Dataset Information

Effect of Machine Learning on Dispatcher Recognition of Out-of-Hospital Cardiac Arrest During Calls to Emergency Medical Services: A Randomized Clinical Trial.

ABSTRACT:

Importance

Emergency medical dispatchers fail to identify approximately 25% of cases of out-of-hospital cardiac arrest (OHCA), resulting in lost opportunities to save lives by initiating cardiopulmonary resuscitation.

Objective

To examine how a machine learning model trained to identify OHCA and alert dispatchers during emergency calls affected OHCA recognition and response.

Design, setting, and participants

This double-masked, 2-group, randomized clinical trial analyzed all calls to emergency number 112 (equivalent to 911) in Denmark. Calls were processed by a machine learning model using speech recognition software. The machine learning model assessed ongoing calls, and calls in which the model identified OHCA were randomized. The trial was performed at Copenhagen Emergency Medical Services, Denmark, between September 1, 2018, and December 31, 2019.

Intervention

Dispatchers in the intervention group were alerted when the machine learning model identified out-of-hospital cardiac arrest, and those in the control group followed normal protocols without alert.

Main outcomes and measures

The primary end point was the rate of dispatcher recognition of subsequently confirmed OHCA.

Results

A total of 169?049 emergency calls were examined, of which the machine learning model identified 5242 as suspected OHCA. Calls were randomized to control (2661 [50.8%]) or intervention (2581 [49.2%]) groups. Of these, 336 (12.6%) and 318 (12.3%), respectively, had confirmed OHCA. The mean (SD) age among of these 654 patients was 70 (16.1) years, and 419 of 627 patients (67.8%) with known gender were men. Dispatchers in the intervention group recognized 296 confirmed OHCA cases (93.1%) with machine learning assistance compared with 304 confirmed OHCA cases (90.5%) using standard protocols without machine learning assistance (P?=?.15). Machine learning alerts alone had a significantly higher sensitivity than dispatchers without alerts for confirmed OHCA (85.0% vs 77.5%; P?Conclusions and relevanceThis randomized clinical trial did not find any significant improvement in dispatchers' ability to recognize cardiac arrest when supported by machine learning even though artificial intelligence did surpass human recognition.

Trial registration

ClinicalTrials.gov Identifier: NCT04219306.

SUBMITTER: Blomberg SN

PROVIDER: S-EPMC7788469 | biostudies-literature | 2021 Jan

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Effect of Machine Learning on Dispatcher Recognition of Out-of-Hospital Cardiac Arrest During Calls to Emergency Medical Services: A Randomized Clinical Trial.

Blomberg Stig Nikolaj SN Christensen Helle Collatz HC Lippert Freddy F Ersbøll Annette Kjær AK Torp-Petersen Christian C Sayre Michael R MR Kudenchuk Peter J PJ Folke Fredrik F

JAMA network open 20210104 1

<h4>Importance</h4>Emergency medical dispatchers fail to identify approximately 25% of cases of out-of-hospital cardiac arrest (OHCA), resulting in lost opportunities to save lives by initiating cardiopulmonary resuscitation.<h4>Objective</h4>To examine how a machine learning model trained to identify OHCA and alert dispatchers during emergency calls affected OHCA recognition and response.<h4>Design, setting, and participants</h4>This double-masked, 2-group, randomized clinical trial analyzed al ...[more]

