From Staff Messages to Actionable Insights: A Multi-Stage LLM Classification Framework for Healthcare Analytics

Franziska Jovin; Hajar Sakai; Joshua Bosire; Mohammadsadegh Mikaeili; Yi-En Tseng

REVIEW 4 major objections 5 minor 18 references

From Staff Messages to Actionable Insights: A Multi-Stage LLM Classification Framework for Healthcare Analytics

T0 review · 4 major / 5 minor · reviewed 2026-08-05 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read Hospital staff messages can be auto-classified into actionable topics with 78.4% weighted F1 using a three-stage LLM pipeline.

desk verdict A useful applied LLM classification paper for hospital staff messages, but the gold labels are not documented well enough to trust the model ranking. read the letter →

arxiv 2509.05484 v1 pith:7NH37C4F submitted 2025-09-05 cs.CL

Hajar Sakai , Yi-En Tseng , Mohammadsadegh Mikaeili , Joshua Bosire , Franziska Jovin This is my paper · ORCID

classification cs.CL

keywords staffmessageshealthcareanalyticslargelanguagemodelsmulti-classclassificationtopicmodelingdecisionsupportdashboardcallcenterHIPAA

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

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

The reading

This paper argues that the unstructured messages hospital call-center navigators send while handling patient requests can be automatically turned into structured, operationally meaningful categories without traditional supervised training. The proposed framework first catches easy cases with keyword-based rules, then sends the leftovers through an LLM, and finally feeds still-unclassified cases to the LLM with the full message thread for context. Across 17 LLMs evaluated on 500 manually labeled real staff messages, the best result is a weighted F1-score of 78.4% and accuracy of 79.2% with the o3 model. If this holds, hospitals can mine a high-volume, previously underused data source to spot training needs, staffing imbalances, and emerging patient-access issues.

What carries the argument

The load-bearing mechanism is the multi-stage classification function F(m), which routes each message through three increasingly expensive stages: f1 applies manually identified keywords and key phrases; f2 applies a prompted LLM to messages the first stage flagged as 'Other'; f3 applies a contextual prompt containing the entire encounter thread to the remaining 'Other' messages. This cascade is what lets the framework keep computation low, because only hard cases reach the LLM, while still producing a single multi-class label per message.

What would settle it

Take a random sample of the 500 evaluation messages and have two annotators who have never seen the GPT-4o topic hierarchy label them independently; if pairwise agreement is low, or if the labels differ from the original set, the reported 78.4% weighted F1 would not survive re-evaluation.

Watch

Extended reading notes

Core claim

The central claim is that a three-stage cascade—keyword-based primary classification, LLM-based secondary classification, and full-thread contextual tertiary classification—can classify hospital staff messages into a multi-class topic hierarchy with useful accuracy, and that the choice of LLM matters enormously, with weighted F1 ranging from 46.4% to 78.4%. The taxonomy of topics is itself derived by querying GPT-4o over 1,000 messages with a retrieval-augmented prompt, then the cascade is evaluated on 500 manually labeled messages from a 2,000-message dataset. The best model, o3, reaches 78.4% weighted F1 and 79.2% accuracy, ahead of gpt-5 (75.3% and 76.2%) and DeepSeek-R1 (74.1% and 74.8%)

Load-bearing premise

The 500 manually labeled messages used as the evaluation set are accurate and were labeled without being influenced by the LLM-generated topic hierarchy; the paper gives no annotator details, no agreement measure, and no procedure ensuring independence.

Editorial extensions

If this is right

Hospital call-center staff messages can be monitored continuously as structured topic distributions rather than read individually.
Recurring non-clinical topics such as rescheduling, insurance, and billing can be linked to specific teams and offices, targeting workflow changes and training.
The roughly 40-point accuracy gap between best and worst models means LLM choice, not just pipeline design, drives whether such a system is usable.
The same HIPAA-compliant pipeline should transfer to other healthcare communication settings with minimal retraining, since it relies on prompts rather than fine-tuned weights.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

Because the topic taxonomy was produced by GPT-4o, part of o3's edge may come from family alignment with the taxonomy; a taxonomy built by a non-OpenAI model would test whether the ranking is robust.
The paper does not ablate the three stages separately; measuring the marginal gain of stage 2 and stage 3 would show whether the keyword prefilter helps or accidentally buries hard cases.
The ground-truth labels need independent reannotation: if the manual labels were shaped by the same hierarchy, all reported scores would be upper bounds on real-world performance.
The framing suggests a natural next test: compare o3 on this task against a fine-tuned smaller model with the same 500 labels, to see if flagship reasoning cost is actually necessary.

Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Colleague, quick take on arXiv:2509.05484. It's an applied paper from Cooper University Health Care: they build a three-stage pipeline to categorize internal call-center staff messages into an operational topic hierarchy, deploy it in a Power BI dashboard, and compare 17 Azure OpenAI models on a private 500-message gold set. o3 gets 78.4 weighted F1, gpt-5 75.3, DeepSeek-R1 74.1. If you are looking for a real-world deployment story, this is one of the more concrete ones I've seen: actual PHI constraints, HIPAA-compliant workflow, a dashboard that managers use, and an honest acknowledgment of bias/hallucination risk.

What's new: the domain and the private benchmark. The method is not novel—keyword stage plus LLM stages is a standard funnel—but the paper doesn't claim otherwise. The 17-model comparison on this type of operational text is useful evidence for people choosing models in similar settings.

The soft spot is exactly where the stress test lands, and it can't be waved away. The gold labels are the load-bearing part, and §3.1 gives no annotator details and no inter-annotator agreement. §3.2.1 says the topic hierarchy was derived by prompting GPT-4o on 1,000 messages. If the 500 manual labels were assigned using that same hierarchy, the evaluation is measuring agreement with the LLM's coding scheme rather than independent ground truth. I don't know the answer, because the paper doesn't say. That's the problem. On top of that, there are no confidence intervals or significance tests; the 3.1-point gap between o3 and gpt-5 is about 6 misclassified messages out of 500, so it's plausibly within noise. The stage 3 sample size also varies by LLM, which makes the per-stage comparison hard to interpret for the tail of the funnel. Prompts are not included, and the data can't be shared, though de-identified examples would have helped.

So: the engineering and the deployment work are credible, and the paper reads like a genuine operational report. But as a scientific evaluation, the headline ranking is not yet supported. It's a solid candidate for peer review, not a desk reject, because the application matters and the evaluation issues are addressable. I'd ask for annotator details, a small de-identified sample, exact prompts, and a supervised baseline (even a simple SVM on TF-IDF, using the same 500 labels) to calibrate the LLM numbers. Also worth telling the authors to report the raw confusion matrix or per-class counts so reviewers can see where the gap comes from.

For your reading group: maybe, as an example of the gap between deployment value and evaluation rigor. I wouldn't cite it for the numbers, but I'd cite it as an example of applied LLM classification in healthcare if that's your angle.

Referee Report

4 major / 5 minor

Summary. The paper presents a three-stage LLM-based classification framework for hospital staff encounter messages. Stage 1 applies a keyword-based primary classifier; Stage 2 sends unclassified messages to an LLM with prompt P2; Stage 3 applies a contextual LLM prompt P3 to the remaining messages using the full encounter thread. The topic hierarchy is derived by prompting GPT-4o on 1,000 messages. The pipeline is evaluated on 500 manually labeled messages drawn from a 2,000-message dataset, with 17 Azure OpenAI models compared. The reported best model is o3, with 78.4% weighted F1 and 79.2% accuracy, followed by gpt-5 (75.3% / 76.2%) and DeepSeek-R1 (74.1% / 74.8%). The outputs are integrated into a Power BI dashboard for hospital decision support.

Significance. The application is genuinely useful and comparatively unusual: hospital staff messages are an underutilized, high-volume data source, and the multi-stage design reduces LLM inference costs. The paper reports a real 500-message manual evaluation, 17 model comparisons, and deployment considerations including HIPAA. If the evaluation is valid, the o3 result and the multi-stage framework would be a solid applied contribution for healthcare analytics. However, the central performance claims currently rest on label provenance and evaluation protocol details that are not documented. The annotation methodology, uncertainty quantification, and reproducibility materials must be strengthened before the reported rankings can be relied upon.

major comments (4)

