A Health Technology Assessment of the Strokefinder MD100 For Early Detection of Stroke and Traumatic Brain Injury in the Western Cape Healthcare System, South Africa Degree project report in Biomedical Engineering Ebba Alvaeus Tynnerstål Alice Thornander DEPARTMENT OF ELECTRICAL ENGINEERING CHALMERS UNIVERSITY OF TECHNOLOGY Gothenburg, Sweden 2025 www.chalmers.se www.chalmers.se Degree project report 2025 A Health Technology Assessment of the Strokefinder MD100 For Early Detection of Stroke and Traumatic Brain Injury in the Western Cape Healthcare System, South Africa Ebba Alvaeus Tynnerstål Alice Thornander Department of Electrical Engineering Chalmers University of Technology Gothenburg, Sweden 2025 A Health Technology Assessment of the Strokefinder MD100 For Early Detection of Stroke and Traumatic Brain Injury in the Western Cape Healthcare System, South Africa Ebba Alvaeus Tynnerstål Alice Thornander © Ebba Alvaeus Tynnerstål, Alice Thornander, 2025. Supervisor: Professor Mikael Persson, Head of Division at Signal processing and Biomedical Engineering Examiner: Andreas Fhager, Head of Unit at Signal Processing and Biomedical Engineering Degree project report 2025 Department of Electrical Engineering Chalmers University of Technology SE-412 96 Gothenburg Sweden Telephone +46 73 849 0900 Cover: Image of the Strokefinder MD100 by Medfield Diagnostics AB, used with permission Typeset in LATEX, template by Kyriaki Antoniadou-Plytaria Gothenburg, Sweden 2025 iv A Health Technology Assessment of the Strokefinder MD100 For Early Detection of Stroke and Traumatic Brain Injury in the Western Cape Healthcare System, South Africa Ebba Alvaeus Tynnerstål, Alice Thornander Department of Electrical Engineering Chalmers University of Technology Abstract Stroke is the second leading cause of death worldwide, with particularly high preva- lence in low- and middle-income countries (LMIC) such as South Africa. Traumatic Brain Injury (TBI) accounts for approximately six million deaths annually and is commonly caused by trauma, an especially significant issue in South Africa (SA), where rates of interpersonal violence and traffic accidents are notably high. The South African healthcare system consists of a private and a public sector. The pri- vate sector are profit-driven hospital groups, while the public sector is government- funded. The public system operates on a referral-based model, which is not well- suited to managing time-sensitive medical conditions such as stroke and TBI. This report presents a Health Technology Assessment (HTA) of the Strokefinder MD100 device developed by the Swedish company Medfield Diagnostics AB, within the context of the healthcare system in the Western Cape, South Africa. The HTA is based on an extensive literature review combined with semi-structured qualita- tive interviews conducted with relevant stakeholders. The findings support the final recommendations for optimal implementation sites of the MD100 with the main purpose of minimising time-to-treatment and thereby improve patient outcomes. With the identified healthcare needs in the Western Cape, the most appropriate implementation sites are identified to be in pre-hospital units, within both private and public healthcare sectors, or in level 1 hospital facilities lacking CT imaging capabilities within the public healthcare sector. Keywords: Strokefinder MD100, HTA, Assessment, Western Cape South Africa, Triage, TBI, Medfield Diagnostics, Stoke, Truma, Healthcare. v Acknowledgements We would like to begin by expressing our sincere gratitude to our supervisor, Profes- sor Emeritus Mikael Persson, for his steadfast support throughout the project, his participation in interviews, and his invaluable insights at every stage. We are also deeply grateful to Professor Sara Grobblear for her support during our time at Stel- lenbosch University. Our heartfelt thanks go to John Paul Kulumba and Tinashe Chikunichawa for generously sharing their experiences and reflections from their own work, which significantly enriched our understanding. We extend our deepest appreciation to the Global Mentorship Program, without your financial support, this thesis would not have been possible. We are also thankful to our mentors, Karin and Hanna, for their ongoing encouragement and guidance throughout the project. Finally, we would like to express our sincere thanks to the remarkable individuals who participated in our interviews. We are truly grateful for the time you dedicated and the invaluable knowledge you shared with us. Ebba Alvaeus Tynnerstål, Alice Thornander Gothenburg, June 2025 vii List of Acronyms Below is the list of acronyms that have been used throughout this thesis listed in alphabetical order: CDC Community Day Centre CHC Community Health Center CMS Council of Medical Schemes CT Computed Tomography CUR Health Problem and Current Use of Technology DALYs Disability Adjusted Life Years ED Emergency Department ECO Costs and economic evaluation EFF Clinical Effectiveness ETH Ethical analysis FAST Face, Arm, Speech Test GCS Glasgow Coma Scale HTA Health Technology Assessment IV r-tPA Intravenous Recombinant Tissue Plasminogen Activator LEG Legal aspects LMICs Low- and Middle Income Countries MRI Magnetic Resonance Imaging MWT Microwave Technology NHI National Health Insurance ORG Organisational aspects PMB Prescribed Minimum Benefit SA South Africa SAHPRA South African Health Products Regulatory Authority SAF Safety SATS South African Triage Scale SOC Patients and Social aspects SOP Standard Operating Procedure TBI Traumatic Brain Injury TEC Description and technical characteristics of technology QoL Quality of Life QALYs Quality Adjusted Life Years WC Western Cape ix WC EMS Western Cape Emergency Medical Services WHO World Health Organization WTP Willingness-To-Pay x xii Contents List of Acronyms ix List of Figures xv List of Tables xvii 1 Introduction 1 1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1.1 Stroke . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1.1.1.1 Ischaemic stroke . . . . . . . . . . . . . . . . . . . . 2 1.1.1.2 Haemorrhagic stroke . . . . . . . . . . . . . . . . . . 2 1.1.2 Stroke treatments . . . . . . . . . . . . . . . . . . . . . . . . . 3 1.1.2.1 Thrombolysis . . . . . . . . . . . . . . . . . . . . . . 3 1.1.2.2 Thrombectomy . . . . . . . . . . . . . . . . . . . . . 4 1.1.3 Stroke rehabilitation . . . . . . . . . . . . . . . . . . . . . . . 4 1.1.4 Traumatic brain injury . . . . . . . . . . . . . . . . . . . . . . 4 1.1.5 TBI treatments . . . . . . . . . . . . . . . . . . . . . . . . . . 5 1.1.6 South African Healthcare System . . . . . . . . . . . . . . . . 6 1.1.6.1 Public Healthcare . . . . . . . . . . . . . . . . . . . . 6 1.1.6.2 Private Healthcare . . . . . . . . . . . . . . . . . . . 10 1.1.6.3 Emergency Medical Services . . . . . . . . . . . . . . 10 1.1.6.4 Triage Scales . . . . . . . . . . . . . . . . . . . . . . 11 1.2 Purpose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 1.3 Goals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 1.4 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 2 Methods 15 2.1 Health Technology Assessment . . . . . . . . . . . . . . . . . . . . . . 15 2.1.1 Literature review process . . . . . . . . . . . . . . . . . . . . . 15 2.1.2 Interview study methodology . . . . . . . . . . . . . . . . . . 16 2.2 Health economy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 3 Results 19 3.1 Health problem and current use of technology . . . . . . . . . . . . . 19 3.2 Description and technical characteristics of technology . . . . . . . . 25 3.3 Safety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 3.4 Clinical Effectiveness . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 xiii Contents 3.5 Cost and economic evaluation . . . . . . . . . . . . . . . . . . . . . . 36 3.6 Ethical analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 3.7 Organisational aspects . . . . . . . . . . . . . . . . . . . . . . . . . . 42 3.8 Patients and social aspects . . . . . . . . . . . . . . . . . . . . . . . . 46 3.9 Legal aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 4 Discussion 53 4.1 Clinical study . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 4.2 Suggestion of future implementation sites . . . . . . . . . . . . . . . . 54 4.2.1 Public healthcare . . . . . . . . . . . . . . . . . . . . . . . . . 54 4.2.2 Private healthcare . . . . . . . . . . . . . . . . . . . . . . . . 57 4.3 Health economy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 4.4 Limitations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 4.4.1 Future outlook . . . . . . . . . . . . . . . . . . . . . . . . . . 60 5 Conclusion 61 Bibliography 65 References 65 A Appendix 1 I B Appendix 2 III C Appendix 3 VII xiv List of Figures 1.1 Associated risk factors of a stroke in Africa, based on two separate studies INTERSTROKE Afrika and SIREN [8]. . . . . . . . . . . . . 2 1.2 Hierarchy of public healthcare levels in SA with district at the bottom, followed by regional and tertiary at the top. . . . . . . . . . . . . . . 7 1.3 Overview of the general patient referral pathways between the hospi- tal facilities in the WC. . . . . . . . . . . . . . . . . . . . . . . . . . . 7 1.4 South African Triage Scale (SATS) chart showing priority levels and associated colour codes. Source: Emergency Medicine Society of South Africa (EMSSA), SATS Manual (2012). Shared under a CC BY-NC-SA 3.0 license [54]. No modifications were made. . . . . . . . 12 3.1 Strokefinder™ MD100 sourced from Medfield Diagnostics [78] . . . . 26 4.1 The current patient referral system of stroke and the possible imple- mentation site of the MD100 in the pre-hospital care. . . . . . . . . . 55 4.2 The current patient referral system of stroke and the possible imple- mentation site of the MD100 at the level 1 facility. . . . . . . . . . . . 56 xv List of Figures xvi List of Tables 1.1 Levels of healthcare facilities in the Western Cape . . . . . . . . . . . 8 1.2 Table of all hospitals in the WC region with their respective level, CT availability, and referral distances . . . . . . . . . . . . . . . . . . . . 9 1.3 The different parameters of GCS and its respective score. . . . . . . . 12 2.1 List of interviewees, their roles, relevance, and meeting context . . . . 17 3.1 Overview of the clinical trials involving Medfield Diagnostics Strokefinder MD100 and its current status. . . . . . . . . . . . . . . . . . . . . . . 23 3.2 Sensitivity and specificity of triage tools for stroke. * For acute is- chaemic stroke. ** Dependent on time of stroke onset and localisation of the haemorrhage. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 3.3 Sensitivity and specificity of triage tools for trauma * For GCS score ≤ 12 in prehospital setting . . . . . . . . . . . . . . . . . . . . . . . . . 