Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; 69 of the 78 references it cites were found in a registry or an index; 55 of the 78 sources its claims cite were checked for support.

Use of artificial intelligence in critical care: opportunities and obstacles

Open access

Abstract

BACKGROUND: Perhaps nowhere else in the healthcare system than in the intensive care unit environment are the challenges to create useful models with direct time-critical clinical applications more relevant and the obstacles to achieving those goals more massive. Machine learning-based artificial intelligence (AI) techniques to define states and predict future events are commonplace activities of modern life. However, their penetration into acute care medicine has been slow, stuttering and uneven. Major obstacles to widespread effective application of AI approaches to the real-time care of the critically ill patient exist and need to be addressed. MAIN BODY: Clinical decision support systems (CDSSs) in acute and critical care environments support clinicians, not replace them at the bedside. As will be discussed in this review, the reasons are many and include the immaturity of AI-based systems to have situational awareness, the fundamental bias in many large databases that do not reflect the target population of patient being treated making fairness an important issue to address and technical barriers to the timely access to valid data and its display in a fashion useful for clinical workflow. The inherent "black-box" nature of many predictive algorithms and CDSS makes trustworthiness and acceptance by the medical community difficult. Logistically, collating and curating in real-time multidimensional data streams of various sources needed to inform the algorithms and ultimately display relevant clinical decisions support format that adapt to individual patient responses and signatures represent the efferent limb of these systems and is often ignored during initial validation efforts. Similarly, legal and commercial barriers to the access to many existing clinical databases limit studies to address fairness and generalizability of predictive models and management tools. CONCLUSIONS: AI-based CDSS are evolving and are here to stay. It is our obligation to be good shepherds of their use and further development.

Overview

CoChat Summary

TL;DR

This review argues that AI-based clinical decision support systems (CDSS) in critical care have substantial potential, but real-world use is slowed by bias, poor situational awareness, workflow integration problems, and governance/regulatory concerns. The authors emphasize that AI in the ICU should augment, not replace, clinicians, and that successful deployment requires rigorous validation, human-factors design, and ongoing monitoring.

Key findings

  • ICU AI/CDSS face multiple obstacles: immature situational awareness, biased databases that may not reflect target populations, technical barriers to timely data access/display, and the “black-box” nature of many models that undermines trust.
  • The paper notes that models for clinical deterioration and sepsis can predict events hours before they occur and have improved outcomes in intervention studies, but the “implementation gap” remains substantial.
  • Real-world EHR data offer large sample sizes and unselected populations, but the paper highlights issues of bias, non-random missingness, semantic ambiguity across EHR vendors, and privacy/legal barriers to secondary use.
  • The authors state that only a few AI-based algorithms have shown evidence for improved clinician performance or patient outcomes in clinical studies, and that evaluating performance, safety, fairness, and human factors is essential before wider deployment.

Methods

  • Narrative review / perspective article synthesizing challenges and opportunities for AI in critical care; not an original empirical study.
  • Draws on real-world ICU/EHR datasets, prior model-development and implementation studies, and published guidance/frameworks such as DECIDE-AI and FDA/WHO/regulatory documents.
  • Main analysis is conceptual: organizes the discussion around real-world data, bias/governance, situational awareness, implementation, workforce training, and adoption roadmap.

Limitations

  • Not stated in the paper.
  • Evident from the text: this is a non-systematic narrative review, so its claims are interpretive rather than based on a single new dataset or formal meta-analysis.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.1186/s13054-024-04860-z

Cites

69 of 69 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Fig. 1 The process of creating, testings, and launching an effective Clinical Decision Support System (CDSS) is multifaceted and ongoing. The interaction of multiple processes and involvement of various stakeholders along the way improve the likelihood of final adoption during real-world deployment (dashed vertical line). Importantly, as illustrated in this work flow diagram, is ongoing assessment refining models and information transfer options. At the start one uses a model card which is a short document that provides key information about a machine learning model. This is central to maintaining focus throughout the workflow cycle of CDSS development
Fig. 1 The process of creating, testings, and launching an effective Clinical Decision Support System (CDSS) is multifaceted and ongoing. The interaction of multiple processes and involvement of various stakeholders along the way improve the likelihood of final adoption during real-world deployment (dashed vertical line). Importantly, as illustrated in this work flow diagram, is ongoing assessment refining models and information transfer options. At the start one uses a model card which is a short document that provides key information about a machine learning model. This is central to maintaining focus throughout the workflow cycle of CDSS developmentCC BY · PubMed Central

Verification

Claims the paper makes

28 Valid · 3 Not valid · 2 Not checked · 28 Own results and interpretations

16 more claims · 9 with their cited sources checked

Create a free account to see all 61 claims and every source verdict.