[§3.1, §3.2.1] The 500 manually labeled messages are the sole ground truth for every accuracy/F1 number, but the paper provides no annotation protocol, annotator qualifications, or inter-annotator agreement. More importantly, the topic hierarchy in Fig. 4 was created by prompting GPT-4o on 1,000 messages, and the paper does not state whether the manual labels were assigned independently of that hierarchy. If annotators used the GPT-4o-generated taxonomy, the evaluation measures agreement with an LLM-derived coding scheme rather than an independent ground truth, and the o3-vs-gpt-5 ranking could reflect prompt sensitivity or label noise. Please report the annotation instrument, whether annotators were blinded to the LLM-derived taxonomy, and IAA statistics (e.g., Cohen's kappa) on a double-coded subset.
[§4.1, Fig. 5] No confidence intervals or significance tests are reported for the 500-sample evaluation. The headline difference between o3 (78.4% weighted F1) and gpt-5 (75.3%) is only 3.1 points; with n=500 this may be within sampling noise, especially for a heavily skewed multi-class problem. Please provide bootstrap confidence intervals, McNemar's test for pairwise model comparisons, or per-class standard errors so the ranking can be assessed statistically.
[§4.1, Eq. (4)] The text states that Stage 3 uses 'a variable sample set size at stage 3 depending on the LLM.' Because Stage 2 routing depends on the model, different models may send different subsets of messages to Stage 3. If the 500 evaluation messages are not a fixed set with the same final-label path for every model, the cross-model metrics are not directly comparable. Please clarify whether all 500 gold messages are scored by the same pipeline path for every model, or describe how variable Stage 3 composition is controlled in the reported metrics.
[§3.2.2, §4.1] The method is defined by the keyword-based function f1 and prompts P2/P3, but neither the exact prompts nor the keyword/key-phrase list are included in the paper or appendix. Without these, the 'reproducible pipeline' claim cannot be verified, and prompt sensitivity cannot be assessed. Please include the full prompts, the keyword list, and the model version/API snapshot (e.g., dates and model identifiers) in an appendix.

minor comments (5)

[§3.2.2] Typo: 'hierarichal' should be 'hierarchical'. Also, 'Weighted F1-score' capitalization is inconsistent across the abstract and results.
[§4.1] The text says 'Open-source models like grok-3 and gpt-oss-120b'; grok-3 is not open-source. Only gpt-oss-120b is open-weight in the listed table. Please correct this.
[§2] In the related work, the sentence 'developed a fusion framework...' is missing the author attribution for the Ren et al. work; the citation appears as [12] but the subject of the sentence is absent.
[§4.1] The paper would benefit from one supervised baseline (e.g., a fine-tuned lightweight transformer) trained on the same 500 labels to contextualize the absolute 78.4% F1. This is not essential to the LLM comparison, but it would strengthen the practical claim.
[Fig. 5] Figure 5 would be more informative with error bars or confidence intervals; as presented, the reader cannot distinguish real differences from noise. This is related to major comment 2.

Circularity Check

0 steps flagged · score 0.0 of 10

No circularity: the paper is an empirical LLM benchmark against a manually labeled gold set, not a derivation from fitted inputs.

full rationale

The paper's central claim is an empirical comparison of 17 LLMs on a 500-message manually labeled subset (§3.1, §4.1). The topic hierarchy in Figure 4 was generated by GPT-4o from 1,000 messages (§3.2.1), but the evaluation measures agreement with human labels, not with the LLM's own taxonomy. No equation defines a reported metric in terms of a fitted parameter; no parameter is fitted; the cascade in Eq. (1)–(4) is a deterministic routing rule, not a circular construction. The only self-citation ([14]) is background in Related Work and is not load-bearing for the reported results. The absence of annotator details or inter-annotator agreement is a legitimate external-validity concern, but the paper does not state that the manual labels were derived from the GPT-4o taxonomy, and no text-level reduction can be exhibited. Therefore, under the standards requiring a specific exhibited reduction, no circularity is present.

Assumptions & free parameters 0 free parameters · 3 assumptions · 0 invented entities

The paper contributes no fitted mathematical constants or new entities. Its load-bearing assumptions are the quality of the manual labels, the validity of the LLM-generated topic taxonomy, and the equivalence of the undocumented prompt setups across 17 models.

assumptions (3)

domain assumption The 500-message manual labels are accurate and reliable.
No annotation protocol, annotator qualifications, or agreement statistics are reported; the labels carry the entire evaluation (Section 3.1).
ad hoc to paper The GPT-4o-derived topic hierarchy is complete and meaningful for all staff messages.
The hierarchy was created by chatting with GPT-4o on 1,000 messages and internal brainstorming, with no external validation or error analysis (Section 3.2.1).
domain assumption LLM API outputs were parsed correctly and each model was given equivalent prompts and content.
Exact prompts, temperature settings, and output parsing are not included, so prompt equivalence across models is assumed (Sections 3.2.2, 4.1).

