Aim:The Danish Drowning Cohort has collected nationwide prehospital data on fatal and non-fatal drowning incidents in Denmark. The database contains 47 variables: 22 variables from the existing Utstein Drowning Template and 25 additional variables based on the availability in the Danish prehospital medical records. This study aimed to estimate the availability of data (ie, data completeness) for the predefined prehospital registry variables in the Danish Drowning Cohort and the availability of Utstein variables in the subpopulations of drowning patients with and without out-of-hospital cardiac arrest (OHCA). Methods:This study included all fatal and non-fatal drownings in the Danish Drowning Cohort from January 1, 2016, to December 31, 2023. Data availability rates were reported for all variables in the Danish Drowning Cohort and the Utstein variables in the subpopulations of patients with and without OHCA, presented as medians with interquartile ranges (IQR). Results:A total of 1,664 drowning patients were analysed (697, 42% with OHCA). The median data availability rate in the Danish Drowning Cohort was 100% (IQR: 88-100%): 100% (IQR: 81-100%) for non-OHCA and 100% (IQR: 87-100%) for OHCA patients. Among the subset of 22 available Utstein variables, the overall data availability was 95% (IQR: 75-100%). For the 17 Utstein variables applicable to both non-OHCA and OHCA, the availability rate was 92% (IQR 76-100%) in non-OHCA cases and 88% (IQR 78-100%) in OHCA cases, with no difference between groups. Overall, 21 of 79 (26%) Utstein variables were unavailable in the Danish setting. Conclusion:The Danish Drowning Cohort contained nationwide data on fatal and non-fatal drowning patients with high degrees of availability. Data availability was high for most variables defined in the Utstein Drowning Template, both for drowning patients with and without OHCA. However, 21 of 79 Utstein variables were unavailable despite the Danish high-resource healthcare setting.

Introduction: Early prehospital identification of stroke and triage to the right level of care may result in more patients receiving acute treatment. Approximately 25% of patients in Denmark receive revascularisation after an acute stroke. Further, 20 % of patients with stroke are not being recognized by paramedics when assessed on scene in a telephone conference with the stroke centre. Accurate prehospital stroke identification is critical for treatment and patient outcomes. The National Institutes of Health Stroke Scale (NIHSS) is an established tool for assessing stroke severity, but it is not commonly used by paramedics. In this study, paramedics are using an app to assess the patient, guiding the paramedic through the NIHSS, while each step is documented in a short video, which is transferred to the centralized stroke center. Methods: A stepped wedge randomized trial is performed, with paramedics being the subject which is randomized. First, paramedics are enrolled in October 2024. Data collection is terminated in December 2025. A paramedic in an intervention cluster uses a mobile app for patients with suspected stroke, recording videos of patients. The app analyses the videos, highlighting potential stroke predictors. Recorded videos and NIHSS scores are transmitted to the stroke center for thrombolysis decision-making. In the control group, standard procedures are followed. Results: The primary outcome measure is for prehospital recognition of patients with acute stroke, defined as the proportion of patients accepted for stroke evaluation and discharged with a final stroke diagnosis. A power calculation yields that no less than 1200 patients will be enrolled in each arm of the trial for 95% power to detect a 5% improvement, with 85% sensitivity assumed in the control group. Conclusion: The results will show the intervention’s potential to improve stroke recognition rates and its ability to reduce on-scene times.