PMID: 33404620

Similar Datasets

Project description:Importance:Emergency medical services (EMS) deliver essential initial care for patients with out-of-hospital cardiac arrest (OHCA), but the extent to which patient outcomes vary between different EMS agencies is not fully understood. Objective:To quantify variation in patient outcomes after OHCA across EMS agencies. Design, Setting, and Participants:This observational cohort study was conducted in the Resuscitation Outcomes Consortium (ROC) Epistry, a prospective multicenter OHCA registry at 10 sites in North America. Any adult with OHCA treated by an EMS from April 2011 through June 2015 was included. Data analysis occurred from May 2017 to March 2018. Exposure:Treating EMS agency. Main Outcomes and Measures:The primary outcome was survival to hospital discharge. Secondary outcomes were return of spontaneous circulation at emergency department arrival and favorable functional outcome at hospital discharge (defined as a modified Rankin scale score ≤3). Multivariable hierarchical logistic regression models were used to adjust confounders and clustering of patients within EMS agencies, and calculated median odds ratios (MORs) were used to quantify the extent of residual variation in outcomes between EMS agencies. Results:We identified 43 656 patients with OHCA treated by 112 EMS agencies. At EMS agency level, we observed large variations in survival to hospital discharge (range, 0%-28.9%; unadjusted MOR, 1.43 [95% CI, 1.34-1.54]), return of spontaneous circulation on emergency department arrival (range, 9.0%-57.1%; unadjusted MOR, 1.53 [95% CI, 1.43-1.65]), and favorable functional outcome (range, 0%-20.4%; unadjusted MOR, 1.54 [95% CI, 1.40-1.73]). This variation persisted despite adjustment for patient-level and EMS agency-level factors known to be associated with outcomes (adjusted MOR for survival 1.56 [95% CI 1.44-1.73]; adjusted MOR for return of spontaneous circulation at emergency department arrival, 1.50 [95% CI, 1.41-1.62]; adjusted MOR for functionally favorable survival, 1.53 [95% CI, 1.37-1.78]). After restricting analysis to those who survived more than 60 minutes after hospital arrival and including hospital treatment characteristics, the variation persisted (adjusted MOR for survival, 1.49 [95% CI, 1.36-1.69]; adjusted MOR for functionally favorable survival, 1.34 [95% CI, 1.20-1.59]). Conclusions and Relevance:We found substantial variations in patient outcomes after OHCA between a large group of EMS agencies in North America that were not explained by documented patient-level and EMS agency-level variables.

Project description:BackgroundThe aim of this study was to identify key indicator symptoms and patient factors associated with correct out of hospital cardiac arrest (OHCA) dispatch allocation. In previous studies, from 3% to 62% of OHCAs are not recognised by Emergency Medical Service call handlers, resulting in delayed arrival at scene.MethodsRetrospective, mixed methods study including all suspected or confirmed OHCA patients transferred to one acute hospital from its associated regional Emergency Medical Service in England from 1/7/2013 to 30/6/2014. Emergency Medical Service and hospital data, including voice recordings of EMS calls, were analysed to identify predictors of recognition of OHCA by call handlers. Logistic regression was used to explore the role of the most frequently occurring (key) indicator symptoms and characteristics in predicting a correct dispatch for patients with OHCA.ResultsA total of 39,136 dispatches were made which resulted in transfer to the hospital within the study period, including 184 patients with OHCA. The use of the term 'Unconscious' plus one or more of symptoms 'Not breathing/Ineffective breathing/Noisy breathing' occurred in 79.8% of all OHCAs, but only 72.8% of OHCAs were correctly dispatched as such. 'Not breathing' was associated with recognition of OHCA by call handlers (Odds Ratio (OR) 3.76). The presence of key indicator symptoms 'Breathing' (OR 0.29), 'Reduced or fluctuating level of consciousness' (OR 0.24), abnormal pulse/heart rate (OR 0.26) and the characteristic 'Female patient' (OR 0.40) were associated with lack of recognition of OHCA by call handlers (p-values < 0.05).ConclusionsThere is a small proportion of calls in which cardiac arrest indicators are described but the call is not dispatched as such. Stricter adherence to dispatch protocols may improve call handlers' OHCA recognition. The existing dispatch protocol would not be improved by the addition of further terms as this would be at the expense of dispatch specificity.

