Stroke Prehospital Decision Support Systems Based on Artificial

Intelligence: Grey Literature Scoping Review

Hoor Jalo

, Eunji Lee

, Mattias Seth

, Anna Bakidou

1,2 c

, Minna Pikkarainen

1,3

Katarina Jood

4,5 d

, Bengt Arne Sjöqvist

and Stefan Candefjord

Department of Electrical Engineering, Chalmers University of Technology, Gothenburg, Sweden

PreHospen – Centre for Prehospital Research, University of Borås, Borås, Sweden

Department of Occupational Therapy, Oslo Metropolitan University, Oslo, Norway

Department of Clinical Neuroscience, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden

Department of Neurology, Sahlgrenska University Hospital, Gothenburg, Sweden

katarina.jood@neuro.gu.se

Keywords: Artificial Intelligence (AI), Clinical Decision Support Systems (CDSSs), Grey Literature, Machine Learning

(ML), Prehospital Care, Stroke.

Abstract: Stroke is a leading cause of mortality and disability worldwide. Therefore, there is a growing interest in

prehospital point-of-care stroke clinical decision support systems (CDSSs), which with improved precision

can identify stroke and decrease the time to optimal treatment, thereby improving clinical outcomes. Artificial

intelligence (AI) may be a route to improve CDSSs for clinical benefit. Deploying AI in the area of prehospital

stroke care is still in its infancy. There are several existing systematic and scoping reviews summarizing the

progress of AI methods for stroke assessment. None of these reviews include grey literature, which could be

a valuable source of information, especially when analysing future research and development directions. This

paper aims to use grey literature to investigate stroke assessment CDSSs based on AI. The study adheres to

PRISMA guidelines and presents seven records showcasing promising technologies. These records included

three clinical trials, two smartphone applications, one master thesis and one PhD dissertation, which identify

electroencephalogram (EEG), video analysis and voice and facial recognition as potential data sources for

early stroke identification. The integration of these technologies may offer the prospect of faster and more

accurate CDSSs in the future.

1 INTRODUCTION

Stroke is a leading cause of death and disability

worldwide (Chennareddy et al., 2022). It is caused by

either a bleeding, called haemorrhagic stroke, or a

clot in one or more of the brain’s blood vessels, called

ischemic stroke which accounts for approximately

85% of cases (Meyran et al., 2020). For ischemic

stroke, two types of treatment are deployed:

thrombolysis (clot-dissolving drugs) or

thrombectomy (mechanically removing the clot)

http://orcid.org/0000-0001-6975-8520

http://orcid.org/0000-0002-3737-3316

http://orcid.org/0000-0002-4288-8146

http://orcid.org/0000-0001-8746-1771

http://orcid.org/0000-0002-6564-737X

http://orcid.org/0000-0001-7942-2190

(Lumley et al., 2020; Shlobin et al., 2022). The latter

is an advanced treatment that requires a specialist at a

comprehensive stroke centre (CSC) and is suited for

thrombectomy candidates, i.e., patients with large

vessel occlusion (LVO) with still viable brain tissue

that can be restored (Chennareddy et al., 2022;

Nicholls et al., 2022). Brain cells die as time passes

without treatment; hence, quick treatment is a critical

factor (Shlobin et al., 2022).

Most stroke patients are initially handled by a

prehospital team in an ambulance according to

458

Jalo, H., Lee, E., Seth, M., Bakidou, A., Pikkarainen, M., Jood, K., Sjöqvist, B. and Candefjord, S.

Stroke Prehospital Decision Support Systems Based on Artiﬁcial Intelligence: Grey Literature Scoping Review.

