/arxiv.2506.06626 2025
[19]

Medicine6, 14, DOI: 10.1038/s43856-025-01321-8 (2026)

McFadyen, J.et al.Increasing engagement with cognitive-behavioral therapy (CBT) using generative AI: a randomized controlled trial (RCT).Commun. Medicine6, 14, DOI: 10.1038/s43856-025-01321-8 (2026). Online ahead of print

work page doi:10.1038/s43856-025-01321-8 2026
[20]

& Dinesh, D

Sinha, C., Thakkar, R., Meheli, S. & Dinesh, D. Exploring the role of app features in providing continuity of care to users on a digital mental health platform (Wysa): Retrospective mixed methods observational study.JMIR Form. Res.10, e73033, DOI: 10.2196/73033 (2026)

work page doi:10.2196/73033 2026
[21]

Zhang, C.et al.A survey on multi-turn interaction capabilities of large language models.arXiv preprint arXiv:2501.09959 DOI: 10.48550/arXiv.2501.09959 (2025)

work page doi:10.48550/arxiv.2501.09959 2025
[22]

A., Schers, H

Uijen, A. A., Schers, H. J., Schellevis, F. G. & van den Bosch, W. J. How unique is continuity of care? a review of continuity and related concepts.Fam. practice29, 264–271, DOI: 10.1093/fampra/cmr104 (2012). 11/14

work page doi:10.1093/fampra/cmr104 2012
[23]

Saultz, J. W. & Lochner, J. Interpersonal continuity of care and care outcomes: a critical review.Annals Fam. Medicine3, 159–166, DOI: 10.1370/afm.285 (2005)

work page doi:10.1370/afm.285 2005
[24]

Gray, D. J. P., Sidaway-Lee, K., White, E., Thorne, A. & Evans, P. H. Continuity of care with doctors—a matter of life and death? A systematic review of continuity of care and mortality.BMJ Open8, e021161, DOI: 10.1136/bmjopen-2017-021161 (2018)

work page doi:10.1136/bmjopen-2017-021161 2017
[25]

& Forster, A

Van Walraven, C., Oake, N., Jennings, A. & Forster, A. J. The association between continuity of care and outcomes: a systematic and critical review.J. evaluation clinical practice16, 947–956, DOI: 10.1111/j.1365-2753.2009.01235.x (2010)

work page doi:10.1111/j.1365-2753.2009.01235.x 2009
[26]

InFindings of the Association for Computational Linguistics: EMNLP 2022, 3395–3407, DOI: 10.18653/v1/2022.findings-emnlp.247 (2022)

Zhang, T.et al.History-aware hierarchical transformer for multi-session open-domain dialogue system. InFindings of the Association for Computational Linguistics: EMNLP 2022, 3395–3407, DOI: 10.18653/v1/2022.findings-emnlp.247 (2022)

work page doi:10.18653/v1/2022.findings-emnlp.247 2022
[28]

the nation remains a key unit of shared experience and its educational and cultural institutions shape the values of almost everyone in that society

Ge, Y .et al.TReMu: Towards neuro-symbolic temporal reasoning for LLM-agents with memory in multi-session dialogues. InFindings of the Association for Computational Linguistics: ACL 2025, 18974–18988, DOI: 10.18653/v1/ 2025.findings-acl.972 (2025)

work page doi:10.18653/v1/ 2025
[29]

LongMemEval: Benchmarking Chat Assistants on Long-Term Interactive Memory

Wu, D.et al.Longmemeval: Benchmarking chat assistants on long-term interactive memory. InProceedings of the International Conference on Learning Representations (ICLR 2025), DOI: 10.48550/arXiv.2410.10813 (2025)

work page internal anchor Pith review doi:10.48550/arxiv.2410.10813 2025
[30]

Reynolds, R.et al.A systematic review of chronic disease management interventions in primary care.BMC Fam. Pract. 19, 11, DOI: 10.1186/s12875-017-0692-3 (2018)

work page doi:10.1186/s12875-017-0692-3 2018
[31]

& Macleod, U

Jones, D., Dunn, L., Watt, I. & Macleod, U. Safety netting for primary care: evidence from a literature review.Br. J. Gen. Pract.69, e70–e79, DOI: 10.3399/bjgp18X700193 (2019)

work page doi:10.3399/bjgp18x700193 2019
[32]

L., Westbrook, J

Callen, J. L., Westbrook, J. I., Georgiou, A. & Li, J. Failure to follow-up test results for ambulatory patients: A systematic review.J. Gen. Intern. Medicine27, 1334–1348, DOI: 10.1007/s11606-011-1949-5 (2011)

work page doi:10.1007/s11606-011-1949-5 1949
[33]

Rothman, A. A. & Wagner, E. H. Chronic illness management: What is the role of primary care?Annals Intern. Medicine 138, 256–261, DOI: 10.7326/0003-4819-138-3-200302040-00034 (2003)

work page doi:10.7326/0003-4819-138-3-200302040-00034 2003
[34]

& Thompson, M

Almond, S., Mant, D. & Thompson, M. Diagnostic safety-netting.Br. J. Gen. Pract.59, 872–874, DOI: 10.3399/ bjgp09X472971 (2009)

work page 2009
[35]

InProceedings of the SIGCHI Conference on Human Factors in Computing Systems (CHI ’10)

Li, I., Dey, A. & Forlizzi, J. A stage-based model of personal informatics systems. InProceedings of the SIGCHI Conference on Human Factors in Computing Systems (CHI ’10), 557–566, DOI: 10.1145/1753326.1753409 (2010)

work page doi:10.1145/1753326.1753409 2010
[36]

Nahum-Shani, I., Hekler, E. B. & Spruijt-Metz, D. Building health behavior models to guide the development of just-in-time adaptive interventions: A pragmatic framework.Heal. Psychol.34, 1209–1219, DOI: 10.1037/hea0000306 (2015)

work page doi:10.1037/hea0000306 2015
[37]

C.et al.Personalized interventions for behaviour change: A scoping review of just-in-time adaptive interventions

Hsu, T.-C. C.et al.Personalized interventions for behaviour change: A scoping review of just-in-time adaptive interventions. Br. J. Heal. Psychol.30, e12766, DOI: 10.1111/bjhp.12766 (2024)

work page doi:10.1111/bjhp.12766 2024
[38]

Bosschaerts, K.et al.Designing a just-in-time adaptive intervention with trigger detection and a generative chatbot: Smoking cessation use case.DIGITAL HEALTH11, DOI: 10.1177/20552076251381747 (2025)

work page doi:10.1177/20552076251381747 2025
[39]

Lu, T., Lin, Q., Yu, B. & Hu, J. A systematic review of strategies in digital technologies for motivating adherence to chronic illness self-care.npj Heal. Syst.2, 13, DOI: 10.1038/s44401-025-00017-4 (2025)

work page doi:10.1038/s44401-025-00017-4 2025
[40]

Chen, C.et al.Followupbot: An llm-based conversational robot for automatic postoperative follow-up.arXiv preprint arXiv:2507.15502DOI: 10.48550/arXiv.2507.15502 (2025)

work page doi:10.48550/arxiv.2507.15502 2025
[41]

Mamykina, L., Smaldone, A. M. & Bakken, S. R. Adopting the sensemaking perspective for chronic disease self- management.J. Biomed. Informatics56, 406–417, DOI: 10.1016/j.jbi.2015.06.006 (2015)

work page doi:10.1016/j.jbi.2015.06.006 2015
[42]

Medicine1, 20172, DOI: 10.1038/s41746-017-0002-4 (2018)

Noah, B.et al.Impact of remote patient monitoring on clinical outcomes: an updated meta-analysis of randomized controlled trials.npj Digit. Medicine1, 20172, DOI: 10.1038/s41746-017-0002-4 (2018)

work page doi:10.1038/s41746-017-0002-4 2018
[43]

Hamine, S., Gerth-Guyette, E., Faulx, D., Green, B. B. & Ginsburg, A. S. Impact of mHealth chronic disease management on treatment adherence and patient outcomes: A systematic review.J. Med. Internet Res.17, e52, DOI: 10.2196/jmir.3951 (2015). 12/14

work page doi:10.2196/jmir.3951 2015
[44]

N., Truong, K

Lin, G., Le, M. N., Truong, K. N. & Mariakakis, A. The cognitive strategies behind multimodal health sensemaking: A menstrual health tracking case study.Proc. ACM on Interactive, Mobile, Wearable Ubiquitous Technol.9, DOI: 10.1145/3749482 (2025)

work page doi:10.1145/3749482 2025
[45]

InDeep learning techniques for biomedical and health informatics, 231–255 (Springer, 2019)

Mulani, J.et al.Deep reinforcement learning based personalized health recommendations. InDeep learning techniques for biomedical and health informatics, 231–255 (Springer, 2019). 46.Vegesna, A., Tran, M., Angelaccio, M. & Arcona, S. Remote patient monitoring via non-invasive digital technologies: A systematic review.Telemedicine e-Health23, 3–17, DOI: 10....

work page doi:10.1089/tmj.2016.0051 2019
[46]

& Hestevik, C

Smedslund, G., Osteras, N. & Hestevik, C. H. Effects of remote patient monitoring on health care utilization in patients with noncommunicable diseases: Systematic review and meta-analysis.JMIR mHealth uHealth13, e68464, DOI: 10.2196/68464 (2025)

work page doi:10.2196/68464 2025
[47]

Merrill, Akshay Paruchuri, Naghmeh Rezaei, Geza Kovacs, Javier Perez, Jake Chan, Shyam Dash, Xin Liu, Daniel McDuff, and Tim Althoff

Merrill, M. A.et al.Transforming wearable data into personal health insights using large language model agents.Nat. Commun.17, 1143, DOI: 10.1038/s41467-025-67922-y (2026)

work page doi:10.1038/s41467-025-67922-y 2026
[48]

Abbasian, M., Azimi, I., Rahmani, A. M. & Jain, R. Conversational health agents: a personalized large language model-powered agent framework.JAMIA open8, ooaf067, DOI: 10.1093/jamiaopen/ooaf067 (2025)

work page doi:10.1093/jamiaopen/ooaf067 2025
[49]

Mamykina, L.et al.Personal discovery in diabetes self-management: Discovering cause and effect using self-monitoring data.J. Biomed. Informatics76, 1–8, DOI: 10.1016/j.jbi.2017.09.013 (2017)

work page doi:10.1016/j.jbi.2017.09.013 2017
[50]

Reports15, 18215 (2025)

Su, J.et al.Investigating the factors influencing users’ adoption of artificial intelligence health assistants based on an extended utaut model.Sci. Reports15, 18215 (2025)

work page 2025
[51]

& Alambeigi, H

Afroogh, S., Akbari, A., Malone, E., Kargar, M. & Alambeigi, H. Trust in ai: progress, challenges, and future directions. Humanit. Soc. Sci. Commun.11, 1568 (2024)

work page 2024
[52]

A., Levin, J., Kahn, J

Sivaraman, V ., Bukowski, L. A., Levin, J., Kahn, J. M. & Perer, A. Ignore, trust, or negotiate: understanding clinician acceptance of ai-based treatment recommendations in health care. InProceedings of the 2023 CHI Conference on Human Factors in Computing Systems, 1–18 (2023)

work page 2023
[53]

Yu, C.et al.From passive to proactive: A multi-agent system with dynamic task orchestration for intelligent medical pre-consultation.arXiv preprint arXiv:2511.01445DOI: 10.48550/arXiv.2511.01445 (2025)

work page doi:10.48550/arxiv.2511.01445 2025
[54]

M., van Keizerswaard, J., Gholiof, M

Mick, I., Freger, S. M., van Keizerswaard, J., Gholiof, M. & Leonardi, M. Comprehensive endometriosis care: a modern multimodal approach for the treatment of pelvic pain and endometriosis.Ther. Adv. Reproductive Heal.18, 26334941241277759, DOI: 10.1177/26334941241277759 (2024)

work page doi:10.1177/26334941241277759 2024
[55]

M., Gattrell, W

Becker, C. M., Gattrell, W. T., Gude, K. & Singh, S. S. Reevaluating response and failure of medical treatment of endometriosis: a systematic review.Fertility Steril.108, 125–136, DOI: 10.1016/j.fertnstert.2017.05.004 (2017)

work page doi:10.1016/j.fertnstert.2017.05.004 2017
[56]

& Perry, M

Devan, H., Hale, L., Hempel, D., Saipe, B. & Perry, M. A. What works and does not work in a self-management intervention for people with chronic pain? Qualitative systematic review and meta-synthesis.Phys. Ther.98, 381–397, DOI: 10.1093/ptj/pzy029 (2018)

work page doi:10.1093/ptj/pzy029 2018
[57]

W., Saunders, P

Edgley, K., Horne, A. W., Saunders, P. T. K. & Tsanas, A. Symptom tracking in endometriosis using digital technologies: Knowns, unknowns, and future prospects.Cell Reports Medicine4, 101192, DOI: 10.1016/j.xcrm.2023.101192 (2023)

work page doi:10.1016/j.xcrm.2023.101192 2023
[58]

Trepanier, L. C. M.et al.Smartphone apps for menstrual pain and symptom management: A scoping review.Internet Interv.31, 100605, DOI: 10.1016/j.invent.2023.100605 (2023)

work page doi:10.1016/j.invent.2023.100605 2023
[59]