Project description:BackgroundThe European resuscitation council have highlighted emergency medical dispatch centres as an important key player for early recognition of Out-of-Hospital Cardiac Arrest (OHCA) and in providing dispatcher assisted cardiopulmonary resuscitation (CPR) before arrival of emergency medical services. Early recognition is associated with increased bystander CPR and improved survival rates. The aim of this study is to describe OHCA call handling in emergency medical dispatch centres in Copenhagen (Denmark), Stockholm (Sweden) and Oslo (Norway) with focus on sensitivity of recognition of OHCA, provision of dispatcher-assisted CPR and time intervals when CPR is initiated during the emergency call (NO-CPRprior), and to describe OHCA call handling when CPR is initiated prior to the emergency call (CPRprior).MethodsBaseline data of consecutive OHCA eligible for inclusion starting January 1st 2016 were collected from respective cardiac arrest registries. A template based on the Cardiac Arrest Registry to Enhance Survival definition catalogue was used to extract data from respective cardiac arrest registries and from corresponding audio files from emergency medical dispatch centres. Cases were divided in two groups: NO-CPRprior and CPRprior and data collection continued until 200 cases were collected in the NO-CPRprior-group.ResultsNO-CPRprior OHCA was recognised in 71% of the calls in Copenhagen, 83% in Stockholm, and 96% in Oslo. Abnormal breathing was addressed in 34, 7 and 98% of cases and CPR instructions were started in 50, 60, and 80%, respectively. Median time (mm:ss) to first chest compression was 02:35 (Copenhagen), 03:50 (Stockholm) and 02:58 (Oslo). Assessment of CPR quality was performed in 80, 74, and 74% of the cases. CPRprior comprised 71 cases in Copenhagen, 9 in Stockholm, and 38 in Oslo. Dispatchers still started CPR instructions in 41, 22, and 40% of the calls, respectively and provided quality assessment in 71, 100, and 80% in these respective instances.ConclusionsWe observed variations in OHCA recognition in 71-96% and dispatcher assisted-CPR were provided in 50-80% in NO-CPRprior calls. In cases where CPR was initiated prior to emergency calls, dispatchers were less likely to start CPR instructions but provided quality assessments during instructions.

Project description:BackgroundAlthough studies have reported variation in out-of-hospital cardiac arrest (OHCA) survival by geographic location, little is known about variation in OHCA survival at the level of emergency medical service (EMS) agencies-which may have modifiable practices, unlike counties and regions. We quantified the variation in OHCA survival across EMS agencies and explored whether variation in 2 specific EMS resuscitation practices were associated with survival to hospital admission.MethodsWithin the Cardiac Arrest Registry to Enhance Survival, a prospective registry representing ≈51% of the US population, we identified 258 342 OHCAs from 764 EMS agencies with >10 OHCA cases annually during 2015 to 2019. Using hierarchical logistic regression, risk-standardized rates of survival to hospital admission were computed for each EMS agency. We quantified inter-agency variation in survival with median odds ratios and assessed the association of 2 resuscitation practices (EMS response time and the proportion of OHCAs with termination of resuscitation without meeting futility criteria) with EMS agency survival rates to hospital admission.ResultsAcross 764 EMS agencies comprising 258 342 OHCAs, the median risk-standardized rate of survival to hospital admission was 27.3% (interquartile range, 24.5%-30.1%; range: 16.0%-45.6%). The adjusted median odds ratio was 1.35 (95% CI, 1.32-1.39), denoting that the odds of survival of 2 patients with identical covariates varied by 35% at 2 randomly selected EMS agencies. EMS agencies in the lowest quartile of risk-standardized survival had longer EMS response times when compared with the highest quartile (12.0±3.4 versus 9.0±2.6 minutes; P<0.001), and a higher proportion of OHCAs with termination of resuscitation without meeting futility criteria (27.9±16.1% versus 18.9±11.4%; P<0.001).ConclusionsSurvival after OHCA varies widely across EMS agencies. EMS response times and termination of resuscitation practices were associated with agency-level rates of survival to hospital admission, suggesting potentially modifiable practices which can improve OHCA survival.