DOI: 10.5220/0012380400003657

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 17th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2024) - Volume 2, pages 458-465

ISBN: 978-989-758-688-0; ISSN: 2184-4305

Figure 1: Drip-and-ship and mothership strategies.

predefined procedures including medical history,

patient assessment, glucose and vital signs

monitoring. In addition, various clinical stroke scales

are used to predict the likelihood of stroke and its

severity. These scales are symptom-based tests that

assess the patient’s stroke symptoms such as speech,

facial expression and arm and leg movements

(Nicholls et al., 2022). In hospital settings, the most

commonly used scale is called the National Institutes

of Health Stroke Scale (NIHSS); it is a validated scale

with a sensitivity of 86% and a specificity of 60% for

LVO (Nicholls et al., 2022). In prehospital settings,

the observation for NIHSS is too complicated to carry

out for the ambulance teams, therefore simpler,

mostly unvalidated scales, e.g., Los Angeles Motor

Scale, are frequently used. The accuracy of

prehospital stroke diagnosis, when relying solely on

stroke scales, remains low, with a 64% sensitivity

(Chennareddy et al., 2022), suggesting the need for

considering additional diagnostic methods or factors

in enhancing diagnostic accuracy.

It is crucial to streamline prehospital stroke

management (Fassbender et al., 2020). Currently,

stroke is typically confirmed at hospitals after brain

imaging is performed, such as computed tomography

(CT) and magnetic resonance imaging (MRI)

(Magnusson et al., 2022). According to international

stroke management guidelines, such as those

recommended by the American Stroke Association,

patients with suspected stroke are typically

transported to the nearest hospital. If the nearest

hospital is not a CSC and LVO is confirmed, the

patient should be transferred to a CSC where

thrombectomy can be performed (Nicholls et al.,

2022). This strategy is called drip-and-ship (Figure

1). Patients with a high probability of LVO may be

transported directly to the CSC, which is called the

mothership strategy (Fassbender et al., 2020; Nicholls

et al., 2022). Identification of more LVO patients in

the prehospital setting is crucial to reducing

prehospital delays and improving patient outcomes

(Nicholls et al., 2022).

The research group Care@Distance–Remote and

Prehospital Digital Health at Chalmers University of

Technology is focused on developing AI-based

prehospital clinical decision support systems

(CDSSs) for acute diseases, such as stroke. Our aim

is to enhance the performance in the identification of

LVO and provide support to reduce treatment delays.

This is in line with the American Heart Association

Guidelines (2019), which called for further research

to identify effective prehospital procedures for

triaging patients to the appropriate centers, including

hospital bypass algorithms (Nicholls et al., 2022).

Artificial intelligence (AI) and machine learning

(ML) deployed in CDSS can play an important role

in improving stroke assessment (Murray et al., 2020;

Shlobin et al., 2022). AI encompasses computer tasks

challenging for humans, including CDSSs, which

could be defined to fall within the AI realm although

Stroke Prehospital Decision Support Systems Based on Artiﬁcial Intelligence: Grey Literature Scoping Review

459

employing basic algorithms influenced by clinical

experience. The focus of this study is however

specifically on advanced ML algorithms,

characterized by their complex model architectures

and data-driven techniques, as we investigate their

potential for enhancing early stroke characterization.

Several recent scoping and systematic reviews

have summarized the progress of AI methods for

stroke assessment; however, none included grey

records (Chennareddy et al., 2022; Lumley et al.,

2020; Murray et al., 2020; Nicholls et al., 2022;

Ruksakulpiwat et al., 2021; Shlobin et al., 2022).

Grey literature can be a rich source of information

about solutions in premature stages, and the Institute

of Medicine Standards for Systematic Review (Berg

et al., 2011) and the Cochrane Handbook for

Systematic Reviews of Interventions (Higgins et al.,

2023) recommend incorporating grey literature in

systematic reviews. Our group is currently working

on a scoping review that is centered on the peer-

reviewed literature (Jalo et al., 2023). This study

therefore aims to complement our review and existing

reviews by specifically focusing on publicly available

grey literature that includes, among others, academic

papers (theses and dissertations), non-peer-reviewed

conference proceedings, research and committee

reports, government reports, clinical trials and

ongoing studies to identify promising AI-based

CDSSs for prehospital stroke assessment.

