Pub Date : 2023-09-01DOI: 10.1007/s43678-023-00574-3
Francois Gravel, Valérie Bélanger, Sophie Gosselin
{"title":"Offload ambulance delays: a small piece of a bigger puzzle.","authors":"Francois Gravel, Valérie Bélanger, Sophie Gosselin","doi":"10.1007/s43678-023-00574-3","DOIUrl":"10.1007/s43678-023-00574-3","url":null,"abstract":"","PeriodicalId":55286,"journal":{"name":"Canadian Journal of Emergency Medicine","volume":null,"pages":null},"PeriodicalIF":2.4,"publicationDate":"2023-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"10237088","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-01DOI: 10.1007/s43678-023-00582-3
{"title":"Global Research Highlights.","authors":"","doi":"10.1007/s43678-023-00582-3","DOIUrl":"10.1007/s43678-023-00582-3","url":null,"abstract":"","PeriodicalId":55286,"journal":{"name":"Canadian Journal of Emergency Medicine","volume":null,"pages":null},"PeriodicalIF":2.4,"publicationDate":"2023-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"10274557","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-01Epub Date: 2023-08-03DOI: 10.1007/s43678-023-00555-6
Jessica Poliwoda, Debra Eagles, Krishan Yadav, Marie-Joe Nemnom, Charlotte Grace Walmsley, Lisa Mielniczuk, Ian G Stiell
Background: Acute heart failure is a serious condition commonly seen in the emergency department (ED). The HEARTRISK6 Scale has been recently developed to identify the risk of poor outcomes but has not been tested. We sought to describe the management and outcomes of ED patients with acute heart failure and to evaluate the potential impact of the HEARTRISK6 Scale.
Methods: We conducted a health records review of 300 consecutive acute heart failure patients presenting to two tertiary care EDs. Two evaluators abstracted clinical variables, ED management and treatment details, and patient outcomes using the electronic health records platform (EPIC) and attending physicians verified the data. The primary outcome measure was a short-term serious outcome, as shown in Results. In addition, the HEARTRISK6 score was calculated retrospectively.
Results: We included 300 patients with mean age of 78.5 years, 51.0% male, 56.3% arrival by ambulance, and 67.0% admitted to hospital. 25.3% experienced a short-term serious outcome 1) after admission (N = 201): non-invasive ventilation 14.9%, intubation 1.5%, major cardiac procedure 5.0%, myocardial infarction 2.0%, death 8.5%; 2) after ED discharge (N = 99): return to ED 21.2%, death 4.0%. Those initially admitted experienced a much higher proportion of serious outcomes compared to those discharged (29.9% vs. 16.2%). A HEARTRISK6 Scale cut-point score of ≥ 1 would have had a sensitivity of 91.0%, specificity 24.5%, and negative likelihood ratio 0.37 for short-term serious outcomes and suggested hospital admission for 80.7% of cases.
Conclusion: There was a large range of severity of illness of acute heart failure patients and a wide variety of treatments were administered in the ED. Both admitted and discharged patients experienced a high proportion of poor outcomes. The HEARTRISK6 Scale showed a high sensitivity for short-term serious outcomes but with the potential to increase hospital admissions. Further validation of the HEARTRISK6 Scale is required before routine clinical use.
{"title":"Outcomes of acute heart failure patients managed in the emergency department.","authors":"Jessica Poliwoda, Debra Eagles, Krishan Yadav, Marie-Joe Nemnom, Charlotte Grace Walmsley, Lisa Mielniczuk, Ian G Stiell","doi":"10.1007/s43678-023-00555-6","DOIUrl":"10.1007/s43678-023-00555-6","url":null,"abstract":"<p><strong>Background: </strong>Acute heart failure is a serious condition commonly seen in the emergency department (ED). The HEARTRISK6 Scale has been recently developed to identify the risk of poor outcomes but has not been tested. We sought to describe the management and outcomes of ED patients with acute heart failure and to evaluate the potential impact of the HEARTRISK6 Scale.</p><p><strong>Methods: </strong>We conducted a health records review of 300 consecutive acute heart failure patients presenting to two tertiary care EDs. Two evaluators abstracted clinical variables, ED management and treatment details, and patient outcomes using the electronic health records platform (EPIC) and attending physicians verified the data. The primary outcome measure was a short-term serious outcome, as shown in Results. In addition, the HEARTRISK6 score was calculated retrospectively.</p><p><strong>Results: </strong>We included 300 patients with mean age of 78.5 years, 51.0% male, 56.3% arrival by ambulance, and 67.0% admitted to hospital. 25.3% experienced a short-term serious outcome 1) after admission (N = 201): non-invasive ventilation 14.9%, intubation 1.5%, major cardiac procedure 5.0%, myocardial infarction 2.0%, death 8.5%; 2) after ED discharge (N = 99): return to ED 21.2%, death 4.0%. Those initially admitted experienced a much higher proportion of serious outcomes compared to those discharged (29.9% vs. 16.2%). A HEARTRISK6 Scale cut-point score of ≥ 1 would have had a sensitivity of 91.0%, specificity 24.5%, and negative likelihood ratio 0.37 for short-term serious outcomes and suggested hospital admission for 80.7% of cases.</p><p><strong>Conclusion: </strong>There was a large range of severity of illness of acute heart failure patients and a wide variety of treatments were administered in the ED. Both admitted and discharged patients experienced a high proportion of poor outcomes. The HEARTRISK6 Scale showed a high sensitivity for short-term serious outcomes but with the potential to increase hospital admissions. Further validation of the HEARTRISK6 Scale is required before routine clinical use.</p>","PeriodicalId":55286,"journal":{"name":"Canadian Journal of Emergency Medicine","volume":null,"pages":null},"PeriodicalIF":2.4,"publicationDate":"2023-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"10568155","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-01Epub Date: 2023-05-19DOI: 10.1007/s43678-023-00521-2
I E Blanchard, T S Williamson, B E Hagel, D J Niven, D J Lane, S Dean, M N Shah, E S Lang, C J Doig
Objective: To address an important care issue in Canada, we tested the association between paramedic system hospital offload and response time, while considering the impact of other system-level factors.
Methods: Data from Calgary, Alberta (2014-2017), included median offload (exposure) and response (outcome) time aggregated by hour, with covariates paramedic system episodes of care-dispatch and arrival of a response unit-and hospital transport arrivals (collectively called volume), time of day, and season. Analyses used linear regression and modified Poisson models.
Results: 301,105 EMS episodes of care over 26,193 1-h periods were included. For any given 1-h period, the median (IQR) across all episodes of care for offload time, response time, episodes of care, and hospital transport arrivals were 55.3 (45.7, 66.3) min, 8.6 (7.6, 9.8) min, 12 (8, 16) episodes, and 8 (5, 10) hospital arrivals, respectively. Multivariable modelling revealed a complex association differing over levels of exposure and covariates, requiring description using "light stress" and "heavy stress" system scenarios. The light scenario was defined as median offload of 30 min and volume < 10th percentile (six episodes and four hospital arrivals), in the summer, and the heavy scenario as median offload of 90 min and volume > 90th percentile (17 episodes and 13 hospital arrivals), in the winter. An increase is reported in minutes:seconds for median hourly response time between scenarios by time of day: 1:04-4:16 (0000-0559 h.), 0:42-2:05 (0600-1159 h.), 0:57-3:01 (1200-1759 h.), and 0:18-2:21 (1800-2359 h.).
Conclusions: Increasing offload is associated with increased response time; however the relationship is complex, with a greater impact on response time noted in select situations such as high volume in the winter. These observations illustrate the interdependence of paramedic, ED, and inpatient systems and provide high-yield targets for polices to mitigate the risk to community availability of paramedic resources at times of high offload delay/system stress.
{"title":"The association between paramedic service system hospital offload time and response time.","authors":"I E Blanchard, T S Williamson, B E Hagel, D J Niven, D J Lane, S Dean, M N Shah, E S Lang, C J Doig","doi":"10.1007/s43678-023-00521-2","DOIUrl":"10.1007/s43678-023-00521-2","url":null,"abstract":"<p><strong>Objective: </strong>To address an important care issue in Canada, we tested the association between paramedic system hospital offload and response time, while considering the impact of other system-level factors.</p><p><strong>Methods: </strong>Data from Calgary, Alberta (2014-2017), included median offload (exposure) and response (outcome) time aggregated by hour, with covariates paramedic system episodes of care-dispatch and arrival of a response unit-and hospital transport arrivals (collectively called volume), time of day, and season. Analyses used linear regression and modified Poisson models.</p><p><strong>Results: </strong>301,105 EMS episodes of care over 26,193 1-h periods were included. For any given 1-h period, the median (IQR) across all episodes of care for offload time, response time, episodes of care, and hospital transport arrivals were 55.3 (45.7, 66.3) min, 8.6 (7.6, 9.8) min, 12 (8, 16) episodes, and 8 (5, 10) hospital arrivals, respectively. Multivariable modelling revealed a complex association differing over levels of exposure and covariates, requiring description using \"light stress\" and \"heavy stress\" system scenarios. The light scenario was defined as median offload of 30 min and volume < 10th percentile (six episodes and four hospital arrivals), in the summer, and the heavy scenario as median offload of 90 min and volume > 90th percentile (17 episodes and 13 hospital arrivals), in the winter. An increase is reported in minutes:seconds for median hourly response time between scenarios by time of day: 1:04-4:16 (0000-0559 h.), 0:42-2:05 (0600-1159 h.), 0:57-3:01 (1200-1759 h.), and 0:18-2:21 (1800-2359 h.).</p><p><strong>Conclusions: </strong>Increasing offload is associated with increased response time; however the relationship is complex, with a greater impact on response time noted in select situations such as high volume in the winter. These observations illustrate the interdependence of paramedic, ED, and inpatient systems and provide high-yield targets for polices to mitigate the risk to community availability of paramedic resources at times of high offload delay/system stress.</p>","PeriodicalId":55286,"journal":{"name":"Canadian Journal of Emergency Medicine","volume":null,"pages":null},"PeriodicalIF":2.4,"publicationDate":"2023-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"10272031","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-01DOI: 10.1007/s43678-023-00539-6
Jesse T T McLaren, Tahara D Bhate, Ahmed K Taher, Lucas B Chartier
{"title":"Return visit audits, quality improvement infrastructure, and a culture of safety: a theoretical model and practical assessment tool.","authors":"Jesse T T McLaren, Tahara D Bhate, Ahmed K Taher, Lucas B Chartier","doi":"10.1007/s43678-023-00539-6","DOIUrl":"https://doi.org/10.1007/s43678-023-00539-6","url":null,"abstract":"","PeriodicalId":55286,"journal":{"name":"Canadian Journal of Emergency Medicine","volume":null,"pages":null},"PeriodicalIF":2.4,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10425292/pdf/","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"10385624","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-01DOI: 10.1007/s43678-023-00549-4
Michaela McGillis, Danielle Roy, David Savage, Sarah McIsaac, Jenna Nicholls, Danielle Waltenbury, Robert Ohle
{"title":"Overuse of pharmacological treatments for patients with benign paroxysmal positional vertigo in the emergency department.","authors":"Michaela McGillis, Danielle Roy, David Savage, Sarah McIsaac, Jenna Nicholls, Danielle Waltenbury, Robert Ohle","doi":"10.1007/s43678-023-00549-4","DOIUrl":"https://doi.org/10.1007/s43678-023-00549-4","url":null,"abstract":"","PeriodicalId":55286,"journal":{"name":"Canadian Journal of Emergency Medicine","volume":null,"pages":null},"PeriodicalIF":2.4,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"9988013","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-01DOI: 10.1007/s43678-023-00529-8
David Jerome, David W Savage, Matthew Pietrosanu
Objective: Triage is the process of identifying patients with both the greatest clinical need and the greatest likelihood of benefit in the setting of limited clinical resources. The primary objective of this study was to assess the ability of formal mass casualty incident triage tools to identify patients requiring urgent lifesaving interventions.
Methods: Data from the Alberta Trauma Registry (ATR) was used to assess seven triage tools: START, JumpSTART, SALT, RAMP, MPTT, BCD and MITT. Clinical data captured in the ATR was used to determine which triage category each of the seven tools would have applied to each patient. These categorizations were compared to a reference standard definition based on the patients' need for specific urgent lifesaving interventions.
Results: Of the 9448 records that were captured 8652 were included in our analysis. The most sensitive triage tool was MPTT, which demonstrated a sensitivity of 0.76 (0.75, 0.78). Four of the seven triage tools evaluated had sensitivities below 0.45. JumpSTART had the lowest sensitivity and the highest under-triage rate for pediatric patients. All the triage tools evaluated had a moderate to high positive predictive value (> 0.67) for patients who had experienced penetrating trauma.
Conclusions: There was a wide range in the sensitivity of triage tools to identify patients requiring urgent lifesaving interventions. MPTT, BCD and MITT were the most sensitive triage tools assessed. All of the triage tools assessed should be employed with caution during mass casualty incidents as they may fail to identify a large proportion of patients requiring urgent lifesaving interventions.
{"title":"An assessment of mass casualty triage systems using the Alberta trauma registry.","authors":"David Jerome, David W Savage, Matthew Pietrosanu","doi":"10.1007/s43678-023-00529-8","DOIUrl":"https://doi.org/10.1007/s43678-023-00529-8","url":null,"abstract":"<p><strong>Objective: </strong>Triage is the process of identifying patients with both the greatest clinical need and the greatest likelihood of benefit in the setting of limited clinical resources. The primary objective of this study was to assess the ability of formal mass casualty incident triage tools to identify patients requiring urgent lifesaving interventions.</p><p><strong>Methods: </strong>Data from the Alberta Trauma Registry (ATR) was used to assess seven triage tools: START, JumpSTART, SALT, RAMP, MPTT, BCD and MITT. Clinical data captured in the ATR was used to determine which triage category each of the seven tools would have applied to each patient. These categorizations were compared to a reference standard definition based on the patients' need for specific urgent lifesaving interventions.</p><p><strong>Results: </strong>Of the 9448 records that were captured 8652 were included in our analysis. The most sensitive triage tool was MPTT, which demonstrated a sensitivity of 0.76 (0.75, 0.78). Four of the seven triage tools evaluated had sensitivities below 0.45. JumpSTART had the lowest sensitivity and the highest under-triage rate for pediatric patients. All the triage tools evaluated had a moderate to high positive predictive value (> 0.67) for patients who had experienced penetrating trauma.</p><p><strong>Conclusions: </strong>There was a wide range in the sensitivity of triage tools to identify patients requiring urgent lifesaving interventions. MPTT, BCD and MITT were the most sensitive triage tools assessed. All of the triage tools assessed should be employed with caution during mass casualty incidents as they may fail to identify a large proportion of patients requiring urgent lifesaving interventions.</p>","PeriodicalId":55286,"journal":{"name":"Canadian Journal of Emergency Medicine","volume":null,"pages":null},"PeriodicalIF":2.4,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"9985840","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-01Epub Date: 2023-05-28DOI: 10.1007/s43678-023-00524-z
Suneel Upadhye
{"title":"Code at home: chaos and crisis.","authors":"Suneel Upadhye","doi":"10.1007/s43678-023-00524-z","DOIUrl":"10.1007/s43678-023-00524-z","url":null,"abstract":"","PeriodicalId":55286,"journal":{"name":"Canadian Journal of Emergency Medicine","volume":null,"pages":null},"PeriodicalIF":2.4,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"9982093","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Objective: Health economic evaluations are used in decision-making regarding resource allocation and it is imperative that they are completed with rigor. The primary objectives were to describe the characteristics and assess the quality of economic evaluations published in emergency medicine journals.
Methods: Two reviewers independently searched 19 emergency medicine-specific journals via Medline and Embase from inception until March 3, 2022. Quality assessment was completed using the Quality of Health Economic Studies (QHES) tool, and the primary outcome was the QHES score out of 100. Additionally, we identified factors that may contribute to higher-quality publications.
Results: 7260 unique articles yielded 48 economic evaluations that met inclusion criteria. Most studies were cost-utility analyses and of high quality, with a median QHES score of 84 (interquartile range, IQR: 72, 90). Studies based on mathematical models and those primarily designed as an economic evaluation were associated with higher quality scores. The most commonly missed QHES items were: (i) providing and justifying the perspective of the analysis, (ii) providing justification for the primary outcome, and (iii) selecting an outcome that was long enough to allow for relevant events to occur.
Conclusions: The majority of health economic evaluations in the emergency medicine literature are cost-utility analyses and are of high quality. Decision analytic models and studies primarily designed as economic analyses were positively correlated with higher quality. To improve study quality, future EM economic evaluations should justify the choice of the perspective of the analysis and the selection of the primary outcome.
{"title":"Quality of health economic evaluations in emergency medicine journals: a systematic review.","authors":"Shawn Chhabra, Austin Cameron, Kednapa Thavorn, Lindsey Sikora, Krishan Yadav","doi":"10.1007/s43678-023-00535-w","DOIUrl":"10.1007/s43678-023-00535-w","url":null,"abstract":"<p><strong>Objective: </strong>Health economic evaluations are used in decision-making regarding resource allocation and it is imperative that they are completed with rigor. The primary objectives were to describe the characteristics and assess the quality of economic evaluations published in emergency medicine journals.</p><p><strong>Methods: </strong>Two reviewers independently searched 19 emergency medicine-specific journals via Medline and Embase from inception until March 3, 2022. Quality assessment was completed using the Quality of Health Economic Studies (QHES) tool, and the primary outcome was the QHES score out of 100. Additionally, we identified factors that may contribute to higher-quality publications.</p><p><strong>Results: </strong>7260 unique articles yielded 48 economic evaluations that met inclusion criteria. Most studies were cost-utility analyses and of high quality, with a median QHES score of 84 (interquartile range, IQR: 72, 90). Studies based on mathematical models and those primarily designed as an economic evaluation were associated with higher quality scores. The most commonly missed QHES items were: (i) providing and justifying the perspective of the analysis, (ii) providing justification for the primary outcome, and (iii) selecting an outcome that was long enough to allow for relevant events to occur.</p><p><strong>Conclusions: </strong>The majority of health economic evaluations in the emergency medicine literature are cost-utility analyses and are of high quality. Decision analytic models and studies primarily designed as economic analyses were positively correlated with higher quality. To improve study quality, future EM economic evaluations should justify the choice of the perspective of the analysis and the selection of the primary outcome.</p>","PeriodicalId":55286,"journal":{"name":"Canadian Journal of Emergency Medicine","volume":null,"pages":null},"PeriodicalIF":2.4,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"9985527","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-01DOI: 10.1007/s43678-023-00545-8
Julia Sarty, Eleanor A Fitzpatrick, Majid Taghavi, Peter T VanBerkel, Katrina F Hurley
Purpose: To characterize patients who left without being seen (LWBS) from a Canadian pediatric Emergency Department (ED) and create predictive models using machine learning to identify key attributes associated with LWBS.
Methods: We analyzed administrative ED data from April 1, 2017, to March 31, 2020, from IWK Health ED in Halifax, NS. Variables included: visit disposition; Canadian Triage Acuity Scale (CTAS); triage month, week, day, hour, minute, and day of the week; sex; age; postal code; access to primary care provider; visit payor; referral source; arrival by ambulance; main problem (ICD10); length of stay in minutes; driving distance in minutes; and ED patient load. The data were randomly divided into training (80%) and test datasets (20%). Five supervised machine learning binary classification algorithms were implemented to train models to predict LWBS patients. We balanced the dataset using Synthetic Minority Oversampling Technique (SMOTE) and used grid search for hyperparameter tuning of our models. Model evaluation was made using sensitivity and recall on the test dataset.
Results: The dataset included 101,266 ED visits where 2009 (2%) records were excluded and 5800 LWBS (5.7%). The highest-performing machine learning model with 16 patient attributes was XGBoost which was able to identify LWBS patients with 95% recall and 87% sensitivity. The most influential attributes in this model were ED patient load, triage hour, driving minutes from home address to ED, length of stay (minutes since triage), and age.
Conclusion: Our analysis showed that machine learning models can be used on administrative data to predict patients who LWBS in a Canadian pediatric ED. From 16 variables, we identified the five most influential model attributes. System-level interventions to improve patient flow have shown promise for reducing LWBS in some centres. Predicting patients likely to LWBS raises the possibility of individual patient-level interventions to mitigate LWBS.