35 3.4 Number of facilities for each suggested implementation site and the purpose of the MD100 in each instance. . . . . . . . . . . . . . . . . . 37 3.5 Identified locations in the public healthcare system for potential adop- tion of the MD100. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 3.6 Classification risk according to guidelines from SAHPRA [110] . . . . 51 xvii List of Tables xviii 1 Introduction 1.1 Background The background introduces the medical conditions stroke and traumatic brain injury (TBI), along with an overview of current treatment and rehabilitation approaches. It also outlines the structure of the South African (SA) healthcare system in the Western Cape (WC) region, covering both the public and private sectors, as well as emergency medical services (EMS). In addition, the current standard operating procedures (SOPs) for stroke and TBI management in the WC healthcare system are presented. 1.1.1 Stroke Stroke is a major global public health challenge and the second leading cause of death worldwide, accounting for an estimated 6 million deaths annually [1], [2]. While stroke affects populations globally, its burden is disproportionately higher in low- and middle-income countries (LMICs), including SA. Between 1990-2016, the number of stroke cases doubled in LMICs while it during the same period decreased by 42% in high-income countries [2]. With 70% of all stroke-related deaths and 87% of all stroke-related disabilities occurring in LMICs, stroke places as a significant socio-economic burden on society [2], [3]. Stroke is a neurological disorder in which blood flow in the brain is restricted [2]. It is an acute medical condition, as limited blood circulation progressively leads to oxygen deprivation, causing severe brain tissue damage if not treated promptly [4]. Time is a particularly important factor in this context. 1.9 million neurons and 13.8 billion synapses are lost every minute as a consequence of an untreated stroke which dramatically increases the risk of permanent brain damage, disability or in worst case death [5]. The symptoms of a stroke can vary and depends on both the severity and exact location [4]. As the brain is the central organ of the nervous system, the site of the stroke affects both what neurological functions that are impaired as well as to what extent [6]. Common symptoms of a stroke are numbness, weakness, speech and vision impairments as well as dizziness and difficulties with body coordination such as walking [7]. Often, only one side of the body is affected causing a one-sided 1 1. Introduction facial drop or paralysis [7]. Stroke risk factors are associated with both lifestyle and genetic influences, as well as pre-existing medical conditions [7], see Figure 1.1. In Africa, hypertension is the leading contributor of the stroke burden, followed by dyslipidaemia and obesity [8]. The World Health Organization (WHO) has estimated that the world’s highest proportion of people with hypertension lives in sub-Saharan Africa, where SA is located [8], [9]. Additionally, the Human Immunodeficiency Virus (HIV) is another significant risk factor for stroke. Africa bears the highest global burden of HIV, with nearly 70% of the cases occurring in sub-Saharan Africa [8]. Figure 1.1: Associated risk factors of a stroke in Africa, based on two separate studies INTERSTROKE Afrika and SIREN [8]. There are two main categories of stroke, namely; ischaemic and haemorrhagic [2]. Globally, 80-85 % are ischaemic while the remaining 15-20 % are haemorrhagic, However, a study in SA showed a split of 60 % ischaemic and 40 % haemorrhagic, possibly due to the high rates of hypertension [10]. 1.1.1.1 Ischaemic stroke Ischaemic strokes occur when a blood vessel in the brain becomes blocked, commonly referred to as a blood clot [2]. Blood clots are often a result from gradual build- up of atherosclerosis inside the blood vessels, which over time narrows the vascular pathway and limits blood flow. There are two types of ischaemic strokes depending on its origin. If the blockage forms directly at the site of the occlusion, it is referred to as a thrombotic ischaemic stroke. A blood clot can also form in another part of the body and travel through the vascular system to the brain, where it becomes lodged. This type of stroke is called an embolic ischaemic stroke. 1.1.1.2 Haemorrhagic stroke Haemorrhagic strokes are characterized by an intracranial brain bleeding, either caused from a ruptured blood vessel or a leaking blood vessel [2]. Typically, the 2 1. Introduction haemorrhage expand within 3–12 hours of onset, leading to a build-up of blood that increases intracranial pressure and restricts blood flow, resulting in progressive brain damage [11]. This type of stroke is generally associated with high morbidity and mortality due to the pressure exerted by accumulated blood on surrounding brain tissue [12], [13]. The symptoms of a haemorrhagic stroke are similar to those appearing in case of an ischaemic stroke, usually presented as severe headache, speech and vision difficulties as well as one-sided muscle weakness and facial palsy [12]. 1.1.2 Stroke treatments Treatment strategies for ischaemic and haemorrhagic strokes differ significantly. Management of haemorrhagic strokes typically involves supportive care, such as the administration of antihypertensive medications to control blood pressure or al- ternatively prescription of medication to promote clotting to stop the bleeding [11], [14]. In some cases, surgical interventions may also be considered for haemorrhagic strokes. Coil embolization is a minimally invasive procedure where small platinum coils are used to occlude the bleeding vessel by preventing further expansion. The coils are guided thorough the the blood vessels via a catheter to the site where there are deployed in order to minimize the free bleeding in the brain [15]. To re- lieve intracranial pressure caused by the accumulation of blood, another surgical intervention called decompressive craniotomy can be performed. This procedure alleviates the pressure resulting from cerebral edema induced by the bleeding [14]. In contrast, the two treatment procedures for ischaemic strokes are thrombolysis and thrombectomy. The eligibility for either treatment depends on the time from stroke onset, additional inclusion criteria, as well as the healthcare facility’s capacity and access to necessary medical expertise. 1.1.2.1 Thrombolysis Thrombolysis is a treatment that involves administering a drug to dissolve blood clots. This is typically done by injecting tissue plasminogen activator (tPA), a substance that breaks down fibrin which is the protein that holds the clot together [16]–[18]. Administering thrombolysis to patients with ischaemic stroke has been associated with a significant reduction in death and long-term disability, given that it is ad- ministrated within the therapeutic time-window of 4.5 hours from stroke onset [14], [17]. Outside the recommended time-frame, the risk of intracranial haemorrhage may outweigh the benefits as thrombolytic agents are not targeted medications, but has a degrading property of the entire fibrin matrix and not just the intended blood clot. Therefore, thrombolysis is not suitable for all patients, and should be avoided for patients with an active internal bleeding due to the elevated risk of haemorrhagic complications [17]. 3 1. Introduction The greatest benefits of thrombolysis are observed when it is administered as early as possible [17], [18]. Patients who receive thrombolysis within 90 minutes of stroke onset experience nearly twice the benefit compared to those treated later [18]. There- fore, timely identification of stroke is crucial, yet often the reason patients are ex- cluded from receiving thrombolytic therapy. 1.1.2.2 Thrombectomy When obstruction occurs in the proximal cerebral arteries, known as a large vessel occlusion (LVO) which account for 1/3 of all ischaemic strokes [19], thrombolysis is often ineffective due to the extent of the blood clot [20]. Instead, a more effective approach is the mechanical thrombectomy, a procedure that involves the removal of the clot from a blood vessel under image guidance using endovascular devices [21]. The most common techniques for thrombectomy are catheter-based therapies, which utilize either a stent retriever or direct aspiration [14], [21]. Usually, a catheter is inserted via an artery at the groin and from there guided up to the neck and to the site of the blockage in the brain artery. The catheter works as a passage for the stent-retriever to reach the clot and manually remove it in order to restore blood flow. Thrombectomy is compared to thrombolysis a more recent approved treatment for ischaemic stroke and has been proven to be both safe and effective at randomized controlled trials (RTCs), under the circumstances that the procedure is performed correctly [22]. Currently, the surgical procedure is mostly only performed by spe- cialised surgeons such as interventional radiologist or interventional neurosurgeon [12]. The procedure is approved to be performed up to 24 hours after stroke onset [20], [23], although current evidence suggests that performing the procedure as early as possible yields better clinical outcomes [22]. 1.1.3 Stroke rehabilitation While cell death at the stroke core is irreversible, functional recovery is possible through neuroplasticity, which is the brain’s ability to form new neural connections [24]. This process allows healthy brain regions to take over lost functions and is stim- ulated through targeted, task-specific exercises during rehabilitation [25]. Stroke rehabilitation typically includes physical therapy and strength training to restore mobility and aims to minimize the long-term effects of a stroke. Although recovery timelines vary between individuals, early intervention is crucial, as the most signif- icant improvements usually occur within the first three to six months post-stroke [7]. 