Vision Transformer-based Decision Support for Neurosurgical Intervention in Acute Traumatic Brain Injury: Automated Surgical Intervention Support Tool.

Purpose To develop an automated triage tool to predict neurosurgical intervention for patients with traumatic brain injury (TBI). Materials and Methods A provincial trauma registry was reviewed to retrospectively identify patients with TBI from 2005 to 2022 treated at a specialized Canadian trauma center. Model training, validation, and testing were performed using head CT scans with binary reference standard patient-level labels corresponding to whether the patient received neurosurgical intervention. Performance and accuracy of the model, the Automated Surgical Intervention Support Tool for TBI (ASIST-TBI), were also assessed using a held-out consecutive test set of all patients with TBI presenting to the center between March 2021 and September 2022. Results Head CT scans from 2806 patients with TBI (mean age, 57 years ± 22 [SD]; 1955 [70%] men) were acquired between 2005 and 2021 and used for training, validation, and testing. Consecutive scans from an additional 612 patients (mean age, 61 years ± 22; 443 [72%] men) were used to assess the performance of ASIST-TBI. There was accurate prediction of neurosurgical intervention with an area under the receiver operating characteristic curve (AUC) of 0.92 (95% CI: 0.88, 0.94), accuracy of 87% (491 of 562), sensitivity of 87% (196 of 225), and specificity of 88% (295 of 337) on the test dataset. Performance on the held-out test dataset remained robust with an AUC of 0.89 (95% CI: 0.85, 0.91), accuracy of 84% (517 of 612), sensitivity of 85% (199 of 235), and specificity of 84% (318 of 377). Conclusion A novel deep learning model was developed that could accurately predict the requirement for neurosurgical intervention using acute TBI CT scans. Keywords: CT, Brain/Brain Stem, Surgery, Trauma, Prognosis, Classification, Application Domain, Traumatic Brain Injury, Triage, Machine Learning, Decision Support Supplemental material is available for this article. © RSNA, 2024 See also commentary by Haller in this issue.

Non-task expert physicians benefit from correct explainable AI advice when reviewing X-rays.

Artificial intelligence (AI)-generated clinical advice is becoming more prevalent in healthcare. However, the impact of AI-generated advice on physicians' decision-making is underexplored. In this study, physicians received X-rays with correct diagnostic advice and were asked to make a diagnosis, rate the advice's quality, and judge their own confidence. We manipulated whether the advice came with or without a visual annotation on the X-rays, and whether it was labeled as coming from an AI or a human radiologist. Overall, receiving annotated advice from an AI resulted in the highest diagnostic accuracy. Physicians rated the quality of AI advice higher than human advice. We did not find a strong effect of either manipulation on participants' confidence. The magnitude of the effects varied between task experts and non-task experts, with the latter benefiting considerably from correct explainable AI advice. These findings raise important considerations for the deployment of diagnostic advice in healthcare.

Designs, facilitators, barriers, and lessons learned during the implementation of emergency department led virtual urgent care programs in Ontario, Canada.

Introduction: Virtual patient care has seen incredible growth since the beginning of the COVID-19 pandemic. To provide greater access to safe and timely urgent care, in the fall of 2020, the Ministry of Health introduced a pilot program of 14 virtual urgent care (VUC) initiatives across the province of Ontario. The objective of this paper was to describe the overall design, facilitators, barriers, and lessons learned during the implementation of seven emergency department (ED) led VUC pilot programs in Ontario, Canada. Methods: We assembled an expert panel of 13 emergency medicine physicians and researchers with experience leading and implementing local VUC programs. Each VUC program lead was asked to describe their local pilot program, share common facilitators and barriers to adoption of VUC services, and summarize lessons learned for future VUC design and development. Results: Models of care interventions varied across VUC pilot programs related to triage, staffing, technology, and physician remuneration. Common facilitators included local champions to guide program delivery, provincial funding support, and multi-modal marketing and promotions. Common barriers included behaviour change strategies to support adoption of a new service, access to high-quality information technology to support new workflow models that consider privacy, risk, and legal perspectives, and standardized data collection which underpin overall objective impact assessments. Conclusions: These pilot programs were rapidly implemented to support safe access to care and ED diversion of patients with low acuity issues during the COVID-19 pandemic. The heterogeneity of program implementation respects local autonomy yet may present challenges for sustainability efforts and future funding considerations.

The Anticoagulated trauma patient in the age of the direct oral anticoagulants: a Canadian perspective.

BACKGROUND: The anticoagulated trauma patient presents a particular challenge to the critical care physician. Our understanding of these patients is defined and extrapolated by experience with patients on warfarin pre-injury. Today, many patients who would have been on warfarin are now prescribed the Direct Oral Anticoagulants (DOACs) a class of anticoagulants with entirely different mechanisms of action, effects on routine coagulation assays and approach to reversal. METHODS: Trauma registry data from Toronto's (Ontario, Canada) two Level 1 trauma centres were used to identify patients on oral anticoagulation pre-injury from June 1, 2014 to June 1, 2015. The trauma registry and medical records were reviewed and used to extract demographic and clinical data. RESULTS: We found 81 patients were on oral anticoagulants pre-injury representing 3.2% of the total trauma population and 33% of the orally anticoagulated patients were prescribed a DOAC prior to presentation. Comparison between the DOAC and warfarin groups showed similar age, mechanisms of injury, indications for anticoagulation, injury severity score and rate of intracranial hemorrhage. Patients on DOACs had higher initial mean hemoglobin vs warfarin (131 vs 120) and lower serum creatinine (94.8 vs 129.5). The percentage of patients receiving a blood transfusion in the trauma bay and total in-hospital transfusion was similar between the two groups however patients on DOACs were more likely to receive tranexamic acid vs patients on warfarin (32.1% vs 9.1%) and less likely to receive prothrombin concentrates (18.5% vs 60%). Patients on DOACs were found to have higher survival to discharge (92%) vs patients on warfarin (72%). CONCLUSION: Patients on DOACs pre-injury now represent a significant proportion of the anticoagulated trauma population. Although they share demographic and clinical similarities with patients on warfarin, patients on DOACs may have improved outcomes despite lack of established drug reversal protocols and challenging interpretation of coagulation assays. LEVEL OF EVIDENCE: III; Study Type: Retrospective Review.

Communication between nurses and physicians: strategies to surviving in the emergency department trenches.

The emergency department (ED) is a challenging and stressful work environment where communication lapses can lead to negative health outcomes. This article offers strategies to Emergency Medicine residents, nurses and staff physicians on how to improve communication to optimize patient care.

Tranexamic acid in pediatric trauma: why not?

Trauma is a leading cause of death in pediatrics. Currently, no medical treatment exists to reduce mortality in the setting of pediatric trauma; however, this evidence does exist in adults. Bleeding and coagulopathy after trauma increases mortality in both adults and children. Clinical research has demonstrated a reduction in mortality with early use of tranexamic acid in adult trauma patients in both civilian and military settings. Tranexamic acid used in the perioperative setting safely reduces transfusion requirements in children. This article compares the hematologic response to trauma between children and adults, and explores the potential use of tranexamic acid in pediatric hemorrhagic trauma.

Violence in Canadian amateur hockey: the experience of referees in Ontario.

OBJECTIVE: To determine the perceptions and roles of referees about violence and injury in hockey games. DESIGN: Questionnaire. SETTING: Web-based survey. PARTICIPANTS: We contacted referees across Canada from various leagues and all levels of play, with the majority of respondents from Ontario (92%). MAIN OUTCOME MEASURES: We gathered demographic information anonymously and posed questions on aggression and experience in hockey games. RESULTS: The majority of referees (n = 632) indicated that violence is a serious concern to both players and referees at all levels of hockey. More than 90% of referees responded that they were the recipients of aggression and anger (92.1%, 95% confidence interval, 90.0-94.2), 55% had been involved in hockey games where aggressive behavior resulted in the referee losing control of the game, and 71% said that this increased aggression leads to injury. Referees' opinions are that the coach is the most responsible for managing on-ice safety (63%). To improve hockey safety, referees suggest education and more rigorous enforcement of discipline for all participants. CONCLUSIONS: Referees are important for hockey safety and need to be appropriately supported. Referees believe that increased aggression can lead to injury and that rules need to be enforced more diligently. Referees recommend that increased education about safety is needed to guide parents, coaches, and players to make hockey safer.

Bicyclist deaths and striking vehicles in the USA.

OBJECTIVES: Bicycling is a popular means of transportation that is sometimes associated with injury from collisions. The authors analysed national data for the USA to evaluate bicyclist deaths associated with motor vehicle impacts. METHODS: The authors conducted a population-based case-control analysis of road deaths reported by the National Highway Traffic Safety Administration. The authors included bicyclist deaths from 1 January 2008 to 31 December 2008 (cases), along with the non-bicyclist road deaths immediately before and after the bicyclist death in the same state (controls). Analyses also included linkages to auto appraisal websites to estimate type, size and cost of the motor vehicle involved in each death. RESULTS: A total of 711 bicyclist deaths were included, equivalent to a rate of 2 deaths per million population annually. No state had a rate statistically significantly below the national average whereas Florida was a high outlier with three times the national rate (p<0.001). The typical bicyclist who died was a man travelling in the afternoon or evening. The average estimated resale value of the involved motor vehicle was about one-third higher for bicyclist deaths than control deaths (US$10 603 vs US$8118, p<0.001). Analyses based on median estimated resale value and luxury resale value yielded similar findings. Stratified analyses based on demographics, time and posted speed limits yielded similar discrepancies. Larger motor vehicles were particularly common in bicyclist deaths compared to control deaths, especially freight trucks (11% vs 8%, p=0.008) and large automobiles (43% vs 37%, p=0.004). Conversely, motorcycles were distinctly infrequent in bicyclist deaths compared to control deaths (1% vs 14%, p<0.001). CONCLUSIONS: Large expensive motor vehicles account for a disproportionate share of bicyclist deaths. Bicyclists, motorists, policy-makers and vehicle manufacturers need to consider more imaginative solutions to help prevent future deaths.