Pickering, Brian W.

Background/Objectives: Novel vision-language models (VLMs) can integrate patient textual data with image data to support medical diagnosis. Recent studies reported conflicting results regarding the performance of multimodal VLMs compared to other models and physician performance. This systematic review aims to assess the diagnostic performance of multimodal VLMs integrating both patient textual and image data across diverse real-world hospital settings. Methods: We performed comprehensive searches of eight resources, including Embase, MEDLINE, and SCOPUS, on 17 December 2025. Eligible studies reporting diagnostic performance of VLMs integrating both image and patient history textual data from real-world adult patients compared to that of other models and physicians were included. The review adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. The Prediction model study Risk Of Bias Assessment Tool + AI (PROBAST + AI) was used to assess the quality and risk of bias. The study protocol was registered in the PROSPERO database (CRD420251244054). This review received no external funding. Results: We screened 11,026 records, of which 18 studies met the inclusion criteria. Six studies comparing multimodal and unimodal models demonstrated the consistent superiority of the multimodal models. Four studies evaluating VLM accuracy as standalone agents compared with physician performance reported conflicting evidence. One study assessing VLMs as a clinical copilot demonstrated higher accuracy from the group of physicians using VLM assistance. A meta-analysis could not be performed due to the heterogeneity across study populations and outcomes. The majority of the studies were assessed as having a high risk of bias due to dataset quality. Primary limitations identified across studies include small sample size, a lack of external validation, and the need for prospective clinical deployment studies. No study provided documented considerations regarding model safety or data security. Conclusions: This systematic review suggests that multimodal VLMs consistently outperform unimodal models with access to only image or text. While model performance as standalone agents compared to humans remains inconclusive, a copilot model has demonstrated high diagnostic accuracy. Given substantial methodological concerns across studies, cautious interpretation is required, No firm clinical recommendation can be made regarding the use of standalone VLMs. Further research employing high-quality datasets is needed to ensure the reliability and clinical applicability of future VLMs.

Abstract Introduction Large Language Models (LLMs) have proven effective at analyzing unstructured and structured data from electronic health records (EHR). Their reliance on static EHR data means they miss the dynamic, high-resolution physiological signals that are the first indicators of patient decline, creating a fundamental data lag as clinical documentation of an event may occur hours after its onset. This represents a critical gap in the ability to develop truly effective triage tools. Furthermore, while LLMs can synthesize multi-source data for acuity assessment, they exhibit known limitations, including hallucinations that constrain autonomous use in safety-critical settings. In contrast, a Vision-Language Model (VLM) with strong image-grounded clinical reasoning can interpret vital sign trend images, surpassing the limitations of current EHR-based models. This approach provides immediate triage evaluations that lead to earlier detection of patient deterioration. Methods This mixed-methods study deployed a VLM (Gemini 2.5) integrated within the eICU dashboard. We used near-real-time two-hour trends in heart rate, blood pressure, respiratory rate, and oxygen saturation from consecutive patient cases during alerts in hospital floor units monitored by eICU that admit acute medical and surgical patients. To increase the diversity of the sample, we sought routine cases from patients without alerts and emergent cases from patients admitted to the medical, surgical, trauma, and mixed ICUs. A structured context prompt was used to generate a standardized summary including clinical interpretation, data quality and artifact assessment, triage recommendation, and the model’s chain-of-thought reasoning. Results We included patients from Mayo Clinic Rochester, MN, using an 800-patient cohort for prompt development and a 100-patient cohort for validation against clinically adjudicated results. In the validation cohort the VLM achieved 80% accuracy (80/100). An analysis of the 20 incorrect predictions revealed a clear pattern of conservative over-triage: 80% (16/20) of errors were clinically acceptable adjacent-class discrepancies (routine↔urgent or urgent↔emergent). Only 4/20 discrepancies represented a substantive misclassification. Most errors therefore reflected conservative prioritization rather than incorrect assessment. The model demonstrated a perfect safety profile, underestimating patient acuity in 0% of cases, and thus avoiding the most clinically dangerous error. These results confirm the feasibility of VLM-based approach and provide a strong foundation for the proposed work. Conclusion This VLM performs cross-modal reasoning by fusing deep visual perception with broad clinical context. The result is a fully contextualized and explainable triage prioritization, which provides direct decision support and integrates into the clinical workflow, allowing for faster and more confident acute care assessments. This abstract is funded by: none