1.1.4 Traumatic brain injury TBI is major global health concern, accounting for approximately 6 million deaths annually [26]. It is defined as an acquired cerebral dysfunction resulting from an external force that inflicts trauma on the brain, often as a result from mechanical violence or compression injuries. As with stroke, the burden of TBI is disproportion- 4 1. Introduction ality high in LMICs, which also accounts for 90 % of the TBI-related deaths [27]. SA has one of highest TBI rates in the sub-Saharan Africa with intentional injuries almost seven times as high compared to global rates and road traffic accidents twice the global rates. Similar to stroke, TBI is highly time-sensitive and accurate management is critical to prevent secondary brain injury and mitigate further neurological damage [28]. Secondary brain injuries refer to potential changes over time, sometimes even days after the actual point of injury. Secondary brain injury often results in increased intracranial pressure, which causes brain swelling that restricts blood flow and dam- ages the brain tissue. There are two main types of TBI: penetrating and non-penetrating injuries [28]. Penetrating TBIs are most often the result of interpersonal violence, where objects, such as bullets, pierce the skull and interferes with the brain tissue. Non-penetrating TBIs occur when an external force causes the brain to heavily move within the skull, causing a widespread trauma to the brain. Non-penetrating TBIs are commonly caused by falls, road traffic accidents and sports injuries. Given the broad spectrum of different TBIs, symptoms also vary [28]. Common physical symptoms include severe headache, nausea, sudden seizures, pupil dilation and blurred vision. In some cases, clear fluid may drain from the nose or ears, indicating a more severe brain injury. In addition to the listed physical signs, neuro- logical symptoms are similar to those observed in stroke patents, such as weakness and speech difficulties. 1.1.5 TBI treatments TBI treatment typically involves one or more of three main approaches, medication, surgery and rehabilitation, depending on the injury’s severity. Medication plays a key role in managing TBI patients and there are a variety of drugs that may be prescribed based on the patient’s specific needs [29]. Commonly, medications are used to reduce the accumulated fluid in brain tissue, which can otherwise increase intracranial pressure in the skull, putting patient at a significant risk of further damaging healthy brain tissue. In some cases, the free bleeding compresses blood vessels which puts patients at risk of seizures. To prevent this, anticonvulsants can be administered to reduce the likelihood of seizures and thereby minimize further damage. Besides medication, surgical intervention is another important treatment alternative for particularly moderate to severe TBIs [29]. Surgery may be necessary to manually remove severely damaged brain tissue to relieve pressure and prevent further injury on surrounding healthy tissues. In addition to acute medical and surgical treatments, rehabilitation therapies are considered as equally essential to improve long-term patient outcomes [29]. While 5 1. Introduction rehabilitation does not cure the TBI itself, it focuses on restoring lost brain function and improving quality of life (QoL). Recommended therapies aim to improve blood circulation and thereby oxygen delivery to the brain which helps reduce inflammation and supports patient recovery. Similar to stroke rehabilitation, these may include physical therapy, occupational therapy and speech therapy. Physical rehabilitation are specially important to help relieve muscle spasms and contractures which are common post-TBI symptoms. 1.1.6 South African Healthcare System The healthcare in SA is divided into a public and a private healthcare sector. The public healthcare is funded by the government and used by 84% of the SA popu- lation, offering services that are either entirely free or available at subsidized rates [30]. The private healthcare sector is instead build upon those who afford to pay for their healthcare services via medical aid schemes and contributes for the remain- der of 16% of the population. As stated, the majority of the population heavily rely on public healthcare. Still, almost 80% of the healthcare professionals work in the private sector, highlighting the significant resource disparity within the public healthcare system [31]. In May 2024, a major initiative to make healthcare more equitable and accessible became law in SA [32]. This initiative, known as the National Health Insurance (NHI) Act, represents SA’s strategy to achieve universal health coverage through the establishment of a national insurance fund. Similar to other insurance models, all SA’s citizens will routinely contribute to the NHI fund through taxes, and the government will use these funds to purchase healthcare services from both public and private providers. When healthcare services are needed, the NHI fund will cover the costs on behalf of the patient. However, the transformation of SA’s healthcare system has only just begun, and it is expected to take many years to fully integrate the public and private sectors [32]. The ultimate goal of the NHI is to create a more efficient, fair and effective healthcare system that provides all SA citizens with the right to access high-quality affordable healthcare. There is, as of today, not much collaboration between the public and private health- care documented. However, it has been stated that all pre-hospital units, regardless of being private or public, is bound by law to take patients, regardless of possessing a medical aid scheme, to a hospital in case of injury [33]. In the cases where patients can not afford the private EMS, the government covers this cost [34]. 1.1.6.1 Public Healthcare The public healthcare services are divided into district, regional and tertiary hospi- tals as illustrated in Figure 1.2 [35]. Additionally, clinics, community health centres (CHC) and community day clinics (CDC), provides primary healthcare services sup- porting the pyramid-structure as well as emergency medical services (EMS). 6 1. Introduction Figure 1.2: Hierarchy of public healthcare levels in SA with district at the bottom, followed by regional and tertiary at the top. SA’s healthcare system relies on patient referrals between hospital levels, as spe- cialized care requires advanced facilities. However, this transfer process is time- consuming, costly and often inefficient with an estimated 20% of transfers classified as unnecessary [36]. Figure 1.3 illustrates the current SOP of referral pathways in the WC from lower level facilities to the two tertiary facilities Tygerberg hospi- tal and Groote Schuur hospital. The clinics and district hospitals are distributed throughout both rural and urban areas of the province, while the tertiary care fa- cilities are located in the urban center of Cape Town. This centralized healthcare structure often necessitates multiple referrals, and in some cases, patients may be transferred through as many as four different facilities before reaching the special- ized care they require [33]. For time-critical medical conditions such as stroke and TBI, this structure is particularly challenging. Figure 1.3: Overview of the general patient referral pathways between the hospital facilities in the WC. 7 1. Introduction Public healthcare can also be explained by the level of care the facilities offer. The clinics, that include both the CHCs and CDCs, offer primary healthcare services. These include maternal and child health, chronic disease management, HIV care, rehabilitative and therapeutic services. In some cases, limited emergency care [37]. District hospitals offer primary healthcare services on a 24-hour basis provided by generalist nurses, with four years of education and broader knowledge of basic nurs- ing care, as well as clinical nurse practitioners who have advanced education in a specific area of healthcare and provide specialised care within that scope of prac- tice[38], [39].The size of a district hospital varies between 50 to 600 beds. Common services include in-patient care and emergency care. Some district hospitals are also eligible to perform general surgery [37], [39]. If needed, the district hospitals may receive more specialised support from regional hospitals. The regional hospitals, with 200 to 800 beds, are assigned with broader responsibil- ities as clearly defined in national regulations [39]. A regional hospital operates on a 24-hour basis and offers a range of healthcare services, including general surgery, internal medicine, as well as trauma and emergency care. Some extended services may be offered at regional hospitals but are typically limited to within provincial boundaries. The regional hospitals receive referrals from several district hospitals in the region and can in turn get support from tertiary hospitals. The highest level of healthcare are the tertiary hospitals which offer specialized medical services and intensive care under the supervision of specialists [39]. Ter- tiary hospitals receive referrals from regional hospitals, and these referrals are not restricted to provincial boundaries. In WC there are in total two tertiary hospitals, Tygerberg hospital and Groote Schuur hospital. Both hospitals serve as university hospitals for Stellenbosch University and University of Cape Town respectively. As of today, Tygerberg and Groote Schuur are the only two public hospitals in the WC with the resources required to perform thrombolysis and thrombectomy as well as surgical intervention for treatment of TBI [34], [40]. Table 1.1 groups the different level of hospitals dependent on their accessibility and operation of CT-capabilities. Table 1.1: Levels of healthcare facilities in the Western Cape Level Facility Description Level 1 Clinics and district hospitals without CT capability Level 2 District and regional hospitals with CT capability (typ- ically operating 8a.m. to 4p.m. [34]) Level 3 Tertiary hospitals with CT capability 24/7 There are in total 32 district hospitals, 5 regional hospitals and 2 tertiary hospitals in the WC. Out of these, only 10 facilities are equipped with a CT [34]. Table 1.2 provides an overview of the hospitals equipped with CT scanners, patient referrals and the distances involved. 