In Denmark, 12,000 people experience a stroke every year. It is the fourth largest cause of death, and costs 4,6 billion Danish crowns in healthcare expenses and lost work income. It is also the leading cause of acquired disability for adults. Pre-hospital recognition of stroke is paramount to ensure fast and correct treatment for patients, in turn leading to better outcomes for patients. As the advanced treatment for ischemic stroke, thrombolysis and thrombectomy, is time-sensitive, even short delays in recognition and treatment can have a large effect on the individual stroke patient.

Paramedics on scene have only a few tools to assist them in recognizing stroke, where clinical scales such as the National Institutes of Health Stroke Scale (NIHSS), Face Arm Speech Time (FAST) or Prehospital Stroke Score (PRESS ) are most commonly used. Despite the use of such instruments, patients with stroke still go unrecognized, and as a result, the unrecognized patient might not be hospitalized, be hospitalized in a hospital without stroke facilities or be hospitalized too late for advanced treatment. Lower quality of communication between paramedics and the stroke centre significantly increases prehospital on-scene time. In a consensus statement from the European Academy of Neurology (EAN) and the European Stroke Organisation (ESO), training paramedics to recognise symptoms of all stroke types was strongly recommended. This study aims to explore whether trained paramedics using a mobile application with NIHSS and video communication to the in-hospital stroke physician may improve triage of acute stroke patients. This intervention will be compared to paramedics using standard procedure and communication through regular channels. It is hypothesized that the number of patients brought to the emergency department with suspected acute stroke and discharged with a stroke diagnosis is significantly higher in the app-group (intervention) compared to standard prehospital model (control).

1. Background 1.1 Stroke In Denmark, 12,000 people experience a stroke every year. It is the fourth largest cause of death, and costs 4,6 billion Danish crowns in healthcare expenses and lost work income. It is also the leading cause of acquired disability for adults.

Pre-hospital recognition of stroke is paramount to ensure fast and correct treatment for patients, in turn leading to better outcomes for patients. As the advanced treatment for ischemic stroke, thrombolysis and thrombectomy, is time-sensitive, even short delays in recognition and treatment can have a large effect on the individual stroke patient. Paramedics on scene have only a few tools to assist them in recognizing stroke, where clinical scales such as the National Institutes of Health Stroke Scale (NIHSS), Face Arm Speech Time (FAST) or Prehospital Stroke Score (PRESS ) are most commonly used. Despite the use of such instruments, patients with stroke still go unrecognized, and as a result, the unrecognized patient might not be hospitalized, be hospitalized in a hospital without stroke facilities or be hospitalized too late for advanced treatment.

A combination of early prehospital identification of stroke, triage to the right level of care and improvement of in-hospital measures to reduce door-to-needle time may result in more patients receiving acute treatment. Lower quality of communication between paramedics and the stroke centre significantly increases prehospital on-scene time. In a consensus statement from the European Academy of Neurology (EAN) and the European Stroke Organisation (ESO), training paramedics to recognise symptoms of all stroke types was strongly recommended.

This study aims to explore whether trained paramedics using a mobile application with NIHSS and digital communication may improve triage of acute stroke patients and ensure the standardised transfer of critical patient data to the in-hospital stroke physician. This intervention will be compared to paramedics using standard procedure with PRESS and communication through regular channels.

Investigators hypothesize that the number of patients brought to the emergency department (ED) with suspected acute stroke and discharged with a stroke diagnosis is significantly higher in the app-group (intervention) compared to standard prehospital model (control).

1.1.1 Incidence and mortality In Denmark, 17.647 patients were registered in 2020 with the diagnosis of Stroke and Transient ischemic attack (TIA) according to the national Danish stroke database DanStroke. Costs of Stroke are expected to increase by 44% by 2040 in Europe. A major part of stroke costs is due to post-stroke need of care due to loss of independency. Acute Stroke revascularization treatment using medical thrombolysis or mechanical endovascular treatment is a critical time-dependent situation as increasing time from symptom onset to start of treatment increases independency for patients post-stroke. Every minute there is a loss of approximately 2 MIO brain cells, hence the term "Time is Brain", and corresponds to more loss of independency in clinical outcome with increasing time from onset of symptoms

1.2 Prehospital assessment of patients with suspected stroke 1.2.1 The potential importance of prehospital assessment A combination of early prehospital identification of stroke, triage to the right level of care and improvement of in-hospital measures to reduce door-to-needle time may result in more patients receiving acute treatment. Lower quality of communication between paramedics and stroke centre significantly increases prehospital on-scene time. As mentioned, the European Academy of Neurology (EAN) and the European Stroke Organisation (ESO) strongly recommend training paramedics to recognise symptoms of all stroke types.

