pith. sign in

arxiv: 2605.18768 · v1 · pith:655V5YDLnew · submitted 2026-04-13 · 💻 cs.IR · cs.HC· cs.MA

ClinQueryAgent: A Conversational Agent for Population Health Management

Joseph S. Boyle , Anthony Dranfield , Mike O'Neil , Maria Liakata , Alison Q. Smithard This is my paper

Pith reviewed 2026-05-21 01:27 UTC · model grok-4.3

classification 💻 cs.IR cs.HCcs.MA
keywords conversational agentnatural language to querypopulation health managementsecure AI deploymenthealth informaticsmulti-agent systemdatabase queriesNHS platform
0
0 comments X

The pith

ClinQueryAgent translates natural language health questions into database queries while keeping all patient data inside a secure local environment.

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

ClinQueryAgent is a conversational system that converts everyday questions about population health into executable database queries. Its architecture routes sensitive operations locally so that powerful cloud-based language models can assist without any patient records leaving the secure NHS setting. A dedicated sub-agent manages information retrieval to limit errors that build up over longer exchanges. The tool has been embedded in an existing platform and used by 128 staff across 15 practices serving 148,319 patients. Testing shows that analysts and clinicians can obtain actionable results from health records through plain-language requests without writing code.

Core claim

The paper presents ClinQueryAgent as a multi-agent system that accepts natural language questions on population health, breaks them into steps, and produces executable database queries. A novel separation keeps all patient data local while allowing cloud models to handle general reasoning and language tasks. Information retrieval is handed to a sub-agent to reduce drift in extended dialogues. Deployment inside a live NHS platform and evaluation on real tasks demonstrate that non-programmers can generate useful population-level insights directly from patient records.

What carries the argument

The multi-agent architecture with local data handling and a dedicated sub-agent for information retrieval that keeps cloud model calls separate from patient records.

If this is right

  • Staff without programming skills can produce population health reports and analyses from patient records.
  • The system supports autonomous completion of a range of health informatics tasks inside an existing platform.
  • Real deployment across 148,000 patients shows the approach scales to multiple practices.
  • Actionable information can be extracted from health records through chat without moving data outside the secure environment.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same local-plus-cloud split could be applied to other domains that need powerful models on sensitive records, such as finance or legal case management.
  • Smaller healthcare organizations might adopt advanced query capabilities without building large local AI infrastructure.
  • Adding more external knowledge sources through the sub-agent could further improve accuracy on complex population questions.

Load-bearing premise

Delegating information retrieval to a sub-agent will effectively prevent inaccuracies that arise from context loss during longer conversations.

What would settle it

A controlled test that measures query error rates on conversations of increasing length with and without the sub-agent delegation; a sharp rise in errors beyond a few turns despite the sub-agent would undermine the claim.

Figures

Figures reproduced from arXiv: 2605.18768 by Alison Q. Smithard, Anthony Dranfield, Joseph S. Boyle, Maria Liakata, Mike O'Neil.

Figure 1
Figure 1. Figure 1: ClinQueryAgent parses natural language questions into database queries in an agentic loop. The Query Agent delegates the task of finding concepts to a Retrieval Agent so that it can focus on creating the database query. The Retrieval Agent has access to a lo￾cal database of medical concepts , and the open-access ontology UMLS , which is used to help the retrieval agent parse unusual acronyms and synonyms. … view at source ↗
Figure 2
Figure 2. Figure 2: ClinQueryAgent (CQA) is situated within the eHealthScope application which serves a variety of users [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: An example of computing statistics across [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
read the original abstract

In this paper we introduce ClinQueryAgent, a system for translating natural language population health questions into executable database queries using agents with access to both local and external knowledge bases. Our novel architecture enables the use of powerful cloud-based language models whilst ensuring that no patient data leaves the secure environment. To combat inaccuracies over the course of longer dialogues due to context rot, information retrieval is delegated to a sub-agent. We deploy the system via a chat window embedded within an existing population health management platform where it has been used by 128 staff from 15 healthcare practices covering a total of 148,319 patients in the UK's National Health Service (NHS). We evaluate the system's capacity to autonomously handle a range of health informatics tasks on a constructed dataset and via a beta-testing phase. Our results show that both analysts and clinicians are able to easily generate actionable information from patient health records using natural language requests requiring no programming expertise to verify. We make a public demo of the system available at: https://demo-899965260288.europe-west1.run.app/

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.