Abstract Background Effective prevention of drowning and aquatic incidents requires timely and accurate surveillance supported by high-quality validated data. In Denmark, the use of the free-text fields in the Danish Prehospital Medical Record has proven effective in identifying potentially relevant cases for such surveillance. While these free-text fields contain rich contextual information, manual screening of all records is impractical. This study aimed to develop and internally validate a Danish natural language processing pipeline for identifying drowning and aquatic incidents (AquaAI) from routine prehospital records and prioritizing records for final manual validation by medical experts. Methods This nationwide retrospective cohort study was conducted using Danish prehospital electronic medical records from 2016 to 2024. Medical records were first retrieved using the Danish Drowning Formula, an iteratively developed trigger-word search algorithm, and expert-labelled as Drowning, Aquatic incident, or Non-relevant. A Danish transformer-based language model was then fine-tuned for three-class sequence classification as the second-stage classifier in this workflow and embedded in a hybrid pipeline with rule-based safeguards. Model development used a temporal split with training data from 2016 to 2021 and validation data from 2022. Final performance was evaluated on a temporally separated hold-out dataset comprising records from 2023 to 2024. Primary outcomes were class-wise sensitivity and a binary Relevant (Drowning + Aquatic incident) versus Non-relevant analysis to quantify case finding and workload reduction. Results The dataset comprised 40,876 medical records retrieved with the Danish Drowning Formula. On the hold-out dataset, AquaAI identified 91% of relevant cases and correctly filtered 89% of non-relevant records from manual review. Overall, this reduced the number of records requiring manual review by 77.5%. In the three-class analysis, sensitivity was 84% for Drowning, 83% for Aquatic incident, and 89% for Non-relevant. Only one drowning case (0.3%) was classified as non-relevant. Conclusions A hybrid transformer-based pipeline operating as a second-stage classifier within a two-step retrieval workflow can identify drowning and aquatic incidents from prehospital free-text narratives with high case finding and substantial reduction of manual review. This approach may support national surveillance and targeted prevention initiatives.

OBJECTIVE:The Intravenous versus Intraosseous Vascular Access for Out-of-Hospital Cardiac Arrest (IVIO) trial was a randomised clinical trial that investigated initial vascular access strategy for out-of-hospital cardiac arrest. The current manuscript presents outcomes at 6 months and 1 year. METHODS:Adults with non-traumatic out-of-hospital cardiac arrest, in whom vascular access was indicated, were randomised to initial intraosseous or intravenous access. The allocated method was attempted up to two times. Prespecified 6-months and 1-year outcomes included survival, survival with a favourable neurological outcome, defined as a modified Rankin Scale score of 0-3, and health-related quality-of-life assessed using the EuroQoL 5-Dimension 5-Level questionnaire on domains of mobility, self-care, usual activities, pain/discomfort, and anxiety/depression. RESULTS:Of the 1479 patients included in the main manuscript primary analyses, three were lost to follow-up for 1-year survival. At 1 year, 82 patients (11%) in the intraosseous group and 68 patients (9%) in the intravenous group were alive (risk ratio 1.24; 95% confidence interval 0.91-1.67). Survival with a favourable neurological outcome was observed in 76 patients (10%) and 61 patients (8%), respectively (risk ratio 1.28; 95% confidence interval 0.93-1.77). Among survivors, the mean EQ-5D-5L numeric score was 83 in the intraosseous group and 76 in the intravenous group (mean difference 7; 95% confidence interval 1-13). CONCLUSION:Long-term outcomes were similar between patients who received initial intraosseous versus intravenous vascular access during adult out-of-hospital cardiac arrest. These findings do not support a difference in patient outcomes between the two vascular access strategies. TRIAL REGISTRATION:EU Clinical Trials number 2022-500744-38-00; ClinicalTrials.gov number NCT05205031.

Rapid recognition of stroke/transient ischaemic attack (TIA) in the emergency telephone is crucial to ensure timely intervention and direct transport to stroke centres. Our study investigated the diagnostic accuracy via the 1-1-2 emergency line (1-1-2) call-takers in Region Zealand, Denmark, and examined whether recognised cases had a higher chance of referral to a primary stroke centre and receiving thrombolysis. We retrospectively analysed calls linked to hospital discharge codes for stroke/TIA. Call-taker stroke/TIA suspicion was documented. Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) were calculated. Additionally, we calculated the relative risk of receiving intravenous thrombolysis and direct stroke-centre transport for recognised versus unrecognised calls. Among 196,235 calls to 1-1-2 from 2021 to 2024, 2581 patients had confirmed stroke/TIA. Dispatcher sensitivity was 0.75, with a PPV of 0.18. Recognised patients compared to unrecognised cases had a higher chance of direct admissions to primary stroke units with a RR of 1.46 (95