目的:描述加拿大儿科急诊科(ED)的无诊离开(LWBS)患者的特征,并使用机器学习创建预测模型,以识别与LWBS相关的关键属性。方法:我们分析了2017年4月1日至2020年3月31日来自哈利法克斯IWK Health ED的行政ED数据。变量包括:访问处置;加拿大分诊敏锐度量表(CTAS);分类月、周、日、时、分、日;性;年龄;邮政编码;获得初级保健提供者的服务;访问付款人;推荐来源;救护车到达;主要问题(ICD10);停留时间(以分钟为单位);行车距离(分钟);和急诊科的病人负荷。数据随机分为训练数据集(80%)和测试数据集(20%)。采用五种监督式机器学习二分类算法训练模型预测LWBS患者。我们使用合成少数派过采样技术(SMOTE)平衡数据集,并使用网格搜索进行模型的超参数调整。利用灵敏度和召回率对测试数据集进行模型评价。结果:该数据集包括101,266例ED就诊,其中2009年(2%)的记录被排除,5800例LWBS(5.7%)的记录被排除。具有16个患者属性的表现最好的机器学习模型是XGBoost,它能够以95%的召回率和87%的灵敏度识别LWBS患者。该模型中影响最大的属性是急诊科患者负荷、分诊时间、从家庭住址到急诊科的驾车分钟数、住院时间(分诊后的分钟数)和年龄。结论:我们的分析表明,机器学习模型可以用于管理数据来预测加拿大儿科急诊科的LWBS患者。从16个变量中,我们确定了五个最具影响力的模型属性。在一些中心,改善病人流动的系统级干预措施已显示出减少LWBS的希望。预测可能发生LWBS的患者提高了个体患者水平干预以减轻LWBS的可能性。
{"title":"Machine learning to identify attributes that predict patients who leave without being seen in a pediatric emergency department.","authors":"Julia Sarty, Eleanor A Fitzpatrick, Majid Taghavi, Peter T VanBerkel, Katrina F Hurley","doi":"10.1007/s43678-023-00545-8","DOIUrl":"https://doi.org/10.1007/s43678-023-00545-8","url":null,"abstract":"<p><strong>Purpose: </strong>To characterize patients who left without being seen (LWBS) from a Canadian pediatric Emergency Department (ED) and create predictive models using machine learning to identify key attributes associated with LWBS.</p><p><strong>Methods: </strong>We analyzed administrative ED data from April 1, 2017, to March 31, 2020, from IWK Health ED in Halifax, NS. Variables included: visit disposition; Canadian Triage Acuity Scale (CTAS); triage month, week, day, hour, minute, and day of the week; sex; age; postal code; access to primary care provider; visit payor; referral source; arrival by ambulance; main problem (ICD10); length of stay in minutes; driving distance in minutes; and ED patient load. The data were randomly divided into training (80%) and test datasets (20%). Five supervised machine learning binary classification algorithms were implemented to train models to predict LWBS patients. We balanced the dataset using Synthetic Minority Oversampling Technique (SMOTE) and used grid search for hyperparameter tuning of our models. Model evaluation was made using sensitivity and recall on the test dataset.</p><p><strong>Results: </strong>The dataset included 101,266 ED visits where 2009 (2%) records were excluded and 5800 LWBS (5.7%). The highest-performing machine learning model with 16 patient attributes was XGBoost which was able to identify LWBS patients with 95% recall and 87% sensitivity. The most influential attributes in this model were ED patient load, triage hour, driving minutes from home address to ED, length of stay (minutes since triage), and age.</p><p><strong>Conclusion: </strong>Our analysis showed that machine learning models can be used on administrative data to predict patients who LWBS in a Canadian pediatric ED. From 16 variables, we identified the five most influential model attributes. System-level interventions to improve patient flow have shown promise for reducing LWBS in some centres. Predicting patients likely to LWBS raises the possibility of individual patient-level interventions to mitigate LWBS.</p>","PeriodicalId":55286,"journal":{"name":"Canadian Journal of Emergency Medicine","volume":null,"pages":null},"PeriodicalIF":2.4,"publicationDate":"2023-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"9987313","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":4,"RegionCategory":"医学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}