how reviews work

0 comments

Cite this review

Pith. "Pith review of From Staff Messages to Actionable Insights: A Multi-Stage LLM Classification Framework for Healthcare Analytics." pith.science (2026). https://pith.science/paper/7NH37C4F

@misc{pith2026250905484,
  author       = {Pith},
  title        = {Pith review of: From Staff Messages to Actionable Insights: A Multi-Stage LLM Classification Framework for Healthcare Analytics},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/7NH37C4F}},
  note         = {Machine review of arXiv:2509.05484}
}

read the original abstract

Hospital call centers serve as the primary contact point for patients within a hospital system. They also generate substantial volumes of staff messages as navigators process patient requests and communicate with the hospital offices following the established protocol restrictions and guidelines. This continuously accumulated large amount of text data can be mined and processed to retrieve insights; however, traditional supervised learning approaches require annotated data, extensive training, and model tuning. Large Language Models (LLMs) offer a paradigm shift toward more computationally efficient methodologies for healthcare analytics. This paper presents a multi-stage LLM-based framework that identifies staff message topics and classifies messages by their reasons in a multi-class fashion. In the process, multiple LLM types, including reasoning, general-purpose, and lightweight models, were evaluated. The best-performing model was o3, achieving 78.4% weighted F1-score and 79.2% accuracy, followed closely by gpt-5 (75.3% Weighted F1-score and 76.2% accuracy). The proposed methodology incorporates data security measures and HIPAA compliance requirements essential for healthcare environments. The processed LLM outputs are integrated into a visualization decision support tool that transforms the staff messages into actionable insights accessible to healthcare professionals. This approach enables more efficient utilization of the collected staff messaging data, identifies navigator training opportunities, and supports improved patient experience and care quality.

Figures

Figures reproduced from arXiv: 2509.05484 by the authors.

**Figure 2.** Contribution This research is structured as follows: The "Related Work" section reviews literature on LLM-based text mining and data visualization. The "Data and Methodology" section describes the dataset used and introduces the proposed multi-stage methodology encompassing topic modeling, text classification, and decision support tool development. The 2 [PITH_FULL_IMAGE:figures/full_fig_p002_2.png] view at source ↗

**Figure 3.** Proposed Methodology 3.2.1 Topic Modeling The staff messages collected represent the raw data; however, this data is characterized by not being associated with any topics or labels. Since the main goal of the classification is to shed light on the reasons behind sending these messages to thereafter improve the process’s efficiency by uncovering areas of training improvement, an LLM (i.e., GPT-4o) was first used to i… view at source ↗

Figures from the paper (9 more)

**Figure 4.** Figure 4: Topics’ Hierarchy 3.2.2 Three-stage Classification [PITH_FULL_IMAGE:figures/full_fig_p005_4.png]

**Figure 5.** Figure 5: Metrics Summary On one hand, the top 3 performing LLMs are dominated by OpenAI’s flagship offerings as well as DeepSeek-R1, with the o3 reasoning model achieving the highest weighted F1-score of 78.4% and close accuracy of 79.2%. gpt-5 follows with a weighted F1-score …

**Figure 6.** Figure 6: Inference Time Summary It is important to keep in mind that when deploying LLMs for multi-class classification of staff messaging, careful consideration of the risks of bias and hallucination could significantly impact decision-making accuracy. LLMs are prone to exhibi…

**Figure 7.** Figure 7: Clinical Topic Distribution [PITH_FULL_IMAGE:figures/full_fig_p009_7.png]

**Figure 8.** Figure 8: Non-clinical Topic Distribution 9 [PITH_FULL_IMAGE:figures/full_fig_p009_8.png]

**Figure 9.** Figure 9: Clinical Reason Topic Distribution Over Time [PITH_FULL_IMAGE:figures/full_fig_p010_9.png]

**Figure 10.** Figure 10: Non-Clinical Reason Topic Distribution Over Time By tracking staff messages categories over time, the decision support tool enables stakeholders at multiple organizational levels to address critical business questions and operational challenges. Healthcare professiona…

**Figure 11.** Figure 11: Top 5 LLMs’ Per-class F1-score 12 [PITH_FULL_IMAGE:figures/full_fig_p012_11.png]

**Figure 12.** Figure 12: LLMs’ Performance Heatmap 13 [PITH_FULL_IMAGE:figures/full_fig_p013_12.png]

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

18 extracted references · 17 canonical work pages

[1]