Timely access to specialized stroke treatment is critical for patient outcomes onset of stroke. In Region Zealand, Denmark, stroke care is centralized to a single Stroke Unit. However, it remains unclear whether distance to the Stroke Unit affects prehospital transport decisions and access to reperfusion therapy for these patients. The aim of this study was to examine whether distance to the Stroke Unit was associated with prehospital transport destination and access to reperfusion therapy among patients with suspected stroke. We conducted a retrospective cohort study of all Emergency Medical Services (EMS) dispatches for suspected stroke in Region Zealand from 2018 to 2022. The primary outcomes were transport destination (direct transport to the Stroke Unit versus local emergency department) and receipt of reperfusion treatment. Patients were categorized according to final diagnosis and transport destination: Stroke Unit or local emergency department. We examined stroke severity, time from symptom onset to EMS call, distance to Stroke Unit, and advanced treatment was administered. Descriptive characteristics were calculated for subgroups. Differences between groups were analysed using Fisher’s exact and Kruskal-Wallis tests. Among 18,289 patients with stroke-suspected dispatches, 3,989 had a confirmed stroke diagnosis. Of these, 3,225 (81

Background Emergency cricothyroidotomy is a rare but potentially lifesaving rescue airway procedure in patients with failed conventional airway management. Comprehensive nationwide data on patient characteristics, procedural success and clinical outcomes remain limited, particularly for patients with out-of-hospital cardiac arrest (OHCA). Methods This nationwide registry-based observational cohort study included all patients undergoing prehospital emergency cricothyroidotomy in Denmark between 2016 and 2023. Cases were identified through nationwide prehospital medical records and the Danish Helicopter Emergency Medical Services database using iterative free-text searches and manual validation. OHCA patients undergoing emergency cricothyroidotomy were identified through linkage with the Danish Cardiac Arrest Register. Results A total of 71 validated prehospital emergency cricothyroidotomy cases were identified, including 48 OHCA patients and 23 patients without OHCA. Overall procedural success was achieved in 87.3% of cases, including 83.3% in patients with OHCA and 95.7% in patients without OHCA. Aspiration (32.4%) and oropharyngeal bleeding (33.8%) were the most frequent indications. Among the 48 patients with OHCA, return of spontaneous circulation was achieved in 25.0%, 30-day survival was 4.5%, and one-year survival was 2.3%. Overall, emergency cricothyroidotomy was performed in 48 of 39,885 nationwide OHCA patients (0.12%). Conclusion Prehospital emergency cricothyroidotomy was an exceedingly rare procedure but demonstrated high procedural success in this nationwide physician-staffed Emergency Medical Services system. Survival among patients with OHCA was relatively low, reflecting the critical condition of patients requiring emergency cricothyroidotomy.

BACKGROUND:Stroke and transient ischemic attack (TIA) are clinical diagnoses; however, magnetic resonance imaging (MRI) with diffusion weighted imaging (DWI) is frequently used as a supplement in the diagnostic process and may also assist in differentiating stroke from stroke mimics. We aimed to test strengths and weaknesses of applying MRI as first scan in patients with suspected stroke, using clinical diagnoses as a reference. METHODS:Patients consecutively admitted with suspected stroke to a non-comprehensive stroke emergency medical unit with a concomitant MRI were included. A neuroradiologist and a neurology resident collected imaging and clinical data, respectively. We calculated sensitivity and specificity between clinical diagnosis and radiological findings, and tested associations between bedside characteristics and disagreement in diagnoses. RESULTS:We included 627 patients (mean age 68 years, 56% female) and found 137 (22%) had ischemic stroke, 102 (16%) TIA, and 5 (0.8%) hemorrhagic strokes at discharge. On the MRI, 155 (25%) had ischemic lesions, 8 (1.3%) hemorrhagic lesions. For all ischemic strokes or TIAs there was 58% sensitivity and 96% specificity with MRI where sensitivity increased to 93% when excluding TIA-patients. Persistent symptoms at admission in the Emergency Department and unilateral motor symptoms associated with DWI-positive stroke/TIA whereas decreased level of consciousness associated with DWI-positive stroke/TIA mimics. CONCLUSIONS:In non-comprehensive stroke care with an MRI-first approach, MRI is moderately sensitive and highly specific for stroke and TIA, and its performance varies depending on the clinical context.