See all claims — free account
  • It may state two separate claims as one. 21
  • Probably not a substantial claim. 24

Create a free account to see these claims.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Citations

Cites

Show all 78 references

Research works printed but not linked: 4

  1. Heaven WD. Hundreds of AI tools have been built to catch covid. None of them helped. MIT Technology Review. Retrieved October. 2021. p. 6.
  2. Chen X, Zou D, Xie H, Cheng G, Liu C. Contributors, collaborations, research topics, challenges, and future directions. Educ Technol Soc. 2022;25(1):28–47.
  3. Hancock B, Lazaroff-Puck K, Rutherford S. Getting practical about the future of work. McKinsey Quart. 2020;1:65–73.
  4. Whittaker M, Alper M, Bennett CL, Hendren S, Kaziunas L, Mills M, et al. Disability, bias, and AI. AI Now Inst. 2019;8:9.

Cited by

See all 160

Metadata

Authors
Pinsky, M. R., Bedoya, A., Bihorac, A., Celi, L., Churpek, M., Economou-Zavlanos, N. J., Elbers, P., Saria, S., Liu, V., Lyons, P. G., Shickel, B., Toral, P., Tscholl, D., & Clermont, G.
Affiliations
  • University of Pittsburgh (US)
  • Duke University (US)
  • University of Florida Health (US)
  • Massachusetts Institute of Technology (US)
  • University of Wisconsin–Madison (US)
  • Amsterdam University Medical Centers (NL)
  • Johns Hopkins University (US)
  • Oregon Health & Science University (US)
  • University of Zurich (CH)
Corresponding author
Pinsky, Michael R.
Venue
Critical Care
Published
2024-04-08
Volume
28
Issue
1
Pages
113-113
ISSN
1364-8535 (ISSN-L)
OpenAlex
W4394571531
Cited by
160
References
78
Citations by year
2026: 80 · 2025: 63 · 2024: 17
FWCI
3.65
Type
journal article
Language
en
Open access
Licence
Locations
  • Critical Care (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • Pure Amsterdam UMC (publishedVersion, other-oa)
  • Zurich Open Repository and Archive (University of Zurich) (acceptedVersion, cc-by)
  • DSpace@MIT (Massachusetts Institute of Technology) (submittedVersion, cc-by-nc)
  • Amsterdam UMC (VU Amsterdam) - Institutional Repository (submittedVersion, other-oa)
  • Zurich Open Repository and Archive (University of Zurich) (submittedVersion, cc-by)
  • Zurich Open Repository and Archive (University of Zurich)
Funders
  • National Institutes of Health (HL141916)
  • National Institutes of Health,United States (NR013912)
Topics
  • Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
  • Physical Sciences › Computer Science › Artificial Intelligence › Machine Learning in Healthcare
  • Health Sciences › Health Professions › Health Information Management › Artificial Intelligence in Healthcare
Keywords
artificial intelligence, intensive care unit, clinical decision support systems, predictive modeling, algorithmic bias, black-box algorithms, trust in AI, data access barriers, clinical workflow integration
MeSH terms
Algorithms, Artificial Intelligence, Critical Care, Delivery of Health Care, Humans, Intensive Care Units

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Use of artificial intelligence in critical care: opportunities and obstacles. https://app.cochat.ai/papers/10.1186/s13054-024-04860-z
The paper
Pinsky, M. R., Bedoya, A., Bihorac, A., Celi, L., Churpek, M., Economou-Zavlanos, N. J., Elbers, P., Saria, S., Liu, V., Lyons, P. G., Shickel, B., Toral, P., Tscholl, D., & Clermont, G. (2024). Use of artificial intelligence in critical care: opportunities and obstacles. Critical Care. https://doi.org/10.1186/s13054-024-04860-z

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.