Project description:BackgroundThere are approximately 60,000 out-of-hospital cardiac arrests (OHCA) in the United Kingdom (UK) each year. Within the UK there are well-established clinical practice guidelines that define when resuscitation should be commenced in OHCA, and when resuscitation should cease. Background literature indicates that decision-making in the commencement and cessation of resuscitation efforts in OHCA is complex, and not comprehensively understood. No relevant research from the UK has been published to date and this research study seeks to explore the influences on UK Emergency Medical Service (EMS) provider decision-making when commencing and ceasing resuscitation attempts in OHCA. The aim of this research to explore the influences on UK Emergency Medical Services provider decision-making when commencing and ceasing resuscitation attempts in OHCA.MethodsFour focus groups were convened with 16 clinically active EMS providers. Four case vignettes were discussed to explore decision-making within the focus groups. Thematic analysis was used to analyse transcripts.ResultsThis research found that there are three stages in the decision-making process when EMS providers consider whether to commence or cease resuscitation attempts in OHCA. These stages are: the call; arrival on scene; the protocol. Influential factors present at each of the three stages can lead to different decisions and variability in practice. These influences are: factual information available to the EMS provider; structural factors such as protocol, guidance and research; cultural beliefs and values; interpersonal factors; risk factors; personal values and beliefs.ConclusionsAn improved understanding of the circumstantial, individual and interpersonal factors that mediate the decision-making process in clinical practice could inform the development of more effective clinical guidelines, education and clinical decision support in OHCA. These changes have the potential to lead to greater consistency. and EMS provider confidence, with the potential for improved patient outcome from OHCA.

Project description:BackgroundThe challenges encountered in emergency medical services (EMS) contacts with children are likely most pronounced in infants, but little is known about their out-of-hospital care. Our primary aim was to describe the characteristics of EMS contacts with infants. The secondary aims were to examine the symptom-based dispatch system for nonverbal infants, and to observe the association of unfavorable patient outcomes with patient and EMS mission characteristics.MethodsIn a population-based 5-year retrospective cohort of all 1712 EMS responses for infants (age < 1 year) in Helsinki, Finland (population 643,000, < 1-year old population 6548), we studied 1) the characteristics of EMS missions with infants; 2) mortality within 12 months; 3) pediatric intensive care unit (PICU) admissions; 4) medical state of the infant upon presentation to the emergency department (ED); 5) any medication or respiratory support given at the ED; 6) hospitalization; and 7) surgical procedures during the same hospital visit.Results1712 infants with a median age of 6.7 months were encountered, comprising 0.4% of all EMS missions. The most common complaints were dyspnea, low-energy falls, and choking. Two infants died on-scene. The EMS transported 683 (39.9%) infants. One (0.1%) infant died during the 12-month follow-up period. Ninety-one infants had abnormal clinical examination upon arrival at the ED. PICU admissions (n = 28) were associated with young age (P < 0.01), a history of prematurity or problems in the neonatal period (P = 0.01), and previous EMS contacts within 72 h (P = 0.04). The adult-derived dispatch codes did not associate with the final diagnoses of the infants.ConclusionsInfants form a small but distinct group in pediatric EMS care, with specific characteristics differing from the overall pediatric population. Many EMS contacts with infants were nonurgent or medically unjustified, possibly reflecting an unmet need for other family services. The use of adult-derived symptom codes for dispatching is not optimal for infants. Unfavorable patient outcomes were rare. Risk factors for such outcomes include quickly renewed contacts, young age and health problems in the neonatal period.

Dataset Information

Effect of Machine Learning on Dispatcher Recognition of Out-of-Hospital Cardiac Arrest During Calls to Emergency Medical Services: A Randomized Clinical Trial.

Importance

Objective

Design, setting, and participants

Intervention

Main outcomes and measures

Results

Trial registration

Publications

Effect of Machine Learning on Dispatcher Recognition of Out-of-Hospital Cardiac Arrest During Calls to Emergency Medical Services: A Randomized Clinical Trial.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets