3 results
LO32: Artificial intelligence to predict disposition to improve flow in the emergency department
- L. Grant, X. Xue, Z. Vajihi, A. Azuelos, S. Rosenthal, D. Hopkins, R. Aroutiunian, B. Unger, A. Guttman, M. Afilalo
-
- Journal:
- Canadian Journal of Emergency Medicine / Volume 22 / Issue S1 / May 2020
- Published online by Cambridge University Press:
- 13 May 2020, pp. S18-S19
- Print publication:
- May 2020
-
- Article
-
- You have access Access
- Export citation
-
Introduction: Emergency department (ED) crowding is a major problem across Canada. We studied the ability of artificial intelligence methods to improve patient flow through the ED by predicting patient disposition using information available at triage and shortly after patients’ arrival in the ED. Methods: This retrospective study included all visits to an urban, academic, adult ED between May 2012 and June 2019. For each visit, 489 variables were extracted including triage data that had been collected for use in the Canadian Triage Assessment Scale (CTAS) and information regarding laboratory tests, radiological tests, consultations and admissions. A training set consisting of all visits from April 2012 up to December 2018 was used to train 5 classes of machine learning models to predict admission to the hospital from the ED. The models were trained to predict admission at the time of the patient's arrival in the ED and every 30 minutes after arrival until 6 hours into their ED stay. The performance of models was compared using the area under the ROC curve (AUC) on a test set consisting of all visits from January 2019 to June 2019. Results: The study included 536,332 visits and the admission rate was 15.0%. Gradient boosting models generally outperformed other machine learning models. A gradient boosting model using all available data at 2 hours after patient arrival in the ED yielded a test set AUC 0.92 [95% CI 0.91-0.93], while a model using only data available at triage yielded an AUC 0.90 [95% CI 0.89-0.91]. The quality of predictions generally improved as predictions were made later in the patient's ED stay leading to an AUC 0.95 [95% CI 0.93-0.96] at 6 hours after arrival. A gradient boosting model with 20 variables available at 2 hours after patient arrival in the ED yielded an AUC 0.91 [95% CI 0.89-0.93]. A gradient boosting model that makes predictions at 2 hours after arrival in ED using only variables that are available at all EDs in the province of Quebec yielded an AUC 0.91 [95% 0.89-0.92]. Conclusion: Machine learning can predict admission to a hospital from the ED using variables that area collected as part of routine ED care. Machine learning tools may potentially be used to help ED physicians to make faster and more appropriate disposition decisions, to decrease unnecessary testing and alleviate ED crowding.
P001: Age-adjusted D-dimer and two-site compression point of care ultrasonography to rule out acute deep vein thrombosis - a pilot study
- K. Alqaydi, J. Turner, L. Robichaud, D. Hamad, X. Xue, M. Afilalo
-
- Journal:
- Canadian Journal of Emergency Medicine / Volume 20 / Issue S1 / May 2018
- Published online by Cambridge University Press:
- 11 May 2018, p. S57
- Print publication:
- May 2018
-
- Article
-
- You have access Access
- Export citation
-
Introduction: Deep vein thrombosis (DVT) can lead to significant morbidity and mortality if not diagnosed and treated promptly. Currently, few methods aside from venous duplex scanning can rule out DVT in patients presenting to the Emergency Department (ED). Current screening tools, including the use of the subjective Wells score, frequently leads to unnecessary investigations and anticoagulation. In this study, we sought to determine whether two-site compression point-of-care ultrasound (POCUS) combined with a negative age-adjusted D-dimer test can accurately rule out DVT in ED patients irrespective of the modified Wells score. Methods: This is a single-center, prospective observational study in the ED of the Jewish General Hospital in Montreal. We are recruiting a convenience sample of patients presenting to the ED with symptoms suggestive of DVT. All enrolled patients are risk-stratified using the modified Wells criteria for DVT, then undergo two-site compression POCUS, and testing for age-adjusted D-dimer. Patients with DVT unlikely according to modified Wells score, negative POCUS and negative age-adjusted D-dimer are discharged home and receive a three-month phone follow-up. Patients with DVT likely according to modified Wells score, a positive POCUS or a positive age-adjusted D-dimer, will undergo a venous duplex scan. A true negative DVT is defined as either a negative venous duplex scan or a negative follow-up phone questionnaire for patients who were sent home without a venous duplex scan. Results: Of the 42 patients recruited thus far, the mean age is 56 years old and 42.8% are male. Twelve (28.6%) patients had DVT unlikely as per modified Wells score, negative POCUS and negative age-adjusted D-dimer and were discharged home. None of these patients developed a DVT on three-month follow-up. Thirty patients (71.4%) had either a DVT likely as per modified Wells score, a positive POCUS or a positive age-adjusted D-dimer and underwent a venous duplex scan. Of those, six patients had a confirmed DVT (3 proximal & 3 distal). POCUS detected all proximal DVTs, while combined POCUS and age-adjusted D-dimer detected all proximal and distal DVTs. None of the patients with a negative POCUS and age-adjusted D-dimer were found to have a DVT. Conclusion: Two-site compression POCUS combined with a negative age-adjusted D-dimer test appears to accurately rule out DVT in ED patients without the need for follow-up duplex venous scan. Using this approach would alleviate the need to calculate the Wells score, and also reduce the need for radiology-performed duplex venous scan for many patients.
