Outcomes and Cost-Effectiveness of an EHR-Embedded AI Screener for Identifying Hospitalized Adults at Risk for Opioid Use Disorder.

Hospitalized adults with opioid use disorder (OUD) are at high risk for adverse events and rehospitalizations. This pre-post quasi-experimental study evaluated whether an AI-driven OUD screener embedded in the electronic health record (EHR) was non-inferior to usual care in identifying patients for Addiction Medicine consults, aiming to provide a similarly effective but more scalable alternative to human-led ad hoc consultations. The AI screener analyzed EHR notes in real-time with a convolutional neural network to identify patients at risk and recommend consultation. The primary outcome was the proportion of patients receiving consults, comparing a 16-month pre-intervention period to an 8-month post-intervention period with the AI screener. Consults did not change between periods (1.35% vs 1.51%, p < 0.001 for non-inferiority). The AI screener was associated with a reduction in 30-day readmissions (OR: 0.53, 95% CI: 0.30-0.91, p = 0.02) with an incremental cost of $6,801 per readmission avoided, demonstrating its potential as a scalable, cost-effective solution for OUD care. ClinicalTrials.gov ID : NCT05745480.

Prompt engineering with a large language model to assist providers in responding to patient inquiries: a real-time implementation in the electronic health record.

Background: Large language models (LLMs) can assist providers in drafting responses to patient inquiries. We examined a prompt engineering strategy to draft responses for providers in the electronic health record. The aim was to evaluate the change in usability after prompt engineering. Materials and Methods: A pre-post study over 8 months was conducted across 27 providers. The primary outcome was the provider use of LLM-generated messages from Generative Pre-Trained Transformer 4 (GPT-4) in a mixed-effects model, and the secondary outcome was provider sentiment analysis. Results: Of the 7605 messages generated, 17.5% (n = 1327) were used. There was a reduction in negative sentiment with an odds ratio of 0.43 (95% CI, 0.36-0.52), but message use decreased (P < .01). The addition of nurses after the study period led to an increase in message use to 35.8% (P < .01). Discussion: The improvement in sentiment with prompt engineering suggests better content quality, but the initial decrease in usage highlights the need for integration with human factors design. Conclusion: Future studies should explore strategies for optimizing the integration of LLMs into the provider workflow to maximize both usability and effectiveness.

Call me Dr Ishmael: trends in electronic health record notes available at emergency department visits and admissions.

Objectives: Numerous studies have identified information overload as a key issue for electronic health records (EHRs). This study describes the amount of text data across all notes available to emergency physicians in the EHR, trended over the time since EHR establishment. Materials and Methods: We conducted a retrospective analysis of EHR data from a large healthcare system, examining the number of notes and a corresponding number of total words and total tokens across all notes available to physicians during patient encounters in the emergency department (ED). We assessed the change in these metrics over a 17-year period between 2006 and 2023. Results: The study cohort included 730 968 ED visits made by 293 559 unique patients and a total note count of 132 574 964. The median note count for all encounters in 2006 was 5 (IQR 1-16), accounting for 1735 (IQR 447-5521) words. By the last full year of the study period, 2022, the median number of notes had grown to 359 (IQR 84-943), representing 359 (IQR 84-943) words. Note and word counts were higher for admitted patients. Discussion: The volume of notes available for review by providers has increased by over 30-fold in the 17 years since the implementation of the EHR at a large health system. The task of reviewing these notes has become commensurately more difficult. These data point to the critical need for new strategies and tools for filtering, synthesizing, and summarizing information to achieve the promise of the medical record.

Dashboarding to Monitor Machine-Learning-Based Clinical Decision Support Interventions.

BACKGROUND: Existing monitoring of machine-learning-based clinical decision support (ML-CDS) is focused predominantly on the ML outputs and accuracy thereof. Improving patient care requires not only accurate algorithms but also systems of care that enable the output of these algorithms to drive specific actions by care teams, necessitating expanding their monitoring. OBJECTIVES: In this case report, we describe the creation of a dashboard that allows the intervention development team and operational stakeholders to govern and identify potential issues that may require corrective action by bridging the monitoring gap between model outputs and patient outcomes. METHODS: We used an iterative development process to build a dashboard to monitor the performance of our intervention in the broader context of the care system. RESULTS: Our investigation of best practices elsewhere, iterative design, and expert consultation led us to anchor our dashboard on alluvial charts and control charts. Both the development process and the dashboard itself illuminated areas to improve the broader intervention. CONCLUSION: We propose that monitoring ML-CDS algorithms with regular dashboards that allow both a context-level view of the system and a drilled down view of specific components is a critical part of implementing these algorithms to ensure that these tools function appropriately within the broader care system.

Effectiveness of an Emergency Department-Based Machine Learning Clinical Decision Support Tool to Prevent Outpatient Falls Among Older Adults: Protocol for a Quasi-Experimental Study.