HIMSS. (2020). Digital health: A framework for healthcare transformation. https://www.himss.org/resources/digital-health-framework-healthcare-transformation-white-paper

work page 2020
[2]

R., & Turner, K

Hossain, E., Rana, R., Higgins, N., Soar, J., Datta Barua, P., Pisani, A. R., & Turner, K. (2023). Natural Language Processing in Electronic Health Records in relation to healthcare decision-making: A systematic review. Computers in Biology and Medicine, 155, 106649. https://doi.org/10.1016/j.compbiomed.2023.106649 PubMed

work page arXiv 2023
[3]

Comparative Analysis of Text Classification Approaches in Electronic Health Records

Mascio, A., Kraljevi´c, Ž., Bean, D., Dobson, R., Stewart, R., Bendayan, R., & Roberts, A. (2020, May). Com- parative analysis of text classification approaches in electronic health records. arXiv Preprint arXiv:2005.06624. arXivResearchGate

work page Pith review arXiv 2020
[4]

L., Chen, I

Ghassemi, M., Naumann, T., Schulam, P., Beam, A. L., Chen, I. Y ., & Ranganath, R. (2018). A review of challenges and opportunities in machine learning for health. arXiv Preprint arXiv:1806.00388. arXivHealthy ML

arXiv 2018
[5]

J., Crossley, S., Semere, W., Duran, N

Brown III, W., Balyan, R., Karter, A. J., Crossley, S., Semere, W., Duran, N. D., Lyles, C., Liu, J., Moffet, H. H., Daniels, R., McNamara, D. S., & Schillinger, D. (2021, January). Challenges and solutions to employing natural language processing and machine learning to measure patients’ health literacy and physician writing complexity: The ECLIPPSE stud...

work page arXiv 2021
[6]

Zhuang, M., Concannon, D., & Manley, E. (2022). A framework for evaluating dashboards in healthcare. IEEE Transactions on Visualization and Computer Graphics, 28(4), 1715–1731

work page 2022
[7]

L., Xu, J., Xu, J., Zhou, X., & Zhang, K

He, J., Baxter, S. L., Xu, J., Xu, J., Zhou, X., & Zhang, K. (2019). The practical implementation of artificial intelligence technologies in medicine. Nature Medicine, 25(1), 30–36

work page 2019
[8]

E., Jackson, K., & DeShazo, J

Heisey-Grove, D., Rathert, C., McClelland, L. E., Jackson, K., & DeShazo, J. (2021). Classification of patient- and clinician-generated secure messages using a theory-based taxonomy. Health Science Reports, 4(2), e295

work page 2021

Show all 18 references

[9]

Kalabikhina, I., Moshkin, V ., Kolotusha, A., Kashin, M., Klimenko, G., & Kazbekova, Z. (2024). Advancing semantic classification: A comprehensive examination of machine learning techniques in analyzing Russian- language patient reviews. Mathematics, 12(4), 566

2024
[10]

M., Jackson, G

Sulieman, L., Gilmore, D., French, C., Cronin, R. M., Jackson, G. P., Russell, M., & Fabbri, D. (2017). Classifying patient portal messages using convolutional neural networks. Journal of Biomedical Informatics, 74, 59–70

2017
[11]

W., Redfern, J., Thiagalingam, A., & Chow, C

Klimis, H., Nothman, J., Lu, D., Sun, C., Cheung, N. W., Redfern, J., Thiagalingam, A., & Chow, C. K. (2021). Text message analysis using machine learning to assess predictors of engagement with mobile health chronic disease prevention programs: Content analysis. JMIR mHealth ...

2021
[12]

W., Khurana, A., Fu, S., Wu, D., Hongfang, L., & Huang, M

Ren, Y ., Wu, Y ., Fan, J. W., Khurana, A., Fu, S., Wu, D., Hongfang, L., & Huang, M. (2024). Automatic uncovering of patient primary concerns in portal messages using a fusion framework of pretrained language models. Journal of the American Medical Informatics Association, 31...

2024
[13]

Chen, X., Ma, W., Li, D., Zhu, F., Routray, S., Guduri, M., & Margala, M. (2025). Kalman-based adaptive moment estimation optimisation algorithm to enhance GPT in LLMs for medical sentiment analysis of patient health-related feedback. IEEE Journal of Biomedical and Health Info...