Frequent callers to Emergency Medical Services represent a small group of citizens who account for a substantial number of calls. Despite their impact, little is known about their characteristics and patterns of EMS use. This study aimed to examine frequency, dispatch outcomes, and characteristics of frequent callers in a prehospital setting. We conducted an observational cohort study of calls to the Emergency Medical Dispatch Center in Region Zealand, Denmark, between 2018 and 2025. Calls without a civil register number and individuals younger than 18 years were excluded. Frequent callers were defined as individuals contacting 1-1-2 five or more times within 12 consecutive months. Calls from frequent callers were categorized as pre-frequent calls (before reaching the threshold) and post-threshold calls (from the fifth call onward). Descriptive statistics were used to compare frequent callers with non-frequent callers, as well as subgroups of frequent callers based on call frequency. Overall, 214 091 individuals made 443 977 calls. Of these, 3.0

Early identification and correct triage of stroke in the prehospital setting are essential for timely treatment and improved outcomes. In Region Zealand, Denmark, ambulance personnel are instructed to consult a neurologist when stroke is suspected to facilitate direct admission to a specialised stroke unit. However, little is known about what happens when emergency medical service (EMS) dispatchers correctly suspect stroke, but the patient is not admitted directly to a stroke unit. This includes whether a consultation with a neurologist occurs and on what basis patients are rejected. The primary aim was to determine how often ambulance personnel consulted with a neurologist and examine documented reasons for not referring patients directly to a stroke unit. The secondary aim was to describe symptom presentation in these patients. This retrospective observational study linked prehospital ambulance records, Computer-Aided Dispatch data, and the Danish Stroke Register. The prehospital medical record by ambulance personnel was reviewed for each case to examine the decision-making during ambulance responses using a predefined instrument. The study included patients ≥ 18 years who were initially suspected of having a stroke by the EMS dispatcher, were not transported directly to a stroke unit, and a stroke diagnose in the Danish Stroke Register corresponding to the same clinical event from January 2021 to September 2024 in Region Zealand, Denmark. Descriptive statistics were used to summarise findings. Of the 680 patients included, 583 (86

Background Early and reliable predictions of elevated cardiac troponin levels from electrocardiograms (ECGs) in the prehospital setting could serve as a valuable risk stratification tool, guiding triage and early intervention in patients with suspected acute coronary syndrome, and especially in patients presenting with electrocardiographic non-ST elevation (NSTE). Therefore, the primary objective of this study was to investigate whether machine learning applied to the prehospital ECG can enable early identification of patients at high risk of myocardial infarction. Methods and Results A total of 100,334 patients with a prehospital ECG and in-hospital troponin measurement available were included in this study. A random forest model was developed to predict elevated cardiac troponin T (>14 ng/L) from the prehospital ECG. Mean age was 64.72 (17.05) and 55.13% of the cohort were male. Five-fold cross-validation showed an area under the receiver operating characteristics curve of 0.88 and area under the precision-recall curve of 0.89. Positive predictive value was 0.80 and negative predictive value was 0.79. Results on the internal independent test cohort and achieved similar performance. Supplementary analyses showed that the model was able to identify NSTE patients with elevated troponin T as well as identified a gap in time to definitive treatment for NSTE patients compared to those with ST elevation (STE). Conclusion Machine learning applied to the prehospital ECG can identify patients at high risk of myocardial injury before biomarker results are available, with potential to streamline patient flow and reduce time to definitive treatment in patients with suspected acute coronary syndrome.