8 1. Introduction Table 1.2: Table of all hospitals in the WC region with their respective level, CT availability, and referral distances Hospital Level CT Referral [distance] Referral [distance] Tygerberg Tertiary Yes - - Worcester Regional Yes - Tygerberg [96 km] Paarl Regional Yes - Tygerberg [47 km] Karl Bremer District Yes - Tygerberg [4.6 km] Helderberg District No Tygerberg [41 km] - Khayelitsha District Yes - Tygerberg [27 km] Eerste River District No Tygerberg [22 km] - Swellendam District No Worcester [111 km] - Otto du Plessis District No Worcester [151 km] - Hermanus District No Worcester [119 km] - Caledon District No Worcester [90 km] - Montagu District No Worcester [73 km] - Robertson District No Worcester [47 km] - Ceres District No Worcester [55 km] - Vredendal District No Paarl [270 km] - Clanwiliam District No Paarl [197 km] - Citrusdal District No Paarl [144 km] - Radi Kotze District No Paarl [109 km] - Vredenburg District No Paarl [143 km] - LAPA Munnik District No Paarl [88 km] - Swartland District No Paarl [45 km] - Stellenbosch District No Paarl [35 km] - Groote Schuur Tertiary Yes - - George Regional Yes - Groote Schuur [429 km] New Somerset Regional Yes - Groote Schuur [8.8 km] Mowbray Maternity Regional No Groote Schuur [2.6 km] - Victoria Wynberg District Yes - Groote Schuur [10.7 km] Mitchell’s Plain District Yes - Groote Schuur [22.5 km] Beaufort West District No George [235 km] - Murraysburg District No George [395 km] - Prince Albert District No George [167 km] - Laingsburg District No George [228 km] - Alan Blyth District No George [158 km] - Riversdale District No George [134 km] - Mossel Bay District No George [53 km] - Knysna District No George [63 km] - Uniondale District No George [110 km] - False Bay District No Victoria Wynberg [20 km] - Wesfleur District No New Somerset [55 km] - 9 1. Introduction 1.1.6.2 Private Healthcare Private healthcare is a voluntary service in SA, funded by medical aid schemes, which are commonly distributed with employment or can be bought privately [41]. There are approximately 8.9 million people in SA registered at one of 71 medical aid schemes across the country [41], [42]. There are multiple private hospital groups operating in the WC region of SA, where Mediclinic Southern Africa, Life Healthcare and Netcare are the leading ones [43], [44], [45]. Compared to public healthcare that operates on a provincial basis, private healthcare are offered via an entire hospital group that extends beyond provincial boundaries [46]. Members pay a monthly fee to their chosen medical aid scheme, which grants them access to private healthcare services in SA. The extent of coverage for these services depends on the specific medical aid plans [41]. All medical aid schemes are regulated by the national Council for Medical Schemes (CMS) to ensure compliance with Prescribed Minimum Benefits (PMBs) for their members, regardless of their medical aid plan [41]. 1.1.6.3 Emergency Medical Services Public EMS are a provincial function governed by the Provincial Departments of Health (PDoH), funded via the government [47]. In the WC province this is the Western Cape Emergency Medical Services (WC EMS). For private EMS, there are both privately owned companies and hospital groups that offer EMS, amongst those, ER24 is a part of the private hospital group Mediclinic Southern Africa [48]. ER24 units are staffed with paramedics and equipped with medication and medical equipment for advanced life support [48]. The collaboration between public and private healthcare in terms of pre-hospital services and units are somewhat limited [33]. The WCG EMS operate over 250 ambulances across the WC province of SA [49]. In SA specifically, the national recommendation is to have at least one ambulance per 10,000 citizens [50]. In the context of the WC’s population with 7.2 million people, this would correspond to approximately 720 pre-hospital units [51]. Given that around 84 % of the WC population relies on public healthcare services, this gives a shortage of 350 units across the WC province. It has also been reported that many pre-hospital units are out of service [50]. A study from 2024 found that only 130 units operated in the WC, which represents to 52 % of the resources that should be functional [51], which highlights a significant shortfall. For private EMS operations in the WC, there is no information available on how many units that are available. The WC government have reported that each public pre-hospital unit in WC is equipped with medication and medical equipment as well as two medical practition- ers. It is stated that one or both of the practitioners in the EMS can perform basic life support (BLS), intermediate life support (ILS) and advanced life support (ALS) [49]. However, in a study from 2024 it was shown that the SA pre-hospital care providers knowledge about both BLS and ILS were lacking, especially the knowl- edge about Glasgow Coma Scale (GCS) which is a specifically critical evaluation 10 1. Introduction system used for triaging of stroke and TBI patients [50]. Besides pre-hospital care, the WCG EMS also perform inter-facility transfers of patients between different in- stances of the public healthcare in cases where the patient needs care that extend the capability of the current instance [49]. In an annual report from 2017/2018 it was concluded that approximately 1/3 of the WC EMS activity was dedicated to inter-facility transfers [52]. This further restricts the time allocated to handle primary medical responses, such as stroke and TBI. 1.1.6.4 Triage Scales A definite diagnosis of stroke and TBI is confirmed by a neuroimaging service, either computed tomography (CT) or magnetic resonance imaging (MRI) [1]. These imaging methods not only confirm the presence of a stroke or TBI but also provide critical information about the exact location and the extent to which surrounding brain tissue has been affected. The initial assessment of patients prior access of a neuroimaging service, is usually conducted using standardized triage scales, both in the pre-hospital setting and at the hospitals. Worldwide, commonly used scales for assessment of stroke in the pre-hospital settings are the 3-item stroke scale, the Austrian Prehospital Stroke Scale, the Cincinnati Prehospital Stroke Scale, the Los Angeles Stroke Screen and the Rapid Arterial Occlusion Evaluation scale [53]. All of the mentioned stroke assessment scales are highly based on clinicians interpretation of a patient’s appear- ance and it has been reported that in 30 % of the pre-hospital cases, the stroke assessment scales are insufficient to adequately recognize stroke [53]. The most complete and detailed scoring system in-hospital settings is the National Institutes of Health Stroke Scale (NIHSS). This is a commonly used stroke scale which estimates the stroke severity based on 15 evaluating statements [53]. The assessment does not require neurology expertise, making it suitable as an initial evaluation tool at all hospital levels and contributing to its widespread use. To adapt the scale for use in the time-sensitive emergency setting, two modified versions have been developed, namely the modified NIHSS and the shortened NIHSS-EMS, which both are frequently used in the pre-hospital setting. The specific stroke triage scales in SA are the South African Triage Scale (SATS), the Glasgow Coma Scale (GCS), and the Face-Arm-Speech-Time (FAST) test [54] [34]. Similarly, the initial assessment of a TBI patient is based on SATS and GCS [34]. For performance metrics of the individual assessment tools used, see tables 3.2 and 3.3. South African Triage Scale Triaging patients in SA pre-hospital and emergency centres (ECs) are done using SATS [55], [54]. The priority coding is divided into five different levels and managed accordingly; see Figure 1.4. 11 1. Introduction Figure 1.4: South African Triage Scale (SATS) chart showing priority levels and associated colour codes. Source: Emergency Medicine Society of South Africa (EMSSA), SATS Manual (2012). Shared under a CC BY-NC-SA 3.0 license [54]. No modifications were made. The SATS follows a five step approach where the healthcare personnel begins with looking for emergency signs followed by looking for very urgent or urgent signs. Depending on the level of urgency, the next step is to measure the vital signs and calculate the Triage Early Warning Score. Finally , a triage priority level is assigned according to above [54]. The published manual also states that if the patient has any of the following emergency signs, a TEWS score does not need to be calculated in order to prioritize the patient as red; obstructed airway (not breathing), ongoing seizures, facial burns/inhalation, hypoglycaemia (glucose less than 3mmol/L) or cardiac arrest [54]. Glasgow Coma Scale The GCS is particularly valuable for assessing patients with haemorrhagic stroke or TBI, and it is also applicable for comatose patients [53]. It is commonly used to perform neurological assessments at the hospital by evaluating the three key compo- nents: motor responsiveness, verbal performance and eye opening. Each component is assigned a specific score, which is then summed to yield a total score, as shown in Table 1.3 [56], [57]. Table 1.3: The different parameters of GCS and its respective score. Motor function Verbal Response Eye response Obey commands (6) Oriented (5) Spontaneous reaction (4) Localising to pain (5) Confused (4) React to sound (3) Normal flexion to pain (4) Words (3) React to pressure/pain (2) Abnormal flexion to pain (3) Sounds (2) No eye opening (1) Extension to pain (2) No verbal response (1) - None (1) - - 12 1. Introduction The interpretation of the total scoring is divided into three different severity levels: a score between 13-15 is classified as a minor head injury, 9-12 is classified as a moderate head injury and ≤ 8 is classified as a severe head injury or coma [56]. The GCS is also incorporated into three more detailed neurological scoring systems; the World Federation of Neurological Surgeons (WFNS) scale, the ICH score and the Full Outline of UnResponsiveness (FOUR) score [53]. These scales build upon the GCS to provide a more comprehensive assessment tailored to specific clinical scenarios. Face, Arm, Speech, Time - Test The FAST assessment is an initial screening method used in the pre-hospital set- ting where the clinician evaluates the physical appearance of face drooping (F), arm weakness (A) and speech (S) difficulties [58]. The FAST test is designed to facilitate the early recognition of stroke symptoms and is widely applicable, both within healthcare and by the general public, due to its straightforward and easy-to- remember process. There are several modified versions of the FAST assessment [53]. The G-FAST test takes gaze deviation (G) into consideration while the BE-FAST version incorporates assessments of both balance (B) and eye (E) symptoms. Addi- tionally, the FAST-ED scale includes eye deviation (ED) and signs of anosognosia or neglect, which is specially common for LVOs[53]. 1.2 Purpose Healthcare in SA faces significant challenges. The public sector, which serves the vast majority of the population, is severely under-resourced and overburdened, strug- gling to meet the growing demand for services. Shortages of financial and human re- sources and a high burden of disease place enormous pressure on healthcare providers and facilities. Additionally, the current referral-based healthcare system introduces substantial delays in care, which are especially critical for time-sensitive conditions such as stroke and TBI. Consequently, timely diagnosis, effective treatment, and equitable access to care remain difficult to achieve, resulting in poorer patient outcomes and increased strain on healthcare services. 1.3 Goals This project aims to evaluate the suitability of the Swedish medical innovation, Strokefinder MD100, in the WC public healthcare system, with a focus on its po- tential to shorten time-to-treatment for stroke and TBI patients. By examining current care pathways, the project will identify critical integration points where the MD100 could strengthen clinical decision-making and patient triage. Guided by the EUnetHTA Core Model, the assessment will ensure a structured and multidimen- sional evaluation of the device’s clinical relevance and health system compatibility. 