2 METHODS

Arksey and O'Malley (Arksey et al., 2005) describe a

scoping review as a form of literature review

designed to map the relevant literature within a

specific research domain. This scoping review was

carried out in accordance with Arksey and O’Malley

(Arksey et al., 2005), a five-stage methodology

including (1) identifying the research question(s), (2)

identifying relevant studies, (3) selecting studies, (4)

charting the data and (5) collating, summarizing and

reporting the results. The optional consultation stage

was not conducted in this scoping review because it

is challenging to evaluate new AI-based CDSSs due

to the intricate algorithms and methods involved.

The methodology started with identifying the

research question, and this review aims to answer the

following question: what are the promising AI-based

CDSSs for stroke assessment in prehospital settings?

To answer the identified research question, relevant

studies were identified by searching 12 databases,

focusing only on recent grey literature such as non-

peer-reviewed conference proceedings, theses,

dissertations and reports during the past 10 years. The

searched databases (ClinicalTrial, ProQuest, Arxiv,

EBSCOhost, Networked Digital Library of Theses

and Dissertations (NDLTD), University of

Wollongong (UOW) library, Bielefeld Academic

Search Engine (BASE), World Health Organization

(WHO), Canada’s drug and health technology agency

(CADTH), FiNDit, Open Access Theses and

Dissertations (OATD), University of South

California (USC) Library and Science daily) were

selected based on grey literature recommendations

published by James Cook University’s library (James

Cook University). Google Scholar was not included

in this study because a study found that most of the

not retrieved records in Google Scholar were grey

literature (Yasin et al., 2020).

Search terms included stroke, haemorrhagic

stroke, ischemic stroke, artificial intelligence,

machine learning and decision support systems,

which were systematically combined using Boolean

operators (AND, OR) to capture relevant studies.

Search strings were defined and adopted for each

database.

The third stage was to select studies based on

identified eligibility criteria (Table 1). Article

screening was done in two steps: (1) title and abstract

screening, and (2) full text screening.

Table 1: Eligibility criteria.

Inclusion criteria Exclusion criteria

Non-peer reviewed

records

Peer-reviewed journal and

peer-reviewed conference

articles

The record reports a

method for stroke

assessment

No focus on stroke

Stroke related to animal

studies

The record presents an

AI-based tool for stroke

assessment in the

prehospital setting

No AI-based method is

reported, or the method

cannot be used in the

prehospital setting, e.g.,

the use of brain imaging

is required

Written in English Not written in English

Published between 2012

and 2022

Older records or non-

retrievable

The included records were then summarized,

charted and reported. An Excel sheet was used to

extract the following data from each included study:

key information, aims, population and study subject

characteristics, methodology, main findings and

limitations of the study. By applying a consistent

approach to each included study, common

characteristics and descriptive summary were

provided.

HEALTHINF 2024 - 17th International Conference on Health Informatics

460

3 RESULTS

A total of 1593 grey literature records were identified

during Stage 2 (Figure 2). Initial title and abstract

screening and duplicate removal resulted in 126

records passed for full-text screening. Among these,

119 did not meet the inclusion criteria, including

studies published in peer-reviewed journals, not

retrieved data and records irrelevant to the research

questions. Seven relevant records were included and

are summarized in Table 2. No records were found

between 2012 and 2018.

An ongoing clinical trial in the Netherlands tests

the accuracy of AI-STROKE algorithms: one or more

novel AI-based electroencephalography (EEG)

algorithms to detect LVO in ambulances (Coutinho,

2022). EEG is performed using dry electrode EEG

caps, and investigators expect it to be done in less

than five minutes in ambulances (Coutinho, 2022).

The presence or absence of a clot will be confirmed

using CT scans at the emergency department (ED)

(Coutinho, 2022). The accuracy of the new

algorithms will be tested by comparing the area under

the receiver operating characteristic curve (AUC),

sensitivity, specificity, positive predictive value

(PPV) and negative predictive value (NPV) to

existing EEG algorithms (Coutinho, 2022).

Another ongoing clinical trial is a prospective

observational study in Sweden, which evaluates

measurements from EEG, heart rate variability

(HRV) and near-infrared spectroscopy (NIRS)

separately and combined to provide a specific

detection of cerebral ischemia (Block et al., 2020;

Block, 2022). The investigators hypothesize that

changes in those measurements can indicate cerebral

ischemia and reperfusion after being processed by

ML classification models. The study aims to find

specific patterns in EEG, HRV and NIRS to signify

cerebral ischemia and build a monitoring warning

system for the diagnosis of upcoming cerebral

ischemia (Block et al., 2020; Block, 2022).

Figure 2: Modified PRISMA flowchart to summarize retrieved records (Page et al., 2021).

Stroke Prehospital Decision Support Systems Based on Artiﬁcial Intelligence: Grey Literature Scoping Review

461

An ongoing clinical trial in Norway (Haukeland

University Hospital, 2022) works on an AI-based

prototype for acute stroke detection in emergency

medical calls. The study aims to use audio from

emergency calls and available data in hospital records

to implement an AI-based detection system and test

its performance against the current system by

comparing sensitivity and specificity.

Facial palsy is a common symptom in stroke

patients; facial recognition is thus an important

technique to be used in the detection of stroke. A

master thesis from TU Delft presented a system that

automatically detects facial paralysis based on deep

learning algorithms (Sourlos, 2020). The algorithm

first detects a patient’s face from an image, then

metrics are defined based on landmarks localized on

the face, to classify the input image (Sourlos, 2020).

In total, 203 images were analysed (60 of healthy

subjects, 40 of central palsy patients and 103 of

peripheral palsy patients), and an accuracy of 99.5%

was achieved for the healthy group and 90.9% for

patients with facial palsy (Sourlos, 2020).

Another promising technology is a mobile

application called Fatal Recognition that uses facial

recognition to detect early signs of stroke (Chan,

2019). Every time patients unlock their screen, an

algorithm detects signs of face-drooping, sends alerts

to contact emergency services and shows emergency

services contact details (Chan, 2019).

Table 2: Summary of the data extracted for novel CDSSs included in the scoping review.

Record

Study

design

Year Location Goal Methodology

Algorithm Development

Through AI for the Triage

of Stroke Patients in the

Ambulance with EEG

(AI-STROKE)

(Coutinho, 2022).

Clinical

trial

2022 Netherlands LVO detection in

ambulances

AI-based EEG

algorithms with the

use of dry electrode

caps

Detection of Cerebral

Ischemia with Artificial

Intelligence

(Block, 2022)

Clinical

trial

2022 Sweden Detection of cerebral

ischemia

ML models based on

EEG, HRV and NIRS

Artificial Intelligence

Support in Stroke Calls

(AISIS)

(Haukeland University

Hospital, 2022)

Clinical

trial

2022 Norway Acute stroke detection in

emergency medical calls

AI-based detection

system using audio

calls and health

records

Facial Imaging and

Diagnosis System for

Neurological Disorders

(Sourlos, 2020)

Master

thesis

2020 Netherlands Automatic detection of

facial paralysis

Deep learning-based

system

Fatal Recognition

(Chan, 2019)

Mobile

application

2019 Hong Kong Detection of early signs of

facial drooping in stroke

AI-based facial

drooping detection

AI-Stroke

(AI-Stroke, 2022)

Mobile

application

2022 France Indication of stroke and

characterization the stroke

type

Video analysis and

ML to perform NIHSS

scale

Human-Centred Machine

Learning for Healthcare:

Examples in Neurology

and Pulmonology

(Ramesh, 2020)

PhD

dissertation

2020 USA Hemiparesis detection in

stroke patients

Video analysis and

HEALTHINF 2024 - 17th International Conference on Health Informatics

462

A company called AI-Stroke developed a

smartphone application to perform the NIHSS test,

and it guides the person close to the patient on how to

perform the test while recording a video of the patient

(AI-Stroke, 2022). The AI algorithm calculates the

NIHSS score, indicates the probability of having a

stroke and characterizes the stroke type. They aim to

create a dataset with videos from stroke patients and

healthy volunteers to be used in training the AI

model. An ethical application has been approved in

France, and the first patient was recently added to the

dataset Franc (AI-Stroke, 2022).

Video analysis based on ML was used to detect

hemiparesis in stroke patients in a sitting position as

a part of a PhD dissertation at the University of

California San Diego (Ramesh, 2020). Hemiparesis is

usually identified by the NIHSS test by asking the

patient to move the arm or leg (Ramesh, 2020). The

system was tested by eight stroke specialists, and a

video-based assessment was done as part of the

NIHSS test at rest in the sitting position. The accuracy

of the system was 68% and 61% when moving and at

rest, respectively (Ramesh, 2020).

4 DISCUSSIONS

Integration of innovative technology in prehospital

clinical assessment may lead to early stroke detection,

classification of stroke subtypes and fewer false-

positive stroke diagnosis. AI-based prehospital

CDSSs for stroke play a crucial role in the context of

embedded decision support systems. These

innovative technologies leverage the power of AI to

assist ambulance teams in making quick and accurate

decisions during the critical prehospital phase of

stroke care. By analysing various patient data, these

systems can quickly identify potential stroke cases,

prioritize them based on severity and provide real-

time recommendations for appropriate interventions

and transportation protocols, thereby allowing for

faster treatment and improved clinical outcomes.

The seven included records have highlighted

several methods for early stroke detection in the

prehospital environment, including EEG brain

imaging techniques and automated symptom-based

tools. Two clinical studies have used EEG, which has

a long-established sensitivity for early stroke

detection and has the potential to be used in

prehospital settings (Block et al., 2020; Erani et al.,

2020). EEG is a physiological monitoring technique

that is used to record the brain’s electrical activity and

can immediately detect changes in brain function

(Block et al., 2020; Erani et al., 2020). One of its

limitations is the long time spent in applying gel, but

this could be overcome by using rapidly applied dry

electrodes (Erani et al., 2020). A recent study

indicated that EEG signals contain further diagnostic

information compared to the current clinical

assessment (Erani et al., 2020). The AUC for acute

stroke diagnosis was 87.8 and 86.4 for LVO patients

when EEG was incorporated into the clinical routine

of stroke diagnosis (Erani et al., 2020).

Speech, facial palsy and impaired movement are

early stroke symptoms and are currently evaluated by

symptom-based clinical stroke scales. Many of the

stroke cases remain undetected in the prehospital

settings (Fassbender et al., 2020). For example, a

study compared the accuracy of 13 clinical stroke

scales for detecting thrombectomy candidates and

showed that 20% of thrombectomy candidates

remained undetected by the use of clinical stroke

scales (Fassbender et al., 2020). The recent

technologies focusing on early detection of these

symptoms thus have the potential to provide more

informative CDSS to improve prehospital stroke

detection, allowing for faster treatment and improved

clinical outcomes.

After applying the eligibility criteria, seven papers

were included in this study due to its narrow focus,

highlighting the need for further research in the

critical field of prehospital stroke assessment. This

aligns with a recent scoping review (Nicholls et al.,

2022) that aimed to identify LVO detection

techniques across various settings and included just

nine studies on AI-based methods. Out of these

studies, only three focused on AI-based prehospital

triaging tools. These tools included a smartphone

application designed to aid emergency medical

services professionals in patient assessment and

destination triage (Nogueira et al., 2017), a prediction

model that incorporated various predictors (Chen et

al., 2018) and a three-step triage tool aimed at

reducing prehospital assessment time (Zhao et al.,

2018). Notably, none of the studies presented AI-

based solutions involving EEG, video analysis, or

voice recognition for early stroke identification. This

study has thus presented early-stage research and

innovative concepts.

The main strength of this review is that it provides

information about the innovative technologies

usually not covered in reviews, with the potential to

improve stroke detection in prehospital settings.

Notably, this study pioneers the examination of a

research area that has not been well-explored, as

evidenced by the absence of records from 2012 to

2018. The results may thus serve as a basis for

developing an AI-based CDSS for early assessment

Stroke Prehospital Decision Support Systems Based on Artiﬁcial Intelligence: Grey Literature Scoping Review

463

of stroke. Limitations include that this study has been

conducted by only one reviewer, and many of the

presented results are ongoing clinical trials or yet-to-

be-performed studies. Those studies have different

study design, population and diagnostic accuracy

metrics which makes it difficult to perform

technology comparisons. Grey literature also captures

information related to emerging research areas that

are not yet published in peer-reviewed literature,

which makes a direct comparison with published

literature not feasible. However, it serves as a

valuable source of information by providing

preliminary research findings and insights into

emerging technologies. Future work may include

searching Google Scholar to see if more promising

techniques would be captured and widen the scope of

the review to include promising detection

technologies in hospitals. Future work also includes

searching the peer-reviewed literature to identify AI-

based CDSSs designed for assessing stroke in the

prehospital environment (Jalo et al., 2023).

5 CONCLUSIONS

This review has explored grey literature and

advancements in AI-based stroke assessment

methods in prehospital settings. It was found that in

future stroke assessments CDSS, EEG, audio

recognition, facial recognition and video records may

be used as data sources for decision-making. We

conclude that the reviewed technologies are

promising prehospital tools that have the potential to

aid in the early assessment of stroke, but they are yet

to be tested and validated. There are, however, not

many studies in this research area. More studies are

warranted due to the large clinical need.

Improvements in CDSSs would impact stroke

detection and patient triage, increasing the chance for

earlier initial treatment and the potential for fewer

secondary transfers of patients.

REFERENCES

AI-Stroke | Stroke Detection | Our Solution. (2022). AI-

Stroke. Retrieved from https://www.ai-stroke.com/our-

solution

Arksey, H., & O’Malley, L. (2005). Scoping studies:

towards a methodological framework. International

Journal of Social Research Methodology, 8(1), 19–32.

doi: 10.1080/1364557032000119616

Berg, A., Eden, J., Research, I. of M. (US) C. on S. for S.

R. of C. E., Levit, L., & Morton, S. (2011). Committee

On Standards For Systematic Reviews of Comparative

Effectiveness Research. In Finding What Works in

Health Care: Standards for Systematic Reviews.

National Academies Press (US). Retrieved from

https://www.ncbi.nlm.nih.gov/books/NBK209524/

Block, L. (2022). Detection of Cerebral Ischemia With

Artificial Intelligence- Biomarkers as Indicators in

Controlled Cerebral Ischemia and Reperfusion.

(Clinical Trial Registration No. NCT03919370).

clinicaltrials.gov. Retrieved from

https://clinicaltrials.gov/study/NCT03919370

Block, L., El-Merhi, A., Liljencrantz, J., Naredi, S., Staron,

M., & Odenstedt Hergès, H. (2020). Cerebral ischemia

detection using artificial intelligence (CIDAI)—A

study protocol. Acta Anaesthesiologica Scandinavica,

64(9), 1335–1342. doi: 10.1111/aas.13657

Chen, Z., Zhang, R., Xu, F., Gong, X., Shi, F., Zhang, M.,

& Lou, M. (2018). Novel Prehospital Prediction Model

of Large Vessel Occlusion Using Artificial Neural

Network. Frontiers in Aging Neuroscience, 10, 181.

doi: 10.3389/fnagi.2018.00181

Chennareddy, S., Kalagara, R., Smith, C., Matsoukas, S.,

Bhimani, A., Liang, J., Shapiro, S., De Leacy, R.,

Mokin, M., Fifi, J. T., Mocco, J., & Kellner, C. P.

(2022). Portable stroke detection devices: a systematic

scoping review of prehospital applications. BMC

Emergency Medicine, 22(1), 111. doi: 10.1186/s12873-

022-00663-z

Coutinho, J. (2022). Algorithm Development Through

Artificial Intelligence for the Triage of Stroke Patients

in the Ambulance With Electroencephalography

(Clinical Trial Registration No. NCT05437237).

clinicaltrials.gov. Retrieved from https://clinicaltrials.

gov/study/NCT05437237

Erani, F., Zolotova, N., Vanderschelden, B., Khoshab, N.,

Sarian, H., Nazarzai, L., Wu, J., Chakravarthy, B.,

Hoonpongsimanont, W., Yu, W., Shahbaba, B.,

Srinivasan, R., & Cramer, S. C. (2020). EEG Might

Improve Diagnosis of Acute Stroke and Large Vessel

Occlusion. Stroke, 51(11), 3361–3365. doi:

10.1161/STROKEAHA.120.030150

Fassbender, K., Walter, S., Grunwald, I. Q., Merzou, F.,

Mathur, S., Lesmeister, M., Liu, Y., Bertsch, T., &

Grotta, J. C. (2020). Prehospital stroke management in

the thrombectomy era. The Lancet. Neurology, 19(7),

601–610. doi: 10.1016/S1474-4422(20)30102-2

FATAL RECOGNITION. (2019). Mywebsite (New).

Retrieved from https://www.paulchancreative.com/

fatal-recognition

Haukeland University Hospital. (2022). Artificial

Intelligence (AI) Support in Stroke Calls - “The AISIS-

study” -Can Artificial Intelligence Improve the

Precision in Identifying Acute Stroke in Emergency

Medical Calls? (Clinical Trial Registration No.

NCT04648449). clinicaltrials.gov. Retrieved from

https://clinicaltrials.gov/study/NCT04648449

Higgins, J., J, T., & J, Chandler. (2023). Cochrane

Handbook for Systematic Reviews of Interventions.

Retrieved from https://training.cochrane.org/handbook

HEALTHINF 2024 - 17th International Conference on Health Informatics

464

Jalo, H., Seth, M., Pikkarainen, M., Häggström, I., Jood, K.,

Bakidou, A., Sjöqvist, B., & Candefjord, S. (2023).

Early identification and characterisation of stroke to

support prehospital decision-making using artificial

intelligence: a scoping review protocol. BMJ Open

13(5). doi: 10.1136/bmjopen-2022-069660

Lumley, H. A., Flynn, D., Shaw, L., McClelland, G., Ford,

G. A., White, P. M., & Price, C. I. (2020). A scoping

review of pre-hospital technology to assist ambulance

personnel with patient diagnosis or stratification during

the emergency assessment of suspected stroke. BMC

Emergency Medicine, 20(1), 30. doi: 10.1186/s12873-

020-00323-0

Magnusson, C., Herlitz, J., Sunnerhagen, K. S., Hansson,

P.-O., Andersson, J.-O., & Jood, K. (2022). Prehospital

recognition of stroke is associated with a lower risk of

death. Acta Neurologica Scandinavica, 146(2), 126–

136. doi: 10.1111/ane.13618

Meyran, D., Cassan, P., Avau, B., Singletary, E., &

Zideman, D. A. (2020). Stroke Recognition for First

Aid Providers: A Systematic Review and Meta-

Analysis. Cureus, 12(11), e11386. doi: 10.7759/

cureus.11386

Murray, N. M., Unberath, M., Hager, G. D., & Hui, F. K.

(2020). Artificial intelligence to diagnose ischemic

stroke and identify large vessel occlusions: a systematic

review. Journal of NeuroInterventional Surgery, 12(2),

156–164. doi: 10.1136/neurintsurg-2019-015135

Nicholls, J. K., Ince, J., Minhas, J. S., & Chung, E. M. L.

(2022). Emerging Detection Techniques for Large

Vessel Occlusion Stroke: A Scoping Review. Frontiers

in Neurology, 12, 780324. doi: 10.3389/fneur.2021.780

324

Nogueira, R. G., Silva, G. S., Lima, F. O., Yeh, Y.-C.,

Fleming, C., Branco, D., Yancey, A. H., Ratcliff, J. J.,

Wages, R. K., Doss, E., Bouslama, M., Grossberg, J.

A., Haussen, D. C., Sakano, T., & Frankel, M. R.

(2017). The FAST-ED App: A Smartphone Platform

for the Field Triage of Patients With Stroke. Stroke,

48(5), 1278–1284. doi: 10.1161/STROKEAHA.116.01

6026

Page, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I.,

Hoffmann, T. C., Mulrow, C. D., Shamseer, L.,

Tetzlaff, J. M., Akl, E. A., Brennan, S. E., Chou, R.,

Glanville, J., Grimshaw, J. M., Hróbjartsson, A., Lalu,

M. M., Li, T., Loder, E. W., Mayo-Wilson, E.,

McDonald, S., … Moher, D. (2021). The PRISMA

2020 statement: an updated guideline for reporting

systematic reviews. BMJ, 372, n71. doi:

10.1136/bmj.n71

Ramesh, V. (2020). Human-Centered Machine Learning

for Healthcare: Examples in Neurology and

Pulmonology [ProQuest Dissertations Publishing].

Retrieved from https://search.proquest.com/docview/

2456394283?pq-origsite=primo

Ruksakulpiwat, S., Thongking, W., Zhou, W., Benjasirisan,

C., Phianhasin, L., Schiltz, N. K., & Brahmbhatt, S.

(2021). Machine learning-based patient classification

system for adults with stroke: A systematic review.

Chronic Illness, 17423953211067435. doi:

10.1177/17423953211067435

Shlobin, N. A., Baig, A. A., Waqas, M., Patel, T. R.,

Dossani, R. H., Wilson, M., Cappuzzo, J. M., Siddiqui,

A. H., Tutino, V. M., & Levy, E. I. (2022). Artificial

Intelligence for Large-Vessel Occlusion Stroke: A

Systematic Review. World Neurosurgery, 159, 207-

220.e1. doi: 10.1016/j.wneu.2021.12.004

Sources of Grey Literature - Finding Grey Literature -

Library Guides at James Cook University. Retrieved

from https://libguides.jcu.edu.au/grey-lit/sources

Sourlos, N. (2020). Facial Imaging and Diagnosis System

for Neurological Disorders. Retrieved from

https://repository.tudelft.nl/islandora/object/uuid%3Ae

60a60fc-073c-48b4-83af-d5823f309539

Yasin, A., Fatima, R., Wen, L., Afzal, W., Azhar, M., &

Torkar, R. (2020). On Using Grey Literature and

Google Scholar in Systematic Literature Reviews in

Software Engineering. IEEE Access, 8, 36226–36243.

doi: 10.1109/ACCESS.2020.2971712

Zhao, H., Pesavento, L., Coote, S., Rodrigues, E., Salvaris,

P., Smith, K., Bernard, S., Stephenson, M., Churilov,

L., Yassi, N., Davis, S. M., & Campbell, B. C. V.

(2018). Ambulance Clinical Triage for Acute Stroke

Treatment: Paramedic Triage Algorithm for Large

Vessel Occlusion. Stroke, 49(4), 945–951. doi:

10.1161/STROKEAHA.117.019307

Stroke Prehospital Decision Support Systems Based on Artiﬁcial Intelligence: Grey Literature Scoping Review

465