1.3 Current evidence concerning videoassisted prehospital assessment of patients with suspected stroke 1.3.1 Randomized clinical trials PubMed was searched with the search terms "prehospital stroke scales", "prehospital NIHSS", "non-physician NIHSS", "prehospital stroke assessment", "NIHSS in LVO", and "mobile stroke units". The focus was on studies reporting randomised controlled trials, stepped-wedge cluster randomised trials, clinical trials, and cohort studies as well as systematic reviews and meta-analyses of prehospital stroke care. A recent randomised trial, the ParaNASPP trial explored prehospital NIHSS as a common language in the acute stroke chain. This trial showed that introducing prehospital NIHSS with direct communication to the stroke physician, improved care by reducing in-hospital time to CT and by increasing prehospital identification of patients with low NIHSS and subtle symptoms. However, it did not increase diagnostic accuracy.

Another study, the PASTA trial, explored the implementation of a structured protocol and checklists for paramedic stroke assessment to increase thrombolytic rates in a randomised control design. The PASTA checklist was based on structured handover and clinical assessment with a face, arm, speech, time (FAST) test, and concluded that paramedic training in FAST alone did not significantly influence treatment rate. These prehospital stroke scales are usually modified versions of the National Institutes of Health Stroke Scale (NIHSS) and their main purpose is to enable identification of patients with large vessel occlusions (LVOs), eligible for endovascular thrombectomy.

In the PRESTO trial, Duvecot et al. compared the accuracy of eight prehospital stroke scales in detecting LVOs. Since LVOs occur in at most 30% of the general stroke population, most patients have non-LVO stroke with a heterogeneous symptom presentation. The NIHSS is the scale of choice for identification of both LVO and non-LVO strokes. Our literature search did not find any prehospital studies focusing on non-LVO symptoms in minor to moderate strokes. Several in-hospital conducted cohort and inter-rater agreement studies on the NIHSS have shown high levels of agreement when used by non-physicians. No studies considered whether prehospital NIHSS could be implemented in a large-scale prehospital system; however, promising results were presented in cohort studies from the helicopter emergency medical service and mobile stroke units. The prehospital NIHSS studies had poor methodological robustness because of their small sample size and non-randomised design.

1.5 Current practice Currently, paramedics meeting a patient with suspected stroke assess the patients with the PRESS-scale and when deemed necessary the comprehensive stroke centre is contacted, and a teleconference between paramedics and the attending neurologist is initiated. In case of suspected and prehospitally confirmed stroke suspicion, the patient is transported to the comprehensive stroke centre in Roskilde for further evaluation and treatment.

If a stroke is identified at the hospital, it is registered in The Danish Stroke Registry (DanStroke).

Inclusion Criteria:

• A. Suspected stroke by paramedic or dispatcher

Exclusion Criteria:

• Technical issues (ie. the telephone does not work, the app does not work).
• Patients who are incarcerated
• Patients seen by a physician prior to assessment by paramedics.
• Subarachnoid Haemorrhage strokes

Data will be completely anonymized according to Danish law before analysis. Due to Danish data regulations, data is not publically available, but access to the registry data used in this project can be applied for by appropriate official channels. Anonymised data from the app can be made available upon reasonable request after the publication of the results.

All relevant project-related documents will be made available upon reasonable request. Data will be available for any research purpose to all interested parties who have approval from an independent review committee and who have a methodological sound proposal as determined by the steering committee of the current project. Interested parties will be able to request the project specific data by contacting the principal investigator, and registry data by appropriate official channels.