13 1. Introduction Based on the findings, the project will recommend suitable implementation sites to reach more timely and equitable access to care. 1.4 Limitations To ensure the feasibility of this project, certain limitations were necessary. Firstly, the assessment is geographically restricted to the WC province of SA. The scope includes both public and private healthcare sectors; however, the primary focus is on the public sector due to limited access and communication with stakeholders within the private sector. The assessment addresses both stroke and TBI. However, given that the MD100 is only CE-certified for stroke, this is our primary focus when performing HTA, but extends to TBI when appropriate. Due to the limited time-frame of the project and the current development status of the Strokefinder MD100, not all questions from the EUnetHTA Core Model could be fully addressed, see appendix B for specification. The helicopter emergency medical services are primarily responsible for longer transfers of acute patients to higher instances of care. Given the limited information about the helicopter emergency medical services in the WC, SA, this aspect of the healthcare will not be included in the discussions for the final recommendation of implementation sites. The discussions will also be based on that the standard operating procedures for patient referral pathways are followed, even though it has been stated that this is not always the case [40], [59]. 14 2 Methods This study employs a combination of an extensive literature review combined with qualitative interviews with relevant stakeholders. The following section outlines the specific methodologies respectively by detailing the process of sourcing and analysing the literature as well as describing the execution and analysis of the performed expert interviews. 2.1 Health Technology Assessment Health technology assessment (HTA) is a systematic approach that evaluates health related technologies by applying a multidisciplinary approach that includes both di- rect and indirect aspects [60]. The process shall be transparent and accountable and by providing an evidence-based review of the technology, the HTA aims to support and serve stakeholders in the decision-making process of implementation of health technology [60]. The HTA aims to evaluate properties, effects and impacts of med- ical technologies [61] and is defined by WHO as follows: ”It is a multidisciplinary process to evaluate the social, economic, organizational and ethical issues of a health intervention or health technology. The main purpose of conducting an assessment is to inform a policy decision making.” [61]. The European Network for Health Technology Assessment (EUnetHTA) has cre- ated a structured framework, EUnetHTA Core Model version 3.0 from 2016 [62]. This project is structured as per the EUnetHTA core model, focusing on properly answering each respective question via a comprehensive literature review combined with expert interviews. 2.1.1 Literature review process The literature review focused on topics related to stroke and TBI, both in terms of epidemiology, prevalence, morbidity and mortality, therapeutics and rehabilitation in the specific area WC, SA. Other relevant areas where the Strokefinder MD100 and information regarding microwave technology (MWT). The used databases included Pubmed, Google Scholar and IEEE Xplore. Additional material was supplied by 15 2. Methods professor Emeritus Mikael Persson, one of the founders of Medfield Diagnostics, as well as previous work by PhD student Tinashe A Chikunichawa at Stellenbosch University and M.Sc student Jean Paul Kulumba at Stellenbosch University. 2.1.2 Interview study methodology To gain insights into the specific care of WC, SA, qualitative interviews were carried out through the 87 days in SA. In total, 12 interviews were held with participants selected based on their current or previous professional roles and relevant experience. The interviews were semi-structured to allow for open-ended responses and in-depth conversations, encouraging participants to share their perspectives on the topics discussed rather than providing short or single-sentence answers. The interviews were conducted both in person and online, as outlined in Table 2.1, to accommodate the preferences of each interviewee. The duration of the inter- views ranged from 45 to 150 minutes. Prior to each interview, all participants were informed of the study’s purpose and gave their consent to participate voluntarily. Continuos notes were taken during the discussions. The data gathered from the interviews were transcribed and summarized to capture the key conclusions. These summaries were then sent to the respective interviewees for approval and possible correction. This process ensured that the information was accurately represented, minimizing misunderstandings or subjective interpretations, and instead focusing on preserving the integrity of the insights provided. To form a comprehensive understanding of the healthcare system in relation to stroke and TBI care, and the implementation of medical technologies, interview participants were selected based on their expertise across multiple relevant domains, as outlined in Table 2.1. The diverse professional backgrounds were intentionally chosen to capture the perspectives of all stakeholders including real-world consider- ations such as regulatory frameworks, economic feasibility and market integration. This approach ensured the inclusion of both clinical and non-clinical viewpoints, providing a contextually grounded foundation for the study. 16 2. Methods Table 2.1: List of interviewees, their roles, relevance, and meeting context Name, Location & Date Role Relevance Alan Bryer Groote Schuur Hospital, Cape Town 24.03.2025 Former Head of the Division of Neurology and Stroke Unit at Groote Schuur Hospital. To understand current stroke management practices and SOPs of stroke at one of the tertiary facilities in WC, SA. Daniel Youkee Groote Schuur Hospital, Cape Town 24.03.2025 Postdoctoral researcher with the Stroke group at the Neuroscience Institute and Emergency Medicine doctor at Groote Schuur Hospital, Cape Town. Provided stroke insights in contexts without using CT-scanners. Evan Herbst Online 08.04.2025 Director Public Health and Commercial Excellence with 25 years of experience within the South African healthcare market. To provide perspective on medical technology device implementation and market access in SA. Hendrick J Lategan Tygerberg Hospital, Cape Town 07.03.2025 Operational Head of the Trauma Unit at Tygerberg Hospital and Division of Emergency Medicine. To understand current trauma triage processes at the tertiary facilities and gain insights of current patient flow pathways in WC, SA. John Paul Kulumba Online 27.02.2025 Industrial Engineer, Macrologistics Data Analyst. To explore integration possibilities and challenges of medical technology within the private healthcare sector. Kathleen Bateman Groote Schuur Hospital, Cape Town 24.03.2025 Head of the Division of Neurology and Stroke Unit at Groote Schuur Hospital. To understand current stroke management practices and SOPs of stroke at one of the tertiary facilities in WC, SA. Mark Brand Online 01.04.2025 Managing Director at Brandtech Health Technology. To gain industry perspective on implementation of medical technology in SA and regulatory aspects for the health system entry. Mladen Poluta Online 21.04.2025 Health Technology Director at Southern Right HTM Consulting. To understand procurement procedures and introduction of medical technology for both private and public healthcare in SA. Naseef Abdullah Online 05.05.2025 Head of unit of the ambulance care, Western Cape Government. To provide insights on current ambulance care in the WC and possible integration of the MD100 within this setting. Sa’ad Lahri Tygerberg Hospital, Cape Town 30.04.2025 Head of division of Emergency Medicine at Tygerberg Hospital. To understand current management of stroke and discuss clinical applicability of MD100 in the public healthcare of the WC region. Sudesh Sivarasu University of Cape Town, Cape Town 31.03.2025 Leads the university’s MedTech lab and Senior Lecturer in Biomedical Engineering. To discuss local medical technology development processes with a focus on the regulatory context of SA. Tinashe Chikunichawa Stellenbosch University, Stellenbosch 12.03.2025 PhD Student in Industrial Engineering, Stellenbosch University. To discuss previous conclusions about the MD100 amongst healthcare professionals and discussing cost-effectiveness of the device in different contexts. 2.2 Health economy The health economy aspect is a critical factor in the evaluation of medical technology implementation. In order to analyse this factor in the SA context, we have conducted health economic calculations. The calculation of the Quality Adjusted Life Year (QALY) are derived from the amount of disability free days. In the equation, x is the number of disability free days and the utility value is a score between 0 (death) to 1 (perfect health). 17 2. Methods QALY = x 365 · utility value (2.1) The total QALY gained can further be calculated, where n is the amount of patient the device is used on annually. Total QALY = QALY · lifetime of device · n (2.2) The incremental cost-effective ratio (ICER) is derived by the following equation where C is the expected increase in costs, B is the benefit with respect to QALY and λ is the threshold value, i.e., the maximum per QALY which is the Willingness To Pay (WTP). ICER tells you how much extra cost is required to gain one additional unit of health benefit. ICER = ∆C/∆B < λ (2.3) Health economic is proved beneficial when the Incremental Net Monetary Benefit (INMB) is positive as explained in the following equation. INMB converts both costs and effects into a single monetary value and tells you the net value of switching to a new intervention, given a specific WTP threshold. if: 0 < λ · ∆B − ∆C then: INMB > 0 (2.4) INMB = NMBnow − NMBcomparator (2.5) 18 3 Results This section presents the findings from our research, answering the questions from the EUnetHTA Core Model version 3.0 from 2016 for the Medfield Diagnostics device Strokefinder MD100 in SA. The chosen issues are directly taken from the EUnetHTA and specified in the table for each of the domains. The excluded issues, their respective ID, domain and the reason for exclusion is specified in appendix B. 3.1 Health problem and current use of technology Understanding the health context in which the MD100 is intended to operate is essential for assessing its potential value. This domain provides an overview of the current management of stroke and TBI in the WC, SA. It also discusses the burden of these conditions, the role the MD100 is intended to play in their management and the device’s current regulatory status. Target population The following issues and their corresponding IDs, from the EUnetHTA Core Model, aims to address the target population. ID Issue A0007 What is the target population in this assessment? A0023 How many people belong to this target population? A0002 What is the health condition falling under the scope of this as- sessment? A0003 What are the known risk factors for the health condition? A0004 What is the natural course of the health condition? A0005 What are the symptoms or health condition for the patient? Stroke is a significant public health concern in SA with approximately 75 000 identi- fied cases a year [63]. It is the second leading cause of death in SA, and contributing to an estimated 25,000 deaths annually [1], [63]. The natural course of untreated stroke can result in significant neurological impairments or in some cases death, risks that increases if diagnosis and intervention are delayed [6]. Symptoms and associ- ated risk factors are specified in section 1.1.1. The largest risk factor, hypertension, 19 3. Results is a known consequence of both alcohol and substance abuse, which both are con- siderable public health challenges in SA [64]. The other associated risk factor, HIV, has a 20% prevalence in SA with close to 8 million diagnosed cases [1], [65]. SA has one of the highest rates of TBI globally, with an estimated 89,000 cases reported annually, likely excluding a substantial number of unreported cases [66]. The significant burden of trauma is a national concern in SA, with rural areas being particularly affected [67]. The majority of trauma patients in SA are due to interpersonal violence [33], with national rates reaching up to seven times the global average [68]. Studies link the high prevalence of trauma to factors such as unemployment, alcohol abuse, and gang-related activity [69]. These issues are especially prominent in poorer neighbourhoods and townships, where elevated crime and violence rates contribute to the socio-economic burden. TBI often results in long-term neurological and physiological impairments, placing additional strain on individuals, families, and the healthcare system [68]. Target condition The next topic in this domain is the target condition, where the following issues will be addressed. ID Issue A0006 What are the consequences of the health condition for the society? A0009 What aspects of the consequences/burden of disease are targeted by the technology? Permanent disabilities resulting from a stroke or TBI can significantly reduce a pa- tient’s quality of life (QoL) and impose a considerable burden on healthcare systems due to the long-term demand for medical care and rehabilitation resources [26]. In fact, stroke is among the top ten leading causes of disability in SA, accounting for approximately 95,000 years lived with disability nationwide [70]. The post-stroke outcomes are influenced by multiple factors, including the type, severity, and loca- tion of the stroke, as well as the timing and accessibility of medical intervention. Survivors frequently require multidisciplinary rehabilitation to regain physiological functions, still around 40% of surviving stroke-patients have to live with other severe permanent impairments [7]. Stroke is also associated with substantial direct healthcare costs. A previous study estimated these costs to include expenditures related to hospitalisation, medication, physiotherapy, speech therapy and outpatient care services [71]. Over a five-year period, the total direct cost was estimated at R7.3 trillion. The mean medication cost per patient during this time was approximately R65,700. These findings illus- trate the substantial economic burden that stroke care imposes on the healthcare system and society at large. The MD100 has the potential to reduce time-to-treatment by enabling earlier iden- tification of suspected stroke cases. In line with the "time-is-brain" principle [72], 20 3. Results earlier diagnosis and initiation of appropriate treatment are associated with im- proved clinical outcomes and a higher likelihood of functional recovery. Enhanced recovery increases the potential for patients to regain independence and return to work, thereby contributing to gains in quality-adjusted life years (QALYs) [72]. From a societal perspective, this may lead to reduced long-term rehabilitation needs and associated healthcare costs, delivering both clinical and economic benefits. Linked evidence from a study performed showed no difference in cumulative healthcare costs when earlier thrombectomy where performed on acute ischeamic stroke patients [72]. This, since the total cost for healthcare were spread out across the longer life-time of the patient when earlier treatment were given [72]. Current management of the condition Next, current management of the condition stroke and TBI are adressed by answer- ing the following issues from the topic. ID Issue A0018 What are the other typical or common alternatives to the current technology? A0024 How is the health condition currently diagnosed according to pub- lished guidelines and in practice? A0025 How is the health condition currently managed according to pub- lished guidelines and in practice? Currently, no technology exists to assist in stroke triage outside the hospital setting; clinicians must rely solely on their clinical judgment guided by standard assessment scales as presented in Section 1.1.6.4. The MD100 is not intended to replace these scales neither hospital-based neuroimaging (CT or MRI), but rather to complement these in early decision-making. Groote Schuur hospital has an on-call stroke physician who assists in prioritizing potential stroke patients for the CT-scanner [40], [59]. Stroke patients at Groote Schuur hospital can either be admitted to the dedicated stroke unit or to the emer- gency department (ED) depending on capacity [40], [59]. Groote Schuur Hospital administers thrombolysis within the 4.5-hour window and thrombectomy within the 7-hour window, provided that patients meet the eligibility criteria. In total 35-40 thrombolysis per year are administered at Groote Schuur and an unknown num- ber of thrombectomies, primarily due to the fact that the patients reach the stroke unit outside of the time window for treatment. At the other public tertiary hospital Tygerberg, 120 thrombolysis are administered and 48 thrombectomies are performed annually [34], also a limited number as a consequence of the strict time-window [34], [40], [59]. The Standard operating procedures (SOPs) for trauma cases in the public WC is for the EMS to take the patient to the nearest ED to perform an initial stabilisation before further referral. Those who then require specialised care will after the initial 21 3. Results stabilisation be referred to a tertiary facility for further interventions [33]. However, SOPs are not always strictly followed, resulting in the transfer of non-stabilized pa- tients to tertiary facilities which places an additional strain on already overburdened tertiary hospitals [33]. Utilisation The following section addresses the aspect of utilisation by examining the issues outlined. ID Issue A0001 For which health conditions and populations, and for what pur- poses is the technology used? A0011 How much are the technologies utilised? A0012 What kind of variations in use are there across countries/region- s/settings? G0009 Who decides which people are eligible for the technology and on what basis? F0001 Is the technology a new, innovative mode of care, an add-on to or modification of a standard mode of care or replacement of a standard mode of care? The MD100 is a novel and innovative technology designed to serve as an add-on to existing modes of care and treatment protocols with intended use as a triage support tool for suspected stroke and TBI. The technology does not, at current state, possess the ability to differentiate between the two types of stroke, ischaemic or haemorrhagic. The studies involving the MD100 are listed in the Table 3.1. 22 3. Results Table 3.1: Overview of the clinical trials involving Medfield Diagnostics Strokefinder MD100 and its current status. Clinical Trial Trial ID Country Status Study to Evaluate Performance, Usability, Safety of Microwave Technology When Collecting Data From Patients With Stroke NCT02266459 Sweden Completed 2015 Detecting Chronic Subdural Hematoma With Microwave Technology NCT02282228 Sweden Completed 2016 Clinical Evaluation of a Microwave-Based Device for Detection of Traumatic Intracranial Hemorrhage NCT02291261 Sweden Completed 2017 Detecting Traumatic Intracranial Hemorrhage With Microwave Technology NCT02728908 Sweden Completed 2019 Detecting Stroke at the Emergency Department by a Point of Care Device: A Multicenter Feasibility Study N/A Greece Completed 2024 NSW Ambulance First in the World to Trial New Stroke Care Technology N/A Australia Completed 2024 Microwave Imaging in NeuroTrauma NCT05960279 England Unknown Detecting Traumatic Intracranial Hemorrhage With Microwaves and Biomarkers NCT04666766 Sweden Unknown Evaluating Use of Microwave Technology to Differentiate Hemorrhagic Stroke From Infarction in the Acute Phase NCT02490306 Sweden Unknown Mobile Microwave-based Diagnosis and Monitoring of Stroke (MODS) NCT04257149 Norway Unknown In the public sector of WC, the eligibility of medical technology is determined by the National Department of Health (NDoH) [73]. The public hospitals are simply requesting the technology they need via a basic list. For the academic institutions, such as Tygerberg and Groote Schuur, the list is forwarded to their respective com- mittee inside the hospital unit, whereas other public hospitals in the WC submit the list to a provincial committee. This procedure is done on provincial level as each of 23 3. Results the provinces in SA operates independently, meaning that nation-wide adoption of medical technology requires approval and integration into the healthcare system of each separate province [74]. The provincial committee consists of experts with insights and great knowledge about the respective facilities that they are overseeing [73]. The list follows a three- phase assessment. The first phase involves an initial screening, where only technolo- gies that meet basic approval criteria proceed to the second phase of the evaluation cycle. The exact criteria for this are still uncertain. The second phase involves a comprehensive evaluation and review of technical specifications and clinical rel- evance. Finally, the third phase consists of an economic assessment that analyses cost-effectiveness and financial feasibility, which often is the reason why requested technology are declined. Once medical technology has passed the three-phase assessment and thereby been approved for funding, its implementation within a hospital setting is typically over- seen by the head of the unit, also the person writing the request. The head of unit is also responsible for updating and adapting the SOPs to incorporate the new technol- ogy into clinical workflows [33]. It has been emphasized that modifying established procedures among healthcare professionals requires a high level of trust within the clinical team. The hospital’s hierarchical structure, which is grounded in mutual re- spect and professional trust, plays a key role in facilitating a smooth transition. As a result, when changes are introduced to a department these are generally accepted and well-integrated by the healthcare staff [33]. In the private healthcare sector, decisions regarding the eligibility and funding of medical technologies are made by individual medical schemes and their respective administrators. Supporting this process are the Health Policy Units (HPUs) and Clinical Policy Units (CPUs), which are responsible for conducting the HTAs [46]. HTA play a crucial role in the private healthcare sector as it is used when evaluating the clinical effectiveness and cost-efficiency of new technologies, helping to determine whether a technology should be included in the medical aid scheme’s list of covered services. The process operates under the oversight of the CMS, which is a statutory regulatory body that supervises medical schemes to ensure that the CMS ensures that schemes comply with relevant legislation, particularly the Medical Schemes Act, and that they uphold members’ rights and access to appropriate healthcare services [75]. Regulatory status The last topic in this domain addresses the regulatory status and adressess the following issues. ID Issue A0020 For which indications has the technology received marketing au- thorisation or CE-marking? A0022 Who manufactures the product? 24 3. Results In June 2022, the MD100 received CE certification under the Medical Device Regu- lation (EU) 2017/745 as a Class IIa medical device for the detection of stroke. Ac- cording to the manufacturer, the device is intended to function as a decision-support tool to assist in the clinical evaluation and triage of suspected intracranial injuries in acute care settings. The MD100 is developed and manufactured by Medfield Di- agnostics AB, a company founded and headquartered in Gothenburg, Sweden. As of June 2024, the company has declared bankruptcy and is currently undergoing reconstruction proceedings. 3.2 Description and technical characteristics of technology This HTA domain is centred on the MD100 device in the SA context. To fully understand its potential value, it is essential to first understand the technology itself. This domain provides a brief overview of the device’s features and training required to support its effective use in a clinical setting. Features of technology The first topic in this domain examines the specific features of the technology, an- swering the following issues. ID Issue B0001 What is the technology and the comparator(s)? B0002 What is the claimed benefit of the technology in relation to the comparators? B0003 What is the phase of development and implementation of the technology and the comparator(s) B0004 Who determines the technology and the comparators and in what context and level of care are they provided? B0018 Are reference values or cut-off points clearly established? The MD100, Figure 3.1, utilizes microwave technology (MWT) combined with ar- tificial intelligence (AI) to assist in the triage of potential stroke patients. MWT have been a developing area of research for almost three decades [76]. MWT is a low-cost and non-invasive method, free from ionizing radiation [77]. 25 3. Results Figure 3.1: Strokefinder™ MD100 sourced from Medfield Diagnostics [78] MWT builds upon the existence of different dielectric contrast in human tissues [76]. The difference in dielectric properties of healthy brain tissues and non-healthy brain tissues, including blood and edemas provides the basis of the MWT [77]. The relative permittivity, also known as the dielectric constant, is a material parameter that is important for the wave propagation through the tissue [79]. In the case of a haemorrhagic stroke, the accumulation of blood can be detected due to the difference in relative permittivity between blood and brain tissues [77]. In the case of an ischaemic stroke, the detection relies on that the restricted blood flow to brain tissue causes oxygen deprivation which changes in the dielectric properties of the affected areas through edema formation. The MD100 is portable and designed to fit around the patient’s head, enclosing only the outer portion of the skull without covering the face, and used when the patient lies flat on the back [78]. It operates via eight applicators where low-power electromagnetic waves within a specific frequency range are transmitted through the patient’s head [77], [78]. For a complete measurement, each applicator acts both as a transmitter and a receiver. The applicators pressure around the skull are manually adjusted with a handle for a correct position on the head. Additionally, the device incorporates a sensor designed to reduce variability in head positioning by ensuring that consistent pressure is applied, regardless of individual head size. The scanning process takes two minutes, where AI analyses the backscattered, reflected and transmitted microwaves to detect possible abnormalities that signalizes a stroke. As a result, "stroke" or "no stroke" is displayed on the connecting touch screen tablet which communicates with the MD100 via Bluetooth [78]. In contrast to current triage assessment scales, which rely on clinical observations and subjective evaluation of the patient, the MD100 provides results based on an AI-driven analysis of objectively collected signal data from the MWT. However, as the device requires direct contact with the skull, it is not applicable in all scenarios such as in cases when the skull is deformed or compromised due to penetrating trauma [78]. 26 3. Results Investments and tools required to use the technology The next topic address issues in regards to investments and tools required to use the technology. ID Issue B0007 What material investments are needed to use the technology? B0008 What kind of special premises are needed to use the technology and the comparator(s)? B0009 What equipment and supplies are needed to use the technology and the comparator? B0010 What kind of data/records and/or registry is needed to monitor the use of the technology and the comparator? The full system comprises the MD100 device itself, an associated touchscreen tablet and the backpack in which the entire system is safely stored and easily transported. As the system is powered by a rechargeable battery, it requires access to electricity for charging prior to use. The entire system weighs 6 kg, and with complementary single-use hygiene covers required for each patient, it operates independently and only requiring routine maintenance of the system. Given that the system utilizes AI-based algorithms, periodic maintenance in the form of software updates may be necessary to ensure optimal performance and reliability of the system. The current triage assessment scales as presented in Section 1.1.6.4, does not require any supplies or equipment. Training and information needed to use the technology The last topic in this domain addresses issues regarding training and information needed to use the technology. ID Issue B0012 What kind of requirements in terms of qualification and quality assurance processes are needed for the use or maintenance of the technology? B0013 What kinds of skills and training characteristics and information are needed for the personnel/caregivers using this technology? B0015 What information about the technology should be provided out- side the target group and to the general public? Ensuring the quality, safety and effectiveness of a medical technology in accor- dance with SA’s regulatory requirements is a shared responsibility between hospitals, healthcare professionals and manufacturers [73]. One important quality assurance mechanism is to gain acceptance within the local context, meaning that new medical devices must be validated within SA and not solely rely on international performance data. While this serves as a critical safeguard to ensure suitability and reliability in 27 3. Results local contexts, it also adds an additional layer of complexity when introducing new technologies to the SA market. Integrating a device is not merely a matter of distri- bution, strict protocols must be followed to ensure it is used as intended and delivers on its promised performance. This process includes post-market surveillance, which is overseen by the South African Health Products Regulatory Authority (SAHPRA). The MD100 is intended for use by trained healthcare professionals, but its intu- itive design and automated signal analysis significantly reduce the operational skill requirements. Unlike current triaging scales, which require clinical expertise to inter- pret and draw conclusions based on the patients behaviour, the output when using MD100 is straightforward, "stroke" or "no stroke". As a result, no advanced skills or experience are necessary to operate the MD100 besides the ability to determine when its use is appropriate. Consequently, the estimated training required, to op- erate the device and understand the patient-technology interface, is approximately 12 hours [80]. When introducing new medical technology, it is important to minimize concerns amongst the general public. Therefore, relevant information to share about the MD100 is that it is a radiation-free triage support device that aims to shorten time- to-treatment and thereby improve patient outcomes. 3.3 Safety When implementing the MD100, it is critical to assess the safety aspects for both patients and healthcare professionals to ensure that the benefits outweighs potential harm. This addresses safety considerations from multiple perspectives and outlines relevant safety management practices. Patient safety The first topic in this domain addresses patient safety according to the following issues. ID Issue C0008 How safe is the technology in relation to the comparator(s)? C0002 Are the harms related to dosage or frequency of applying the technology? C0004 How does the frequency or severity of harms change over time or in different settings? C0005 What are the susceptible patient groups that are more likely to be harmed through the use of the technology? C0006 What are the consequences of false positives, false negatives, and incidental findings generated by using the technology from the viewpoint of patient safety? C0007 Are the technology and comparator(s) associated with user- dependent harms? 