GD01: Age-adjusted D-dimer and two-site compression point-of-care ultrasonography to rule out acute deep vein thrombosis
- K. Alqaydi, L. Robichaud, D. Hamad, X. Xue, J. Turner, M. Afilalo
-
- Journal:
- Canadian Journal of Emergency Medicine / Volume 19 / Issue S1 / May 2017
- Published online by Cambridge University Press:
- 15 May 2017, p. S62
- Print publication:
- May 2017
-
- Article
-
- You have access Access
- Export citation
-
Introduction: Undiagnosed deep vein thrombosis (DVT) can lead to significant morbidity and mortality, including death from DVT-associated massive pulmonary embolism (PE). While several validated clinical prediction rules, blood test and imaging modalities exist to investigate a potential DVT, there is currently a lack of rapid, accessible and reliable methods to exclude the possibility of DVT without resorting to formal venous duplex scanning. Currently, the use in the ED of a validated clinical prediction rule combined to either a high-sensitivity D-dimer test or ultrasonography of the lower extremities has a poor predictive value, as 75-90% of patients suspected of DVT have a negative formal venous duplex scan. Compression bedside ultrasound has however recently been shown to be a safe, rapid and accurate method for the diagnosis of proximal DVT in the emergency department with a high sensitivity and specificity (combined sensitivity and specificity of 96.1% and 96.8%, respectively1). Research question: In the present study, we will primarily assess whether two-site compression POCUS combined with a negative age-adjusted D-dimer test can accurately rule out DVT in ED patients regardless of the Wells criteria. Methods: This is a single-center, prospective, observational study carried out over one year in the Emergency Department of the Jewish General Hospital in Montreal, Quebec. We aim to enroll a convenience sample of 475 patients aged 18 years and older presenting to the ED with symptoms suggestive of a DVT. All enrolled patients will receive the standard of care required for a lower leg DVT presentation. After calculating Patients DVT risk using modified wells criteria, all patients will undergo POCUS for DVT followed by a D-dimer test. Based on their results, patients will either undergo formal duplex scanning, or will be discharged without further testing and receive a three-month phone follow-up. A true negative lower leg DVT will be defined as follows: (1) Negative follow-up phone questionnaire for patients who were sent home with no formal duplex venous scanning. (2) Negative formal duplex venous scanning for patients who were deemed likely to have lower leg DVT using the Wells score, with a negative D-dimer and POCUS. Age adjusted DVT was added to account for below knee DVT and avoid the need for patients to return for fellow up duplex study in 1 week. To estimate our technique’s sensitivity with a 4% margin of error with 95% confidence intervals, 92 confirmed DVT patients are needed. We expect to recruit a total 475 patients within one-year period at the JGH (95 DVT-positive patients and 380 DVT-negative patients). Impact: The use of compression bedside ultrasound with a negative age-adjusted D-dimer test to rule out DVT in the ED may accelerate the decision regarding patient disposition and significantly decrease the length of patient stay in the ED. In addition, it may help avoid unnecessary medical interventions and diagnostic tests, thus representing potential quality of care and cost-saving improvements as well.