J., Jacobsen, A

Requadt, E., Nahlik, A. J., Jacobsen, A. & Ross, W. T. Patient experiences of endometriosis diagnosis: A mixed methods approach.BJOG: An Int. J. Obstet. & Gynaecol.131, 941–951 (2024)

work page 2024
[60]

Gracia, E.et al.The vulnerable phase of heart failure.Am. J. Ther.25, e456–e464, DOI: 10.1097/MJT.0000000000000794 (2018)

work page doi:10.1097/mjt.0000000000000794 2018
[61]

J.et al.The vulnerable phase after hospitalization for heart failure.Nat

Greene, S. J.et al.The vulnerable phase after hospitalization for heart failure.Nat. Rev. Cardiol.12, 220–229, DOI: 10.1038/nrcardio.2015.14 (2015)

work page doi:10.1038/nrcardio.2015.14 2015
[62]

S., Chapel, D., Mendez, J

Regalbuto, R., Maurer, M. S., Chapel, D., Mendez, J. & Shaffer, J. A. Joint commission requirements for discharge instructions in patients with heart failure: is understanding important for preventing readmissions?J. Cardiac Fail.20, 641–649, DOI: 10.1016/j.cardfail.2014.06.358 (2014)

work page doi:10.1016/j.cardfail.2014.06.358 2014
[63]

Heidenreich, P. A.et al.2022 aha/acc/hfsa guideline for the management of heart failure: A report of the american college of cardiology/american heart association joint committee on clinical practice guidelines.Circulation145, e895–e1032, DOI: 10.1161/CIR.0000000000001063 (2022). 13/14

work page doi:10.1161/cir.0000000000001063 2022
[64]

Weiss, A. J. & Jiang, H. J. Overview of clinical conditions with frequent and costly hospital readmissions by payer, 2018. HCUP Statistical Brief #278. Agency for Healthcare Research and Quality (2021). Accessed 2026-03-01

work page 2018
[65]

K., Yang, J., Hernandez, A

Lee, K. K., Yang, J., Hernandez, A. F., Steimle, A. E. & Go, A. S. Post-discharge follow-up characteristics associated with 30-day readmission after heart failure hospitalization.Med. Care54, 365–372, DOI: 10.1097/MLR.0000000000000492 (2016)

work page doi:10.1097/mlr.0000000000000492 2016
[66]

URL https://journals.plos.org/plosone/article?id=10.1371/journal.pone

Tung, Y .-C., Chang, G.-M., Chang, H.-Y . & Yu, T.-H. Relationship between early physician follow-up and 30-day readmission after acute myocardial infarction and heart failure.PLOS ONE12, e0170061, DOI: 10.1371/journal.pone. 0170061 (2017)

work page doi:10.1371/journal.pone 2017
[67]

Lainscak, M.et al.Self-care management of heart failure: Practical recommendations from the patient care committee of the heart failure association of the european society of cardiology.Eur. J. Hear. Fail.13, 115–126, DOI: 10.1093/eurjhf/hfq219 (2011)

work page doi:10.1093/eurjhf/hfq219 2011
[68]

M., Cox, A

Balaskas, A., Schueller, S. M., Cox, A. L. & Doherty, G. Ecological momentary interventions for mental health: A scoping review.PLOS ONE16, e0248152, DOI: 10.1371/journal.pone.0248152 (2021)

work page doi:10.1371/journal.pone.0248152 2021
[69]

Torous, J.et al.The growing field of digital psychiatry: current evidence and the future of apps, social media, chatbots, and virtual reality.World Psychiatry20, 318–335, DOI: 10.1002/wps.20883 (2021)

work page doi:10.1002/wps.20883 2021
[70]

Psychol.111, 65–72 (2015)

Haaker, J.et al.Deficient inhibitory processing in trait anxiety: Evidence from context-dependent fear learning, extinction recall and renewal.Biol. Psychol.111, 65–72 (2015)

work page 2015
[71]

& Owen, G

Hindmarch, T., Hotopf, M. & Owen, G. S. Depression and decision-making capacity for treatment or research: a systematic review.BMC Med. Ethics14, 54, DOI: 10.1186/1472-6939-14-54 (2013)

work page doi:10.1186/1472-6939-14-54 2013
[72]

Medicine8, 574, DOI: 10.1038/s41746-025-01956-w (2025)

Si, Y .et al.Quality, safety and disparity of an AI chatbot in managing chronic diseases: simulated patient experiments.npj Digit. Medicine8, 574, DOI: 10.1038/s41746-025-01956-w (2025). Acknowledgements Research reported in this publication was supported by the National Library of Medicine Training Grant and the Columbia University Research Stabilization...

work page doi:10.1038/s41746-025-01956-w 2025