2025
[14]

S., Mikaeili, M., Bosire, J., & Jovin, F

Sakai, H., Lam, S. S., Mikaeili, M., Bosire, J., & Jovin, F. (2024). Large language models for patient comments multi-label classification. arXiv preprint arXiv:2410.23528

2024 arXiv
[15]

B., Jung, K

Cho, S., Lee, M., Yu, J., Yoon, J., Choi, J. B., Jung, K. H., & Cho, J. (2024). Leveraging large language models for improved understanding of communications with patients with cancer in a call center setting: Proof-of-concept study. Journal of Medical Internet Research, 26, e63892

2024
[16]

Batugo, A., Hwang, S., Davoudi, A., Luong, T., Lee, N., & Mowery, D. L. (2024). Textual triage: Assessing GPT-4 for classification of free-text medication-related messages for hypertension management. medRxiv, 2024-09

2024
[17]

Arias, M., Rojas, E., Aguirre, S., Cornejo, F., Munoz-Gama, J., Sepulveda, M., & Capurro, D. (2023). Defining healthcare KPIs using process mining and patient journey maps. In 2023 XLIX Latin American Computer Conference (CLEI). IEEE. https://doi.org/10.1109/CLEI60451.2023.10345781

2023
[18]

Jangland, E., Gunningberg, L., & Carlsson, M. (2009). Patients’ and relatives’ complaints about encounters and communication in health care: Evidence for quality improvement. Patient Education and Counseling, 75(2), 199–204. https://doi.org/10.1016/j.pec.2008.10.007 Appendix F...

2009 doi

Pith tools

Reviewed August 5, 2026 · model on record in the stance chip above.

[1] [1]

HIMSS. (2020). Digital health: A framework for healthcare transformation. https://www.himss.org/resources/digital-health-framework-healthcare-transformation-white-paper

work page 2020

[2] [2]

R., & Turner, K

Hossain, E., Rana, R., Higgins, N., Soar, J., Datta Barua, P., Pisani, A. R., & Turner, K. (2023). Natural Language Processing in Electronic Health Records in relation to healthcare decision-making: A systematic review. Computers in Biology and Medicine, 155, 106649. https://doi.org/10.1016/j.compbiomed.2023.106649 PubMed

work page arXiv 2023

[3] [3]

Comparative Analysis of Text Classification Approaches in Electronic Health Records

Mascio, A., Kraljevi´c, Ž., Bean, D., Dobson, R., Stewart, R., Bendayan, R., & Roberts, A. (2020, May). Com- parative analysis of text classification approaches in electronic health records. arXiv Preprint arXiv:2005.06624. arXivResearchGate

work page Pith review arXiv 2020

[4] [4]

L., Chen, I

Ghassemi, M., Naumann, T., Schulam, P., Beam, A. L., Chen, I. Y ., & Ranganath, R. (2018). A review of challenges and opportunities in machine learning for health. arXiv Preprint arXiv:1806.00388. arXivHealthy ML

arXiv 2018

[5] [5]

J., Crossley, S., Semere, W., Duran, N

Brown III, W., Balyan, R., Karter, A. J., Crossley, S., Semere, W., Duran, N. D., Lyles, C., Liu, J., Moffet, H. H., Daniels, R., McNamara, D. S., & Schillinger, D. (2021, January). Challenges and solutions to employing natural language processing and machine learning to measure patients’ health literacy and physician writing complexity: The ECLIPPSE stud...

work page arXiv 2021

[6] [6]

Zhuang, M., Concannon, D., & Manley, E. (2022). A framework for evaluating dashboards in healthcare. IEEE Transactions on Visualization and Computer Graphics, 28(4), 1715–1731

work page 2022

[7] [7]

L., Xu, J., Xu, J., Zhou, X., & Zhang, K

He, J., Baxter, S. L., Xu, J., Xu, J., Zhou, X., & Zhang, K. (2019). The practical implementation of artificial intelligence technologies in medicine. Nature Medicine, 25(1), 30–36

work page 2019

[8] [8]

E., Jackson, K., & DeShazo, J

Heisey-Grove, D., Rathert, C., McClelland, L. E., Jackson, K., & DeShazo, J. (2021). Classification of patient- and clinician-generated secure messages using a theory-based taxonomy. Health Science Reports, 4(2), e295

work page 2021