Abstract Rationale: Ongoing healthcare staffing shortages have prompted new care paradigms that incorporate remote care delivery to support timely detection of patient physiological deterioration through mechanisms other than clinician bedside evaluations. In partnership with institutional Information Technology, our team has developed a unique clinical informatics environment, known as the Advanced Medical Practice (AMP) platform, that incorporates data from multiple sources and platforms and supports remote observation of patients from the electronic intensive care unit(eICU). AMP takes advantage of technologies outside of vendor-supported EHRs, and its advanced user-interface design reduces information overload while incorporating real-time data from novel devices and sensors. We partnered with the hospital practice to co-develop a score within the AMP platform(Clinical Events Detection and Response[CEDAR]) that incorporates heart rate, blood pressure, respiratory rate, and SpO2 to stratify patient risk and generate deterioration alerts for the electronic ICU(eICU) team. This formative evaluation employed an implementation science approach to refine CEDAR and ensure it is acceptable to users and can be feasibly implemented in practice. Methods: The study team conducted this pilot work in the eICU at Mayo Clinic (Rochester, MN). We observed the eICU workflow and documented feedback from eICU staff in fieldnotes. eICU staff were asked to log issues in a tracking file during shifts. Issues were reviewed daily and discussed at multidisciplinary team meetings weekly. Observation and tracking data were coded to constructs in the Consolidated Framework for Implementation Research(CFIR 2.0) framework and mapped to Expert Recommendations for Implementation Change(ERIC) strategies that targeted key determinants. Results: Between 9/23/24 and 10/25/24 we completed 25 days of observations. Most identified issues related to technical/design issues like unclear vital trigger alerts (CFIR: Innovation trialability and design); staff knowledge of CEDAR functionality and attitudes about functionality, complexity, or relative advantage compared to current systems (CFIR: Inner setting access to knowledge and information and relative advantage); and difficulties incorporating CEDAR into workflow(CFIR: Inner setting/compatibility). Implementation strategies mapped to CFIR determinants included promoting adaptability of customized interface features; staging implementation scale-up with iterative improvements; providing local technical assistance; assessing for readiness and identifying facilitators and barriers; and conducting educational outreach visits. Conclusions: Formative evaluation during pilot studies are a critical step in the process of intervention development and testing. User engagement and frameworks from implementation science can identify issues of feasibility and inform strategies to ensure acceptability prior to intervention scale-up in clinical practice and readiness for conducting larger implementation studies.

Poorly designed Electronic Health Records (EHRs) can lead to increased workloads, decreased efficiency, and medical errors. This cross-sectional study surveyed ICU providers’ perceptions of their EHR using the NASA-TLX and CSUQ surveys. We found that providers report increased temporal demand and effort when using the EHR, and that they are dissatisfied with the EHR’s interface quality.

Abstract Background Intensive care unit (ICU) clinicians encounter frequent challenges with managing vast amounts of fragmented data while caring for multiple critically ill patients simultaneously. This may lead to increased provider cognitive load that may jeopardize patient safety. Objectives This systematic review assesses the impact of centralized multipatient dashboards on ICU clinician performance, perceptions regarding the use of these tools, and patient outcomes. Methods A literature search was conducted on February 9, 2023, using the EBSCO CINAHL, Cochrane Central Register of Controlled Trials, Embase, IEEE Xplore, MEDLINE, Scopus, and Web of Science Core Collection databases. Eligible studies that included ICU clinicians as participants and tested the effect of dashboards designed for use by multiple users to manage multiple patients on user performance and/or satisfaction compared with the standard practice. We narratively synthesized eligible studies following the SWiM (Synthesis Without Meta-analysis) guidelines. Studies were grouped based on dashboard type and outcomes assessed. Results The search yielded a total of 2,407 studies. Five studies met inclusion criteria and were included. Among these, three studies evaluated interactive displays in the ICU, one study assessed two dashboards in the pediatric ICU (PICU), and one study examined centralized monitor in the PICU. Most studies reported several positive outcomes, including reductions in data gathering time before rounds, a decrease in misrepresentations during multidisciplinary rounds, improved daily documentation compliance, faster decision-making, and user satisfaction. One study did not report any significant association. Conclusion The multipatient dashboards were associated with improved ICU clinician performance and were positively perceived in most of the included studies. The risk of bias was high, and the certainty of evidence was very low, due to inconsistencies, imprecision, indirectness in the outcome measure, and methodological limitations. Designing and evaluating multipatient tools using robust research methodologies is an important focus for future research.

Computer vision (CV), a type of artificial intelligence (AI) that uses digital videos or a sequence of images to recognize content, has been used extensively across industries in recent years. However, in the healthcare industry, its applications are limited by factors like privacy, safety, and ethical concerns. Despite this, CV has the potential to improve patient monitoring, and system efficiencies, while reducing workload. In contrast to previous reviews, we focus on the end-user applications of CV. First, we briefly review and categorize CV applications in other industries (job enhancement, surveillance and monitoring, automation, and augmented reality). We then review the developments of CV in the hospital setting, outpatient, and community settings. The recent advances in monitoring delirium, pain and sedation, patient deterioration, mechanical ventilation, mobility, patient safety, surgical applications, quantification of workload in the hospital, and monitoring for patient events outside the hospital are highlighted. To identify opportunities for future applications, we also completed journey mapping at different system levels. Lastly, we discuss the privacy, safety, and ethical considerations associated with CV and outline processes in algorithm development and testing that limit CV expansion in healthcare. This comprehensive review highlights CV applications and ideas for its expanded use in healthcare.

Abstract Objectives Inpatients with language barriers and complex medical needs suffer disparities in quality of care, safety, and health outcomes. Although in-person interpreters are particularly beneficial for these patients, they are underused. We plan to use machine learning predictive analytics to reliably identify patients with language barriers and complex medical needs to prioritize them for in-person interpreters. Materials and methods This qualitative study used stakeholder engagement through semi-structured interviews to understand the perceived risks and benefits of artificial intelligence (AI) in this domain. Stakeholders included clinicians, interpreters, and personnel involved in caring for these patients or for organizing interpreters. Data were coded and analyzed using NVIVO software. Results We completed 49 interviews. Key perceived risks included concerns about transparency, accuracy, redundancy, privacy, perceived stigmatization among patients, alert fatigue, and supply–demand issues. Key perceived benefits included increased awareness of in-person interpreters, improved standard of care and prioritization for interpreter utilization; a streamlined process for accessing interpreters, empowered clinicians, and potential to overcome clinician bias. Discussion This is the first study that elicits stakeholder perspectives on the use of AI with the goal of improved clinical care for patients with language barriers. Perceived benefits and risks related to the use of AI in this domain, overlapped with known hazards and values of AI but some benefits were unique for addressing challenges with providing interpreter services to patients with language barriers. Conclusion Artificial intelligence to identify and prioritize patients for interpreter services has the potential to improve standard of care and address healthcare disparities among patients with language barriers.

Pulmonary auscultation is essential for detecting abnormal lung sounds during physical assessments, but its reliability depends on the operator. Machine learning (ML) models offer an alternative by automatically classifying lung sounds. ML models require substantial data, and public databases aim to address this limitation. This systematic review compares characteristics, diagnostic accuracy, concerns, and data sources of existing models in the literature. Papers published from five major databases between 1990 and 2022 were assessed. Quality assessment was accomplished with a modified QUADAS-2 tool. The review encompassed 62 studies utilizing ML models and public-access databases for lung sound classification. Artificial neural networks (ANN) and support vector machines (SVM) were frequently employed in the ML classifiers. The accuracy ranged from 49.43% to 100% for discriminating abnormal sound types and 69.40% to 99.62% for disease class classification. Seventeen public databases were identified, with the ICBHI 2017 database being the most used (66%). The majority of studies exhibited a high risk of bias and concerns related to patient selection and reference standards. Summarizing, ML models can effectively classify abnormal lung sounds using publicly available data sources. Nevertheless, inconsistent reporting and methodologies pose limitations to advancing the field, and therefore, public databases should adhere to standardized recording and labeling procedures.