BACKGROUND:Not all victims of Out-of-Hospital Cardiac Arrest (OHCA) wish to be resuscitated. Some potential victims of OHCA opt to sign a donotattemptresuscitation (DNAR) order, which mandates that no resuscitation should be attempted upon cardiac arrest. Despite a DNAR order being issued, OHCA patients receive treatment against their wish. AIM:The aim of this study was to conduct an epidemiological analysis of DNAR orders in OHCA cases. METHODS:This nationwide, register-based cohort study included all OHCA patients aged 18 years or older in Denmark from January 1, 2016, to December 31, 2023. RESULTS:From 2016 to 2023, a significant rising trend in the number of OHCAs where a DNAR has been in place was observed. In 2016, the percentage of OHCA with DNAR order was 1%, while in 2023, 3% of OHCA occurred in patients with an existing DNAR. This may influence the survival rates following OHCA. Excluding OHCA with DNAR order, overall survival rates would have been higher, with an additional 0.1% in 2016 and an additional 0.5% in 2023. CONCLUSION:Between 2016 and 2023, in Denmark, there has been a significant increasing trend in the number of OHCAs treated despite DNAR orders. DNAR orders were most frequently observed among older patients, and the prevalence increased progressively with age. National OHCA survival rates are somewhat underestimated, as the inclusion of patients with DNAR orders lowers the reported figures; excluding these cases would yield higher survival estimates.

To describe the patient population, care pathways, and short-term safety outcomes of a regional hospital-at-home model in Region Zealand, Denmark, between 2022 and 2024, and to explore variation in outcomes by diagnosis and referral source. Retrospective observational study using registry-based data and internal monitoring records. A publicly funded, cross-sectoral hospital-at-home service coordinated centrally and delivered via municipal nurses and mobile treatment units. We included 2,511 patients with a median age of 72 years (IQR 60–80). Patients generally presented with substantial comorbidity and represented a broad, clinically diverse population. Length of stay, unplanned hospital contacts during treatment or within 24 h after planned completion, 30-day readmissions, and 30-day survival. The median length of stay increased from 5.1 days in 2022 to 7.3 days in 2024. Rates of unplanned hospital contacts remained stable (27.0–31.1

OBJECTIVES:Out-of-hours services (OOHS) worldwide exhibit diverse organisational models, especially within the European Union. This study aims to describe the transformation of OOHS in Region Zealand, Denmark, from a General Practitioner cooperative (GP-OOHS) to a regional organization/service, known as the 1818 Medical Helpline (1818). STUDY DESIGN:Retrospective cohort study. METHODS:GP-OOHS data (January 2017-October 2023) during the transition from GP-cooperative to Regional Service (October 2022) were analysed. Coded and timestamped services retrieved from the billing system were used to identify patient pathways, defined as services within a 12-hour window. Descriptive statistics were applied. RESULTS:A total of 2,572,951 patient pathways were recorded, with 2,258,072 under GP-out-of-hours service and 314,879 under the 1818. Overall, patient pathway volumes declined from 412,116 in 2017 to 314,879 in 2022, and admissions fell from 64,555 to 59,967. The median patient age was 35 years. The GP-out-of-hours service had a higher average monthly volume of patient pathways (32,726 vs. 25,940), while 1818 showed a higher proportion of emergency department admissions within 24 h (19.0% vs. 17.6%), the number of admissions fell from 8.2 per 100,000 inhabitants to 7.1 per 100,000 inhabitants. Interrupted time series analysis showed that the previously increasing tendency in ED admissions flattened following the transition. Although the absolute number of children admitted decreased, the proportion of admissions increased due to overall decline in patient pathways. Face-to-face consultations (31.4% vs. 27.1%) and home visits (9.4% vs. 3.4%) were more frequent in GP-out-of-hours service, both being associated with higher admission rates. CONCLUSIONS:OOHS patient pathways declined over the study period and continued to decline following the organizational change. The absolute number of emergency department admissions also decreased, while the relative proportion of admissions increased. These findings indicate a continued decline in OOHS activity and a slight shift toward more acute cases after the reorganisation.