BACKGROUND: Emergency department (ED) providers are important collaborators in preventing falls for older adults because they are often the first health care providers to see a patient after a fall and because at-home falls are often preceded by previous ED visits. Previous work has shown that ED referrals to falls interventions can reduce the risk of an at-home fall by 38%. Screening patients at risk for a fall can be time-consuming and difficult to implement in the ED setting. Machine learning (ML) and clinical decision support (CDS) offer the potential of automating the screening process. However, it remains unclear whether automation of screening and referrals can reduce the risk of future falls among older patients. OBJECTIVE: The goal of this paper is to describe a research protocol for evaluating the effectiveness of an automated screening and referral intervention. These findings will inform ongoing discussions about the use of ML and artificial intelligence to augment medical decision-making. METHODS: To assess the effectiveness of our program for patients receiving the falls risk intervention, our primary analysis will be to obtain referral completion rates at 3 different EDs. We will use a quasi-experimental design known as a sharp regression discontinuity with regard to intent-to-treat, since the intervention is administered to patients whose risk score falls above a threshold. A conditional logistic regression model will be built to describe 6-month fall risk at each site as a function of the intervention, patient demographics, and risk score. The odds ratio of a return visit for a fall and the 95% CI will be estimated by comparing those identified as high risk by the ML-based CDS (ML-CDS) and those who were not but had a similar risk profile. RESULTS: The ML-CDS tool under study has been implemented at 2 of the 3 EDs in our study. As of April 2023, a total of 1326 patient encounters have been flagged for providers, and 339 unique patients have been referred to the mobility and falls clinic. To date, 15% (45/339) of patients have scheduled an appointment with the clinic. CONCLUSIONS: This study seeks to quantify the impact of an ML-CDS intervention on patient behavior and outcomes. Our end-to-end data set allows for a more meaningful analysis of patient outcomes than other studies focused on interim outcomes, and our multisite implementation plan will demonstrate applicability to a broad population and the possibility to adapt the intervention to other EDs and achieve similar results. Our statistical methodology, regression discontinuity design, allows for causal inference from observational data and a staggered implementation strategy allows for the identification of secular trends that could affect causal associations and allow mitigation as necessary. TRIAL REGISTRATION: ClinicalTrials.gov NCT05810064; https://www.clinicaltrials.gov/study/NCT05810064. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): DERR1-10.2196/48128.

Deployment of Real-time Natural Language Processing and Deep Learning Clinical Decision Support in the Electronic Health Record: Pipeline Implementation for an Opioid Misuse Screener in Hospitalized Adults.

BACKGROUND: The clinical narrative in electronic health records (EHRs) carries valuable information for predictive analytics; however, its free-text form is difficult to mine and analyze for clinical decision support (CDS). Large-scale clinical natural language processing (NLP) pipelines have focused on data warehouse applications for retrospective research efforts. There remains a paucity of evidence for implementing NLP pipelines at the bedside for health care delivery. OBJECTIVE: We aimed to detail a hospital-wide, operational pipeline to implement a real-time NLP-driven CDS tool and describe a protocol for an implementation framework with a user-centered design of the CDS tool. METHODS: The pipeline integrated a previously trained open-source convolutional neural network model for screening opioid misuse that leveraged EHR notes mapped to standardized medical vocabularies in the Unified Medical Language System. A sample of 100 adult encounters were reviewed by a physician informaticist for silent testing of the deep learning algorithm before deployment. An end user interview survey was developed to examine the user acceptability of a best practice alert (BPA) to provide the screening results with recommendations. The planned implementation also included a human-centered design with user feedback on the BPA, an implementation framework with cost-effectiveness, and a noninferiority patient outcome analysis plan. RESULTS: The pipeline was a reproducible workflow with a shared pseudocode for a cloud service to ingest, process, and store clinical notes as Health Level 7 messages from a major EHR vendor in an elastic cloud computing environment. Feature engineering of the notes used an open-source NLP engine, and the features were fed into the deep learning algorithm, with the results returned as a BPA in the EHR. On-site silent testing of the deep learning algorithm demonstrated a sensitivity of 93% (95% CI 66%-99%) and specificity of 92% (95% CI 84%-96%), similar to published validation studies. Before deployment, approvals were received across hospital committees for inpatient operations. Five interviews were conducted; they informed the development of an educational flyer and further modified the BPA to exclude certain patients and allow the refusal of recommendations. The longest delay in pipeline development was because of cybersecurity approvals, especially because of the exchange of protected health information between the Microsoft (Microsoft Corp) and Epic (Epic Systems Corp) cloud vendors. In silent testing, the resultant pipeline provided a BPA to the bedside within minutes of a provider entering a note in the EHR. CONCLUSIONS: The components of the real-time NLP pipeline were detailed with open-source tools and pseudocode for other health systems to benchmark. The deployment of medical artificial intelligence systems in routine clinical care presents an important yet unfulfilled opportunity, and our protocol aimed to close the gap in the implementation of artificial intelligence-driven CDS. TRIAL REGISTRATION: ClinicalTrials.gov NCT05745480; https://www.clinicaltrials.gov/ct2/show/NCT05745480.

Multisite evaluation of prediction models for emergency department crowding before and during the COVID-19 pandemic.

