Publications

BACKGROUND: This study aimed to demonstrate the feasibility of using computer vision (CV) to unobtrusively extract body motion metrics from videos of emergency medicine (EM) clinicians, and gather validity evidence of these metrics to differentiate POCUS skills between novice and experts, as well as to capture skills gained over time.

METHODS: Prospective cohort study including novice and expert EM clinicians performing echocardiogram (ECHO) and focused assessment with sonography for trauma (FAST) exams on a live simulated patient. Expert observers provided objective structured clinical examination (OSCE) scores (numerical ratings on a scale from 1 to 100), and sonographers' hands and head motion metrics (path length, speed, acceleration, jerk, and smoothness) were extracted via CV using 2-dimensional videos. Data points were captured at baseline, and for novices at baseline and after 12-15 months of residency training.

RESULTS: CV achieved high detection rates (99.52% ECHO, 98.70% FAST). At baseline, experts demonstrated superior OSCE scores (ECHO: 98.6 ± 2.1 vs 63.4 ± 17.0; FAST: 99.2 ± 1.5 vs 68.9 ± 17.7, p < 0.001) and faster task completion (101.8 ± 44.7 vs 240.3 ± 84.1 s, p < 0.001). Experts exhibited smoother hand movements (left hand smoothness: -129.3 ± 47.6 vs -241.3 ± 64.6, p < 0.001) and reduced total path lengths. After 12-15 months of training, novices showed significant improvements in OSCE scores (ECHO: 85.3 ± 10.3; FAST: 84.8 ± 6.5) and task efficiency (134.0 ± 35.6 s), with improvements in motion smoothness and reduced path lengths (p < 0.001). Motion metrics strongly correlated with OSCE scores (r = 0.455-0.783) and task completion time (r = 0.491-0.951).

CONCLUSIONS: CV successfully extracted objective motion metrics that differentiated POCUS skill levels between novices and experts and captured skill development over time. This approach offers a scalable, unobtrusive method for objective POCUS assessment, while supporting competency-based medical education frameworks.

Ultrasound-guided nerve blocks (USGNBs) are increasingly used in the emergency department (ED) as a safe and effective part of multimodal pain management. Their use has been shown to reduce reliance on opioids and procedural sedation, improve pain scores, and enhance functional outcomes for patients. Additionally, USGNBs in the ED have a complication rate of 0.4%, markedly lower than procedural sedation (4-11%), and they significantly reduce opioid requirements, which is critical considering the current opioid epidemic and the risks of persistent opioid use and overdose. Despite these benefits, relevant concerns about medicolegal liability, informed consent, evolving standards of care, may still influence the adoption of USGNBs in clinical practice. To address these issues, this review examines the legal risks associated with USGNBs by drawing on current clinical literature, closed claims data, and case law. We highlight common adverse events such as peripheral nerve injury and local anesthetic systemic toxicity and assess their legal implications. Potential legal risk including liability related to alternatives like opioid use and procedural sedation, are discussed. While the risk of litigation remains low when best practices are followed, failing to offer a USGNB when clearly indicated may increasingly be viewed as a liability if preventable complications occur. This article aims to provide a practical, interdisciplinary framework, including legal risk assessment, training, credentialing, and risk mitigation, to help clinicians, educators, and hospital administrators safely and confidently integrate USGNBs into ED practice.

When given a sample of 100 emergency department discharge instructions, Claude Sonnet, a large language model, produced accurate Spanish translations as evaluated by Spanish-speaking physicians and medical interpreters.

BACKGROUND: Monitoring hemoglobin A1c (HbA1c) is a population health quality metric associated with improved outcomes in diabetic patients. However, the utility of emergency department (ED)-based interventions for monitoring HbA1c for institutional quality metrics is unknown.

OBJECTIVES: The purpose of this study is to evaluate a program to improve measurement of HbA1c for at-risk ED patients.

METHODS: This was a retrospective observational pre-/poststudy at an academic medical center. In November 2017, an alert was programmed into the ED Information System (EDIS) to identify diabetic patients within the health system who were overdue for HbA1c testing and prompt physicians to order HbA1c levels. Data were extracted from the EDIS from January 1, 2017, to September 30, 2023. The primary outcome was placement of an ED order for HbA1c testing. The secondary outcome was change in HbA1c on follow-up testing.

RESULTS: Of 348,490 ED patient encounters during the study period, 13,609 (3.91%) were diabetic patients within the health system who were identified as noncompliant with institutional HbA1c metrics. Following the intervention, HbA1c ordering for target patients increased from 1.5% to 12.1% (p < 0.001), but rates of testing decreased over time. Postintervention, 313 of 1416 patients (23%) who had an HbA1c ordered in the ED had follow-up HbA1c testing available, with a small average decrease in HbA1c of -0.42% (standard deviation 1.7). However, this was not statistically different compared to the preintervention group (-0.60%, p = 0.40).

CONCLUSION: A targeted ED intervention was associated with increased testing for HbA1c; however, this effect decreased over time. The program demonstrated operational feasibility, but did not appear to significantly improve HbA1c levels relative to the overall ED population. Further study is needed to assess optimal approaches to targeted population health interventions in the ED for chronic illnesses like diabetes.

Non-evidence based factors influence post-surgical opioid prescribing practices. Delivering automated near real-time opioid prescribing feedback may encourage providers to prescribe opioid quantities which are more aligned with patient consumption and institutional guidelines. COVID-19 presented unprecedented challenges to healthcare delivery. We observed a substantial deviation in guideline-concordant opioids prescribing during the initial outbreak. However, our institution's pre-existing opioid prescribing feedback system and decision aid may have helped limit the duration and magnitude of the observed deviations by informing prescribers of atypically large opioid prescriptions and encouraging use of institutional data. Combined with provider education, a non-directive decision aid, in the form of near, real-time email feedback, may be an effective mechanism to advance evidence-based opioid prescribing, as it retains flexibility and provider autonomy while encouraging data-driven decision making.

BACKGROUND: Emergency physician-performed ultrasound-guided nerve blocks are becoming more commonplace as an integral component of multimodal analgesia. Ultrasound-guided nerve blocks (UGNBs) can be used to safely and effectively treat pain from commonly encountered pathologies or to facilitate procedures.

OBJECTIVES: There is, currently, variability in the use of nerve blocks in emergency departments (ED) based on training, comfort, and resources; however, it is likely that this method of analgesia will continue to expand into ED practice. In this piece, we review the evidence for several high-utility nerve blocks that have been successfully used in the ED.

DISCUSSION: In this article, we specifically review the superficial cervical plexus block, interscalene brachial plexus nerve block, serratus anterior plane block, erector spinae plane block, pericapsular nerve block, and transgluteal sciatic nerve block.

CONCLUSION: UGNBs are increasingly being incorporated into ED patient care and becoming a critical tool as a part of an opioid reduction strategy. The group of UGNBs listed in this article represent a list of commonly performed blocks and should be considered when offering optimal multimodal analgesia to the acutely injured patient.