Abstract Background Potassium is involved in multiple physiological processes in the body, and hyper-kalemia is a common, potentially life-threatening condition. Objective The aim of our study was to examine the association between plasma potassium levels, and 30-day mortality in patients presenting to an emergency department with normo- or hyperkalemia. Design Retrospective Cohort study. Setting Emergency Departments in the Capital region of Denmark Participants Persons attending Emergency Departments in the Capital Region of Denmark from 2017-2021 with a plasma potassium level of at least 3.5 mM measured within 4 hours after arrival. Measurements The study was based on data from Danish National Registries and electronic patient records. We performed Kaplan-Meier survival analyses and unadjusted and adjusted cox regression analyses utilizing plasma [K+] 3.5–4.4 mM as the reference group for 30-day mortality hazard ratios (HRs). Results A total of 248,453 patients were included with a median age of 60 years (Q1;Q3 42;75), and 6,959 (2.8%) died within 30 days. Mortality was 2.2% for potassium level 3.5–4.4 mM, 6.9% for 4.5–4.9 mM, 17.1% for 5.0–5.9 mM, and 26.9% for ≥6.0 mM. Unadjusted 30-day HRs were 3.2 (95%CI: 3.0–3.4) for [K+] 4.5–4.9 mM, 8.6 (95%CI: 7.9–9.3) for [K+] 5.0–5.9 mM, and 14.7 (95%CI: 12.5–17.0) for [K+] ≥6.0 mM. Adjusted HRs were 1.4 (1.3–1.5), 2.10 (1.9–2.3), and 2.4 (2.0–2.8), respectively. Limitations Risk of residual confounding. Missing data. No access to data regarding in-hospital treatment. Conclusion Plasma potassium levels above 4.4 mM were associated with increased 30-day mortality among patients presenting to emergency departments. Primary funding source Department of Emergency Medicine, Copenhagen University hospital, Bispebjerg and Frederiksberg Hospital.

Eye-related symptoms ranging from mild irritation to acute vision loss can lead to emergency calls. While many cases are benign, others involve chemical exposures or trauma requiring urgent care to prevent permanent damage. Prehospital characteristics, dispatch categorization, and early management of eye-related emergencies remain unexplored. The aim of this study was to provide a systematic description of eye-related emergency calls, including patient characteristics, dispatch categorization, and prehospital management, within the framework of the Danish Index for Emergency Care. We conducted a retrospective cohort study of all emergency calls [1–1–2] in Region Zealand, Denmark, from November 2017 to June 2025, describing eye symptoms. We linked data from the Computer-Aided Dispatch (CAD) system, the Prehospital Patient Records (PPJ), and the Danish National Patient Registry (NPR). Patient demographics, paramedic assessment of primary problem, on-scene symptoms, prehospital treatments and hospital outcomes were analysed using descriptive statistics. A total of 376 patients were identified calling 1-1-2 and registered with eye related symptoms as the main complaint. Of these, 36 (9.6