Referee Report

3 major / 2 minor

Summary. The paper introduces ClinQueryAgent, a conversational agent for translating natural language population health questions into executable database queries. It proposes a novel architecture combining local and external knowledge bases to enable cloud-based LLMs while claiming that no patient data leaves the secure environment. Information retrieval is delegated to a sub-agent to address context rot in longer dialogues. The system is deployed in an NHS population health platform and has been used by 128 staff across 15 practices covering 148,319 patients. Evaluation is described on a constructed dataset and via beta-testing, with claims that analysts and clinicians can easily generate actionable information without programming expertise. A public demo link is provided.

Significance. If the security isolation and query accuracy claims are substantiated, the work could meaningfully advance accessible health data querying for non-technical users in clinical settings, with potential benefits for population health management. The reported real-world NHS deployment with substantial patient coverage is a practical strength that, if paired with rigorous metrics, would strengthen the contribution.

major comments (3)
  1. [Abstract and system architecture] Abstract and architecture description: the headline claim that the novel architecture ensures 'no patient data leaves the secure environment' is load-bearing but unsupported by concrete specifications of prompt templates, tool schemas, message boundaries, or data-flow isolation between local records and cloud LLM calls. Without these details it remains possible that retrieval results containing identifiers or record excerpts are serialized into external prompts.
  2. [Evaluation] Evaluation section: the abstract states successful autonomous handling on a constructed dataset and beta-testing with real users, yet provides no quantitative metrics (accuracy, error rates, success rates, or comparison baselines). This leaves the central claims about accuracy and ease of use without detailed supporting evidence.
  3. [System description] System description: the assumption that delegating information retrieval to a sub-agent effectively combats inaccuracies due to context rot over longer dialogues is presented as a design choice but lacks empirical validation, such as dialogue-length ablation results or error-rate comparisons with and without the sub-agent.
minor comments (2)
  1. [Evaluation] The constructed dataset used for autonomous evaluation should be described in more detail (size, query types, ground-truth construction) to allow reproducibility.
  2. [System architecture] Consider adding a diagram of the agent and sub-agent architecture to improve clarity of the data-flow and security boundaries.

Simulated Author's Rebuttal

3 responses · 0 unresolved

We thank the referee for their constructive and detailed feedback, which highlights important areas for clarification and strengthening. We address each major comment below and describe the revisions we will incorporate to improve the manuscript while preserving its core contributions and real-world deployment evidence.

read point-by-point responses

  1. Referee: [Abstract and system architecture] Abstract and architecture description: the headline claim that the novel architecture ensures 'no patient data leaves the secure environment' is load-bearing but unsupported by concrete specifications of prompt templates, tool schemas, message boundaries, or data-flow isolation between local records and cloud LLM calls. Without these details it remains possible that retrieval results containing identifiers or record excerpts are serialized into external prompts.

    Authors: We agree that the security claim is central and would benefit from explicit technical details to rule out potential leakage paths. In the revised manuscript we will add a new subsection (under System Architecture) that specifies the prompt templates, tool schemas for local versus external knowledge bases, message boundary rules, and the precise data-flow isolation. This will include an updated architecture diagram and concrete examples demonstrating that patient identifiers and raw record excerpts remain confined to the local secure environment, with only de-identified query formulations or aggregated outputs passed to the cloud LLM. revision: yes

  2. Referee: [Evaluation] Evaluation section: the abstract states successful autonomous handling on a constructed dataset and beta-testing with real users, yet provides no quantitative metrics (accuracy, error rates, success rates, or comparison baselines). This leaves the central claims about accuracy and ease of use without detailed supporting evidence.