OBJECTIVE: To develop a machine learning framework to forecast emergency department (ED) crowding and to evaluate model performance under spatial and temporal data drift. MATERIALS AND METHODS: We obtained 4 datasets, identified by the location: 1-large academic hospital and 2-rural hospital, and time period: pre-coronavirus disease (COVID) (January 1, 2019-February 1, 2020) and COVID-era (May 15, 2020-February 1, 2021). Our primary target was a binary outcome that is equal to 1 if the number of patients with acute respiratory illness that were ED boarding for more than 4 h was above a prescribed historical percentile. We trained a random forest and used the area under the curve (AUC) to evaluate out-of-sample performance for 2 experiments: (1) we evaluated the impact of sudden temporal drift by training models using pre-COVID data and testing them during the COVID-era, (2) we evaluated the impact of spatial drift by testing models trained at location 1 on data from location 2, and vice versa. RESULTS: The baseline AUC values for ED boarding ranged from 0.54 (pre-COVID at location 2) to 0.81 (COVID-era at location 1). Models trained with pre-COVID data performed similarly to COVID-era models (0.82 vs 0.78 at location 1). Models that were transferred from location 2 to location 1 performed worse than models trained at location 1 (0.51 vs 0.78). DISCUSSION AND CONCLUSION: Our results demonstrate that ED boarding is a predictable metric for ED crowding, models were not significantly impacted by temporal data drift, and any attempts at implementation must consider spatial data drift.

Operationalizing a real-time scoring model to predict fall risk among older adults in the emergency department.

Predictive models are increasingly being developed and implemented to improve patient care across a variety of clinical scenarios. While a body of literature exists on the development of models using existing data, less focus has been placed on practical operationalization of these models for deployment in real-time production environments. This case-study describes challenges and barriers identified and overcome in such an operationalization for a model aimed at predicting risk of outpatient falls after Emergency Department (ED) visits among older adults. Based on our experience, we provide general principles for translating an EHR-based predictive model from research and reporting environments into real-time operation.

Governance of Clinical AI applications to facilitate safe and equitable deployment in a large health system: Key elements and early successes.

One of the key challenges in successful deployment and meaningful adoption of AI in healthcare is health system-level governance of AI applications. Such governance is critical not only for patient safety and accountability by a health system, but to foster clinician trust to improve adoption and facilitate meaningful health outcomes. In this case study, we describe the development of such a governance structure at University of Wisconsin Health (UWH) that provides oversight of AI applications from assessment of validity and user acceptability through safe deployment with continuous monitoring for effectiveness. Our structure leverages a multi-disciplinary steering committee along with project specific sub-committees. Members of the committee formulate a multi-stakeholder perspective spanning informatics, data science, clinical operations, ethics, and equity. Our structure includes guiding principles that provide tangible parameters for endorsement of both initial deployment and ongoing usage of AI applications. The committee is tasked with ensuring principles of interpretability, accuracy, and fairness across all applications. To operationalize these principles, we provide a value stream to apply the principles of AI governance at different stages of clinical implementation. This structure has enabled effective clinical adoption of AI applications. Effective governance has provided several outcomes: (1) a clear and institutional structure for oversight and endorsement; (2) a path towards successful deployment that encompasses technologic, clinical, and operational, considerations; (3) a process for ongoing monitoring to ensure the solution remains acceptable as clinical practice and disease prevalence evolve; (4) incorporation of guidelines for the ethical and equitable use of AI applications.

Collaborative design and implementation of a clinical decision support system for automated fall-risk identification and referrals in emergency departments.

Of the 3 million older adults seeking fall-related emergency care each year, nearly one-third visited the Emergency Department (ED) in the previous 6 months. ED providers have a great opportunity to refer patients for fall prevention services at these initial visits, but lack feasible tools for identifying those at highest-risk. Existing fall screening tools have been poorly adopted due to ED staff/provider burden and lack of workflow integration. To address this, we developed an automated clinical decision support (CDS) system for identifying and referring older adult ED patients at risk of future falls. We engaged an interdisciplinary design team (ED providers, health services researchers, information technology/predictive analytics professionals, and outpatient Falls Clinic staff) to collaboratively develop a system that successfully met user requirements and integrated seamlessly into existing ED workflows. Our rapid-cycle development and evaluation process employed a novel combination of human-centered design, implementation science, and patient experience strategies, facilitating simultaneous design of the CDS tool and intervention implementation strategies. This included defining system requirements, systematically identifying and resolving usability problems, assessing barriers and facilitators to implementation (e.g., data accessibility, lack of time, high patient volumes, appointment availability) from multiple vantage points, and refining protocols for communicating with referred patients at discharge. ED physician, nurse, and patient stakeholders were also engaged through online surveys and user testing. Successful CDS design and implementation required integration of multiple new technologies and processes into existing workflows, necessitating interdisciplinary collaboration from the onset. By using this iterative approach, we were able to design and implement an intervention meeting all project goals. Processes used in this Clinical-IT-Research partnership can be applied to other use cases involving automated risk-stratification, CDS development, and EHR-facilitated care coordination.