BACKGROUND:The PR interval is a ubiquitous parameter available to all clinicians. Studies on associated risk have been conflicting, often limited by sample sizes and one ECG per individual. The study aimed to investigate the association between the PR interval and its temporal changes with cardiac events in a nationwide cohort. METHODS:Nationwide hospital ECGs were linked with Danish national registries. Associations between index-PR interval and absolute temporal changes in PR interval (ΔPR), and cardiac events were modelled with multivariable cause-specific Cox models. RESULTS:A total of 9 020 051 ECGs (n=2 234 492; 53% female) were available, with 1 213 073 patients having >1 ECG. Median follow-up was 7.1 years. Median index-PR interval was 158 ms. PR interval <120 or >200 ms was found in 2.9% and 7.4%, respectively. Prevalence of prolonged PR interval increased with age. Median ΔPR was 10 ms and highest among the oldest. Cox models between the index-PR interval and events showed the following HR patterns: U-shaped for risk of atrial fibrillation/flutter, heart failure and ventricular arrhythmias (HR 1.03-1.24, 1.04-1.11, 1.08-1.16, respectively, p<0.05); stepwise increase at PR interval ≥170 ms for syncope (HR 1.08-1.36, p<0.001); linear increasing at PR intervals >160 ms (HR 1.07-2.11, p<0.05) for high-degree atrioventricular block/cardiac device implantation; J-shaped with moderate and marginal increases among the shortest (HR 1.44, p<0.001) and longest PR intervals (HR 1.03, p<0.001), respectively, for all-cause mortality. Models on ΔPR showed stepwise increases in HR for all events. CONCLUSION:Prevalence of PR prolongation and temporal PR changes increased with age. Short, long and temporal changes in the PR interval were all associated with increased risk of cardiac events and all-cause mortality.

OBJECTIVE:Traumatic cardiac arrest differs from non-traumatic regarding epidemiology. This study evaluated five machine learning classifiers' ability to identify traumatic cardiac arrest in the Danish Cardiac Arrest Registry, which currently relies on manual review. METHODS:This retrospective study employed split sampling to train and test models using medical records of cardiac arrest patients from 2016 to 2021. Data were preprocessed, resampled, and classified using five classification models. Shapley Additive Explanations values were used to explain the models' predictions except for the BERT model. RESULTS:30,171 medical records included, with 985 involving traumatic cardiac arrests. The histogram gradient boosting model achieved the highest F1-score of 0.62, with precision (positive predictive value) of 0.61 and recall (sensitivity) of 0.62. The BERT model demonstrated a recall of 0.96, a precision of 0.21, and an F1-score of 0.35. For histogram gradient boosting and random forest, the increasing presence of "car", "thorax", and "head" was linked to trauma. Further error analysis revealed that these models exhibited age-related bias by implicitly learning that traumatic cardiac arrest is more common in younger populations and less frequent in older populations. CONCLUSION:This study represents the first steps towards using machine learning algorithms to support manual validation of the Danish Cardiac Arrest Registry. Histogram gradient boosting yielded the best overall results; however, the BERT model demonstrated significantly improved recall in detecting trauma cases. Hence, applying the BERT model to the validation would substantially reduce the present workload.

Background: Risk assessment of patients with syncope does not consider a short PR interval despite its association with increased risk of atrial fibrillation and all-cause mortality. This study aimed to explore the association between the PR interval and all-cause mortality and recurrent syncope in patients admitted to the hospital with syncope. Methods: We included patients with a diagnosis of syncope and an ECG recorded within 24 hours of hospital admission from the Danish Nationwide Electrocardiogram Cohort and divided patients into short (<120 ms), normal (120-200 ms), or long PR interval (>200 ms). Patients with ECG abnormalities or comorbidities influencing the PR interval or outcomes were excluded. Results: A total of 52 038 patients were included. Adjusting for age, sex, and relevant covariates the highest hazard ratio (HR) was observed in patients with short PR interval with an HR of 1.50 (95% CI, 1.24-1.80, P<0.001). A long PR interval did not show an association with all-cause mortality (HR, 1.02 [95% CI, 0.97-1.08], P=0.3566). Adjusted 5-year cumulative incidence of all-cause mortality was 18% for short PR interval, 14% for normal PR interval, and 13% for long PR interval. Regarding recurrent syncope, a HR of 1.14 (95% CI, 1.09-1.20. P<0.001) was seen for long PR interval. Adjusted 5-year cumulative incidence of recurrent syncope was 23% in patients with a long PR interval. Conclusion: In patients with syncope, a short PR interval was associated with higher risk of all-cause mortality; however, a long PR interval was associated with increased rate of recurrent syncope.