    Authors: We acknowledge that the evaluation section would be strengthened by explicit quantitative metrics. The constructed dataset was designed to test autonomous query generation across a range of health informatics tasks, and the beta-testing captured real usage by 128 staff. In the revised manuscript we will expand the Evaluation section to report accuracy rates, error categories and frequencies, task success rates, and any available baseline comparisons from the dataset experiments, together with summary statistics from the beta-testing phase. revision: yes

  3. Referee: [System description] System description: the assumption that delegating information retrieval to a sub-agent effectively combats inaccuracies due to context rot over longer dialogues is presented as a design choice but lacks empirical validation, such as dialogue-length ablation results or error-rate comparisons with and without the sub-agent.

    Authors: The sub-agent design was motivated by observed context degradation in early internal testing of longer dialogues. We agree that dedicated empirical validation would make the rationale more robust. In the revision we will add a short discussion (in System Description or a new Limitations subsection) that explains the design rationale with illustrative examples drawn from the beta-testing logs and, where feasible, reports comparative error observations between dialogues of varying lengths. Full controlled ablation experiments would require additional dedicated studies beyond the current deployment scope. revision: partial

Circularity Check

0 steps flagged

No circularity: system description paper with no derivations, fits, or self-referential predictions

full rationale

This is a system description and deployment report for ClinQueryAgent. The paper introduces an architecture for translating natural language queries into database queries using agents with local and external knowledge bases, claims secure isolation of patient data from cloud LLMs, and delegates retrieval to a sub-agent to mitigate context rot. These are presented as architectural choices and supported by usage statistics from 128 staff across 15 practices covering 148,319 patients, plus evaluation on a constructed dataset and beta-testing. No equations, fitted parameters, predictions, or uniqueness theorems appear. No self-citation chains, ansatz smuggling, or renaming of known results reduce any claim to its own inputs by construction. The derivation chain is absent; the work is self-contained as an engineering report against external benchmarks of real-world usage.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 2 invented entities

The central claim rests on standard assumptions about LLM query translation capabilities and the effectiveness of the sub-agent design; no numerical free parameters are introduced, and the main invented element is the overall system itself.

axioms (1)
  • domain assumption Large language models can translate natural language health questions into correct executable database queries when given access to local and external knowledge bases.
    Implicit in the core translation functionality described in the abstract.
invented entities (2)
  • ClinQueryAgent no independent evidence
    purpose: Conversational interface for secure population health query translation.
    The primary contribution is the introduction of this named system and its architecture.
  • Sub-agent for information retrieval no independent evidence
    purpose: To prevent context rot and inaccuracies in extended dialogues.
    Explicitly introduced in the abstract to address a stated limitation of longer conversations.

pith-pipeline@v0.9.0 · 5731 in / 1417 out tokens · 50332 ms · 2026-05-21T01:27:06.267937+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

295 extracted references · 295 canonical work pages · 23 internal anchors

  1. [1]

    and Dudley, Joel T

    Miotto, Riccardo and Li, Li and Kidd, Brian A. and Dudley, Joel T. , month = may, year =. Deep. Scientific Reports , publisher =. doi:10.1038/srep26094 , abstract =

    work page doi:10.1038/srep26094

  2. [2]

    Journal of the American Medical Informatics Association: JAMIA , author =

    Data from clinical notes: a perspective on the tension between structure and flexible documentation , volume =. Journal of the American Medical Informatics Association: JAMIA , author =. 2011 , keywords =. doi:10.1136/jamia.2010.007237 , abstract =

    work page doi:10.1136/jamia.2010.007237 2011

  3. [3]

    Annual review of biomedical data science , author =

    Advances in. Annual review of biomedical data science , author =. 2018 , pages =. doi:10.1146/annurev-biodatasci-080917-013315 , abstract =

    work page doi:10.1146/annurev-biodatasci-080917-013315 2018

  4. [4]

    and Tzavella, Loukia , month = jan, year =

    Chambers, Christopher D. and Tzavella, Loukia , month = jan, year =. The past, present and future of. Nature Human Behaviour , publisher =. doi:10.1038/s41562-021-01193-7 , abstract =

    work page doi:10.1038/s41562-021-01193-7

  5. [5]

    and Kraljevic, Zeljko and Shek, Anthony and Teo, James and Dobson, Richard J

    Bean, Daniel M. and Kraljevic, Zeljko and Shek, Anthony and Teo, James and Dobson, Richard J. B. , month = may, year =. Hospital-wide natural language processing summarising the health data of 1 million patients , volume =. PLOS Digital Health , publisher =. doi:10.1371/journal.pdig.0000218 , abstract =

    work page doi:10.1371/journal.pdig.0000218

  6. [6]

    Coiera, Enrico , month = may, year =. Recent. BMJ , publisher =. doi:10.1136/bmj.310.6991.1381 , abstract =

    work page doi:10.1136/bmj.310.6991.1381

  7. [7]

    and Schilling, Lisa M

    Gold, Sigfried and Lehmann, Harold P. and Schilling, Lisa M. and Lutters, Wayne G. , month = dec, year =. Value sets and the problem of redundancy in value set repositories , volume =. PLOS ONE , publisher =. doi:10.1371/journal.pone.0312289 , abstract =

    work page doi:10.1371/journal.pone.0312289

  8. [8]

    Continuous

    Caralt, Mireia Hernandez and Ng, Clarence Boon Liang and Rei, Marek , editor =. Continuous. Proceedings of the 23rd. 2024 , pages =. doi:10.18653/v1/2024.bionlp-1.19 , abstract =

    work page doi:10.18653/v1/2024.bionlp-1.19 2024

  9. [9]

    Hierarchical Retrieval with Out-Of-Vocabulary Queries: A Case Study on SNOMED CT

    Dilworth, Jonathon and Yang, Hui and Chen, Jiaoyan and Gao, Yongsheng , month = nov, year =. Hierarchical. doi:10.48550/arXiv.2511.16698 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2511.16698

  10. [10]

    2023 , pages =

    Journal of the American Medical Informatics Association , author =. 2023 , pages =. doi:10.1093/jamia/ocad149 , abstract =

    work page doi:10.1093/jamia/ocad149 2023

  11. [11]

    M3: Conversational LLMs Simplify Secure Clinical Data Access, Understanding, and Analysis

    Attrach, Rafi Al and Moreira, Pedro and Fani, Rajna and Umeton, Renato and Fiske, Amelia and Celi, Leo Anthony , month = nov, year =. Conversational. doi:10.48550/arXiv.2507.01053 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2507.01053

  12. [12]

    Hanna, Michael , year =. How to. doi:10.1007/978-3-030-02955-5 , abstract =

    work page doi:10.1007/978-3-030-02955-5

  13. [13]

    Journal of the American Geriatrics Society , author =

    Automatable end-of-life screening for older adults in the emergency department using electronic health records , volume =. Journal of the American Geriatrics Society , author =. 2023 , note =. doi:10.1111/jgs.18262 , abstract =

    work page doi:10.1111/jgs.18262 2023

  14. [14]

    Journal of General Internal Medicine , author =

    Electronic. Journal of General Internal Medicine , author =. 2019 , pages =. doi:10.1007/s11606-019-05219-9 , abstract =

    work page doi:10.1007/s11606-019-05219-9 2019

  15. [15]

    Advances in Neural Information Processing Systems , author =

    A. Advances in Neural Information Processing Systems , author =. 2024 , pages =. doi:10.52202/079017-1400 , language =

    work page doi:10.52202/079017-1400 2024

  16. [16]

    Journal of Medical Systems , author =

    Data base design for natural language medical data , volume =. Journal of Medical Systems , author =. 1982 , keywords =. doi:10.1007/BF00994122 , abstract =

    work page doi:10.1007/bf00994122 1982

  17. [17]

    Improving

    Radford, Alec and Narasimhan, Karthik and Salimans, Tim and Sutskever, Ilya , month = jun, year =. Improving

    work page

  18. [18]

    Science , author =

    Dissecting racial bias in an algorithm used to manage the health of populations , volume =. Science , author =. 2019 , keywords =. doi:10.1126/science.aax2342 , abstract =

    work page doi:10.1126/science.aax2342 2019

  19. [19]

    Radiology , author =

    The meaning and use of the area under a receiver operating characteristic (. Radiology , author =. 1982 , keywords =. doi:10.1148/radiology.143.1.7063747 , abstract =

    work page doi:10.1148/radiology.143.1.7063747 1982

  20. [21]

    Journal of Pain and Symptom Management , author =

    Systematic. Journal of Pain and Symptom Management , author =. 2012 , keywords =. doi:10.1016/j.jpainsymman.2011.04.012 , abstract =

    work page doi:10.1016/j.jpainsymman.2011.04.012 2012

  21. [22]

    BMJ supportive & palliative care , author =

    Science of forecasting: new methods to predict death , issn =. BMJ supportive & palliative care , author =. 2025 , keywords =. doi:10.1136/spcare-2025-005746 , language =

    work page doi:10.1136/spcare-2025-005746 2025

  22. [23]

    Greenwood, Major , year =. A

    work page

  23. [24]

    Reports on

    A. Reports on. 1927 , note =. doi:10.1001/jama.1927.02680330059037 , abstract =

    work page doi:10.1001/jama.1927.02680330059037 1927

  24. [25]

    The Journal of Ambulatory Care Management , author =

    New. The Journal of Ambulatory Care Management , author =. 2007 , pages =

    work page 2007

  25. [26]

    Risk stratification

    Midlands. Risk stratification

    work page

  26. [27]

    Kaplan, E. L. and Meier, Paul , month = jun, year =. Nonparametric. Journal of the American Statistical Association , publisher =. doi:10.1080/01621459.1958.10501452 , abstract =

    work page doi:10.1080/01621459.1958.10501452 1958

  27. [28]

    Kaplan, E. L. and Meier, Paul , year =. Nonparametric. Journal of the American Statistical Association , publisher =. doi:10.2307/2281868 , abstract =

    work page doi:10.2307/2281868

  28. [29]

    anthropics/claude-code , url =

    Anthropic , month = feb, year =. anthropics/claude-code , url =

    work page

  29. [30]

    Team, Kimi and Bai, Tongtong and Bai, Yifan and Bao, Yiping and Cai, S. H. and Cao, Yuan and Charles, Y. and Che, H. S. and Chen, Cheng and Chen, Guanduo and Chen, Huarong and Chen, Jia and Chen, Jiahao and Chen, Jianlong and Chen, Jun and Chen, Kefan and Chen, Liang and Chen, Ruijue and Chen, Xinhao and Chen, Yanru and Chen, Yanxu and Chen, Yicun and Che...

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.2602.02276

  30. [31]

    Epidemiology , author =

    Understanding algorithmic fairness for clinical prediction in terms of subgroup net benefit and health equity , issn =. Epidemiology , author =. doi:10.1097/EDE.0000000000001949 , abstract =

    work page doi:10.1097/ede.0000000000001949

  31. [32]

    Journal of Pain & Palliative Care Pharmacotherapy , publisher =

    Funding the. Journal of Pain & Palliative Care Pharmacotherapy , publisher =. 2011 , note =. doi:10.3109/15360288.2011.621020 , abstract =

    work page doi:10.3109/15360288.2011.621020 2011

  32. [33]

    Palliative Medicine , author =

    Identification of patients with potential palliative care needs:. Palliative Medicine , author =. 2020 , pages =. doi:10.1177/0269216320929552 , abstract =

    work page doi:10.1177/0269216320929552 2020

  33. [34]

    BMC medicine , author =

    How many people will need palliative care in 2040?. BMC medicine , author =. 2017 , keywords =. doi:10.1186/s12916-017-0860-2 , abstract =

    work page doi:10.1186/s12916-017-0860-2 2040

  34. [35]

    Noutahi, Emmanuel and Beaini, Dominique and Horwood, Julien and Giguère, Sébastien and Tossou, Prudencio , month = apr, year =. Towards. doi:10.48550/arXiv.1905.11577 , abstract =

    work page doi:10.48550/arxiv.1905.11577 1905

  35. [36]

    and Sun, Jimeng , month = aug, year =

    Choi, Edward and Bahadori, Mohammad Taha and Song, Le and Stewart, Walter F. and Sun, Jimeng , month = aug, year =. Proceedings of the 23rd. doi:10.1145/3097983.3098126 , abstract =

    work page doi:10.1145/3097983.3098126

  36. [37]

    JAMA Network Open , author =

    Diagnostic. JAMA Network Open , author =. 2025 , pages =. doi:10.1001/jamanetworkopen.2025.50454 , abstract =

    work page doi:10.1001/jamanetworkopen.2025.50454 2025

  37. [38]

    doi:10.48550/arXiv.2511.17559 , abstract =

    Lee, Gyubok and Chay, Woosog and Choi, Edward , month = dec, year =. doi:10.48550/arXiv.2511.17559 , abstract =

    work page doi:10.48550/arxiv.2511.17559

  38. [39]

    and Kendall, Marilyn and Boyd, Kirsty and Sheikh, Aziz , month = apr, year =

    Murray, Scott A. and Kendall, Marilyn and Boyd, Kirsty and Sheikh, Aziz , month = apr, year =. Illness trajectories and palliative care , volume =. BMJ , publisher =. doi:10.1136/bmj.330.7498.1007 , abstract =

    work page doi:10.1136/bmj.330.7498.1007

  39. [40]

    Understanding

    Pavlovic, Maja , month = sep, year =. Understanding. doi:10.48550/arXiv.2501.19047 , abstract =

    work page doi:10.48550/arxiv.2501.19047

  40. [41]

    Veličković, Petar and Cucurull, Guillem and Casanova, Arantxa and Romero, Adriana and Liò, Pietro and Bengio, Yoshua , month = feb, year =. Graph. doi:10.48550/arXiv.1710.10903 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1710.10903

  41. [42]

    Knowledge and Information Systems , author =

    Provably accurate and scalable linear classifiers in hyperbolic spaces , volume =. Knowledge and Information Systems , author =. 2023 , keywords =. doi:10.1007/s10115-022-01820-3 , abstract =

    work page doi:10.1007/s10115-022-01820-3 2023

  42. [43]

    Combinatorica , author =

    The geometry of graphs and some of its algorithmic applications , volume =. Combinatorica , author =. 1995 , keywords =. doi:10.1007/BF01200757 , abstract =

    work page doi:10.1007/bf01200757 1995

  43. [44]

    , year =

    Maguire, Rebecca and Maguire, Phil and Keane, Mark T. , year =. Making sense of surprise:. Journal of Experimental Psychology: Learning, Memory, and Cognition , publisher =. doi:10.1037/a0021609 , abstract =

    work page doi:10.1037/a0021609

  44. [45]

    Journal of Forecasting , author =

    How probable is probable?. Journal of Forecasting , author =. 1982 , note =. doi:10.1002/for.3980010305 , abstract =

    work page doi:10.1002/for.3980010305 1982

  45. [46]

    Vogel, Hannah and Appelbaum, Sebastian and Haller, Heidemarie and Ostermann, Thomas , year =. The. German. doi:10.3233/SHTI220798 , pages =

    work page doi:10.3233/shti220798

  46. [47]

    2019 , note =

    End of life care for adults: service delivery , shorttitle =. 2019 , note =

    work page 2019

  47. [48]

    Experiences of dying, death and bereavement in motor neurone disease:

    Whitehead, Bridget and O’Brien, Mary R and Jack, Barbara A and Mitchell, Douglas , month = jun, year =. Experiences of dying, death and bereavement in motor neurone disease:. Palliative Medicine , publisher =. doi:10.1177/0269216311410900 , abstract =

    work page doi:10.1177/0269216311410900

  48. [49]

    Learning Continuous Hierarchies in the Lorentz Model of Hyperbolic Geometry

    Nickel, Maximilian and Kiela, Douwe , month = jul, year =. Learning. doi:10.48550/arXiv.1806.03417 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1806.03417

  49. [50]

    Family Practice , author =

    Identifying patients who could benefit from palliative care by making use of the general practice information system: the. Family Practice , author =. 2020 , pages =. doi:10.1093/fampra/cmaa049 , abstract =

    work page doi:10.1093/fampra/cmaa049 2020

  50. [51]

    Age and Ageing , author =

    Prediction of appropriate timing of palliative care for older adults with non-malignant life-threatening disease: a systematic review , volume =. Age and Ageing , author =. 2005 , pages =. doi:10.1093/ageing/afi054 , abstract =

    work page doi:10.1093/ageing/afi054 2005

  51. [52]

    Clopper, C. J. and Pearson, E. S. , year =. The. Biometrika , publisher =. doi:10.2307/2331986 , number =

    work page doi:10.2307/2331986

  52. [53]

    Hastie, Trevor and Tibshirani, Robert and Friedman, Jerome , year =. The. doi:10.1007/978-0-387-84858-7 , urldate =

    work page doi:10.1007/978-0-387-84858-7

  53. [54]

    Developing clinical prediction models: a step-by-step guide , volume =

    Efthimiou, Orestis and Seo, Michael and Chalkou, Konstantina and Debray, Thomas and Egger, Matthias and Salanti, Georgia , month = sep, year =. Developing clinical prediction models: a step-by-step guide , volume =. BMJ , publisher =. doi:10.1136/bmj-2023-078276 , abstract =

    work page doi:10.1136/bmj-2023-078276 2023

  54. [55]

    Medical Care , author =

    Development and. Medical Care , author =. 2022 , pages =. doi:10.1097/MLR.0000000000001699 , abstract =

    work page doi:10.1097/mlr.0000000000001699 2022

  55. [56]

    NEJM Catalyst , author =

    Improving. NEJM Catalyst , author =. doi:10.1056/CAT.24.0392 , language =

    work page doi:10.1056/cat.24.0392

  56. [57]

    BMC Medicine , author =

    How accurate is the ‘. BMC Medicine , author =. 2017 , keywords =. doi:10.1186/s12916-017-0907-4 , abstract =

    work page doi:10.1186/s12916-017-0907-4 2017

  57. [58]

    Designing

    Huyen, Chip , month = may, year =. Designing

    work page

  58. [59]

    Epidemiology , url =

    work page

  59. [60]

    Occupational Medicine , author =

    The. Occupational Medicine , author =. 2008 , pages =. doi:10.1093/occmed/kqm162 , number =

    work page doi:10.1093/occmed/kqm162 2008

  60. [61]

    and McDowell, Ian and Mitnitski, Arnold , month = aug, year =

    Rockwood, Kenneth and Song, Xiaowei and MacKnight, Chris and Bergman, Howard and Hogan, David B. and McDowell, Ian and Mitnitski, Arnold , month = aug, year =. A global clinical measure of fitness and frailty in elderly people , volume =. CMAJ , publisher =. doi:10.1503/cmaj.050051 , abstract =

    work page doi:10.1503/cmaj.050051

  61. [62]

    Epidemiology , author =

    Assessing the. Epidemiology , author =. 2010 , pages =. doi:10.1097/EDE.0b013e3181c30fb2 , abstract =

    work page doi:10.1097/ede.0b013e3181c30fb2 2010

  62. [63]

    JAMA , author =

    Association. JAMA , author =. 2016 , pages =. doi:10.1001/jama.2016.16840 , abstract =

    work page doi:10.1001/jama.2016.16840 2016

  63. [64]

    International Journal for Quality in Health Care , author =

    Developing evidence-based clinical indicators: a state of the art methods primer , volume =. International Journal for Quality in Health Care , author =. 2003 , pages =. doi:10.1093/intqhc/mzg084 , abstract =

    work page doi:10.1093/intqhc/mzg084 2003

  64. [65]

    Sean and Meier, Diane E

    Morrison, R. Sean and Meier, Diane E. , month = jun, year =. Palliative. New England Journal of Medicine , publisher =. doi:10.1056/NEJMcp035232 , abstract =

    work page doi:10.1056/nejmcp035232

  65. [66]

    TRIPOD+AI statement: updated guidance for reporting clinical prediction models that use regression or machine learning methods

    Collins, Gary S. and Moons, Karel G. M. and Dhiman, Paula and Riley, Richard D. and Beam, Andrew L. and Calster, Ben Van and Ghassemi, Marzyeh and Liu, Xiaoxuan and Reitsma, Johannes B. and Smeden, Maarten van and Boulesteix, Anne-Laure and Camaradou, Jennifer Catherine and Celi, Leo Anthony and Denaxas, Spiros and Denniston, Alastair K. and Glocker, Ben ...

    work page doi:10.1136/bmj-2023-078378 2023

  66. [67]

    Hepatology (Baltimore, Md.) , author =

    A. Hepatology (Baltimore, Md.) , author =. 2021 , keywords =. doi:10.1002/hep.32086 , abstract =

    work page doi:10.1002/hep.32086 2021

  67. [68]

    and Boyd, Kirsty and Sheikh, Aziz , month = mar, year =

    Murray, Scott A. and Boyd, Kirsty and Sheikh, Aziz , month = mar, year =. Palliative care in chronic illness , volume =. BMJ , publisher =. doi:10.1136/bmj.330.7492.611 , abstract =

    work page doi:10.1136/bmj.330.7492.611

  68. [69]

    Primary care must be prioritised to achieve universal palliative care , volume =

    Mitchell, Sarah and Fuller, Claire and Doyle, Amanda and Ziegler, Lucy , month = oct, year =. Primary care must be prioritised to achieve universal palliative care , volume =. British Journal of General Practice , publisher =. doi:10.3399/BJGP.2025.0344 , abstract =

    work page doi:10.3399/bjgp.2025.0344 2025

  69. [71]

    International Journal of Epidemiology , author =

    Interrupted time series regression for the evaluation of public health interventions: a tutorial , issn =. International Journal of Epidemiology , author =. 2016 , pages =. doi:10.1093/ije/dyw098 , abstract =

    work page doi:10.1093/ije/dyw098 2016

  70. [72]

    Axiomatic Attribution for Deep Networks

    Sundararajan, Mukund and Taly, Ankur and Yan, Qiqi , month = jun, year =. Axiomatic. doi:10.48550/arXiv.1703.01365 , abstract =

    work page internal anchor Pith review Pith/arXiv arXiv doi:10.48550/arxiv.1703.01365

  71. [73]

    doi:10.1177/0272989X06295361 , language =

    Decision. doi:10.1177/0272989X06295361 , language =

    work page doi:10.1177/0272989x06295361

  72. [74]

    Duke, Jon and Knoll, Chris and Shah, Nigam , year =

    work page

  73. [75]

    Journal of Biomedical Informatics , author =

    Clinical code set engineering for reusing. Journal of Biomedical Informatics , author =. 2017 , keywords =. doi:10.1016/j.jbi.2017.04.010 , abstract =

    work page doi:10.1016/j.jbi.2017.04.010 2017

  74. [76]

    BMJ Open , author =

    Identifying clinical features in primary care electronic health record studies: methods for codelist development , volume =. BMJ Open , author =. 2017 , pages =. doi:10.1136/bmjopen-2017-019637 , abstract =

    work page doi:10.1136/bmjopen-2017-019637 2017

  75. [77]

    PLoS ONE , author =

    Term sets:. PLoS ONE , author =. 2019 , pages =. doi:10.1371/journal.pone.0212291 , abstract =

    work page doi:10.1371/journal.pone.0212291 2019

  76. [78]

    Hosseini-Asl, Ehsan and McCann, Bryan and Wu, Chien-Sheng and Yavuz, Semih and Socher, Richard , year =. A. Advances in

    work page

  77. [79]

    Journal of Machine Learning Research , author =

    A. Journal of Machine Learning Research , author =. 2003 , pages =

    work page 2003

  78. [80]

    2022 , pages =

    Advances in Neural Information Processing Systems , author =. 2022 , pages =

    work page 2022

  79. [81]

    Health and Technology , author =

    Google. Health and Technology , author =. 2017 , pages =. doi:10.1007/s12553-017-0179-1 , abstract =

    work page doi:10.1007/s12553-017-0179-1 2017

  80. [82]

    2024 , pages =

    Journal of Biomedical Informatics , author =. 2024 , pages =. doi:10.1016/j.jbi.2024.104649 , abstract =

    work page doi:10.1016/j.jbi.2024.104649 2024

Showing first 80 references.