28 3. Results Since the MD100 is CE-certified, it complies with the European Union’s directive on Electromagnetic Compatibility (EMC), Directive 2014/30/EU. The directive’s essential meaning is "to ensure the electromagnetic compatibility of electrical and electronic equipment" [81]. It outlines requirements which equipment must fulfil to avoid interference with other equipment. The MD100 does therefore meet the requirements for emission limits, immunity levels, conformity assessment procedures, labelling and documentation. In the European Union, there is another directive, namely Directive 2013/35/EU for Electromagnetic fields (EMF). This directive is stated to be the "minimum health and safety requirements regarding the exposure of workers to the risks arising from electromagnetic fields" [82] and aims to ensure that a risk assessment of the EMF is performed at the workplace ensuring the safety of the workers [82]. Since MD100 operates via MWT and complies with the Directive 2014/30/EU, the EMF is ensured through risk analysis and user instructions, as safety information is provided to ensure that Directive 2013/35/EU will be met. The current AI-based algorithm in the MD100 have been trained to demonstrate the highest point of sensitivity on the ROC-curve at 97%, corresponding to a specificity of 48% for stroke detection [80]. With a larger amount of training data, it is a rea- sonable to assume that the model’s performance will improve over time, reaching a higher sensitivity and specificity [83]. This is because algorithms require substantial data to fully converge and reach optimal performance, making them dependent on more extensive datasets. With a high sensitivity, nearly all true stroke cases are cor- rectly identified, with only approximately 3% of stroke cases missed, supporting the device’s safety in identifying potential stroke patients. However, the low specificity suggests a high rate of false positives, meaning that nearly half of the non-stroke patients are incorrectly classified as having a stroke. While this over-triage does not pose a direct safety risk to the patient, it may lead to unnecessary resource utilization and potential delays in care for other conditions. Nonetheless, from a safety perspective, the high sensitivity ensures that stroke patients are unlikely to go undetected when applying the MD100. The same reasoning goes for TBI where the sensitivity is shown to be 100% and the specificity is 75% for subdural hematomas [84]. For reference, the performance metrics of the MD100 can be compared to the current triage assessment scales. As shown in Sections 3.2 and 3.3, the MD100 demonstrates higher sensitivity, suggesting improved patient safety when used in clinical decision- making. Occupational safety The next topic examines the MD100 in regards to occupational safety by answering the issue specified below. ID Issue C0020 What kind of occupational harms can occur when using the tech- nology? 29 3. Results The MWT used in the MD100 operates within a frequency range of 500 MHz to 1.1 GHz, which is within the limits defined by the device’s CE certification [76]. Importantly, there is no evidence suggesting harmful electromagnetic exposure to the device operator. For context, standard mobile phones operate at even higher frequencies, typically between 700 MHz and 2.6 GHz, depending on the network generation (2G to 4G) [85] This highlights that the MD100 uses lower-frequency microwaves and does not puts either the clinician or the patient at risk during usage of the technology. Environmental safety Additionally, environmental safety considerations associated with the technology are addressed in the following section. ID Issue C0040 What kind of risks for public and environment may occur when using the technology? During operation, the MD100 does not emit or require any harmful substances or chemicals, making it safe for both users and the environment. Unlike imaging technologies that may rely on contrast agents or radioactive tracers, the MD100 operates without consumables that require special handling or disposal. Like all battery-driven electronic medical equipment, MD100 must be disposed in accordance with electronic waste regulations at the end of its life cycle. There are no evidence that the device pose any unusual environmental risks beyond those of standard medical electronics. Similarly, the single-use hygienic covers should be recycled or disposed of according to local medical waste guidelines. In addition, CE certification ensures that all materials in contact with the patient are biocompatible and non-toxic, making it suitable for clinical use. Safety risk management The final topic in the safety domain focuses on safety risk management and examines the issues below. ID Issue C0062 How can one reduce safety risks for patients (including technology-, user-, and patient-dependent aspects)? C0063 How can one reduce safety risks for professionals (including technology-, user-, and patient-dependent aspects)? C0064 How can one reduce safety risks for the environment (including technology-, user-, and patient-dependent aspects)? The primary safety considerations for the patient in regards to the MD100 relates to 30 3. Results its software maintenance as well as proper use. The MD100 is intended to support triage and clinical decision-making rather than to provide a definitive diagnosis. This distinction is crucial for healthcare professionals to understand in order to ensure that the use of the device does not compromise patient safety. Therefore, it is important the MD100 is used in accordance with its CE-certification and the specified intended use. Additionally, maintaining up-to-date software is vital to ensure the AI algorithm remains accurate and effective, thereby minimizing risks associated with outdated or malfunctioning. Safety considerations for healthcare professionals include the need for appropriate training to ensure correct use of the device. Proper handling techniques may reduce the risk of strain or injury during repeated use, particularly in high-volume settings such as emergency care or ambulance services. 3.4 Clinical Effectiveness Evaluating the clinical effectiveness of the MD100 is a crucial step in assessing its potential impact on patient outcomes. This domain explores how future implemen- tation within the WC healthcare system may influence patient management and contribute to improved clinical decision-making of stroke and TBI cases. Mortality and morbidity The two first topics in this domain investigates issues related to mortality and mor- bidity as listed below. ID Issue D0001 What is the expected beneficial effect on the technology on mor- tality? D0026 How does the technology modify the effectiveness of subsequent interventions? Early detection of stroke and TBI has the potential to significantly reduce the time- to-treatment. By assisting in triage and complementing current assessment scales, MD100 may facilitate earlier treatment. Evidence shows that earlier treatment im- proves patient outcomes by minimizing long-term disabilities, reducing rehabilita- tion time and increasing survival rates [86], [87]. However, earlier recognition alone does not guarantee earlier treatment which is what the concept of "time is brain" refers to [72]. Improved patient outcomes ultimately depend on whether patients receive appropriate care within the critical treatment windows[87]. In this context, the MD100 serves as a crucial tool in facilitating timely access to care but for the treatment aspect, it requires healthcare facilities to have the necessary resources and capacity to deliver timely and effective treatment. As of today, there are no direct clinical evidence demonstrating the impact of the MD100 on mortality rates. However, since faster treatment is strongly associated with improved outcomes, the MD100 - if it successfully enables earlier intervention - may indirectly contribute 31 3. Results to reduced mortality [87], [88]. Nonetheless, further studies are needed to establish this direct link. Function The function of the technology is addressed by examining the following issues. ID Issue D0011 What is the effect of the technology on patient’s body functions? D0014 What is the effect of the technology on work ability? D0015 What is the effect of the technology on return to previous living conditions? D0016 How does the use of the technology affect activities of daily living? The technology does not directly have an effect on the patient’s body functions. Indirect evidence however, suggests that earlier intervention of stroke and TBI may facilitate improved patient outcomes. With the assumption that the MD100 would support in triage, contributing to a faster diagnosis and thereby earlier treatment, it would reduce the extent of brain tissue damage and thus decrease rehabilitation and likely improve patient outcomes. This includes a higher probability of func- tional recovery, increased ability to return to work and a greater chance of returning previous living conditions. Health-related quality of life In order to investigate the health-related quality of life, the following issues are adressed in this section. ID Issue D0013 What is the effect of the technology on disease-specific quality of life? D0030 Does the knowledge of the test result affect the patient’s non- health-related quality of life Previous studies estimate that every 10-minute reduction in time-to-treatment for stroke results in an average gain of 39 days of disability-free life [72]. As discussed under mortality and morbidity, the MD100 may support earlier identification of stroke, which in turn could enable earlier treatment. Under the assumption that earlier identification leads to earlier treatment, the MD100 may therefore be indi- rectly associated with improved QoL, as earlier treatment is linked to better clinical outcomes [86], [88]. However, it is important to note that no direct evidence cur- rently demonstrates this specific linkage. 32 3. Results Test-treatment chain The test-treatment chain topic answer the following issue. ID Issue D0024 Is there an effective treatment for the condition the test is detect- ing? The main treatments for stroke are outlined and described in Section 1.1.2. How- ever, the effectiveness of these treatments is limited by narrow therapeutic time windows, which restrict the number of patients eligible for intervention. For ex- ample, at Tygerberg Hospital, patients who arrive outside the 4.5-hour window for thrombolysis are not prioritized for a CT scan, as the opportunity for thrombolytic therapy has already passed [34]. In some cases, patients may not receive CT imag- ing at all. Instead, a clinical assessment by a physician confirms a stroke, after which the patient is referred directly to rehabilitation without any technology-based diagnostic evaluation. Similar, treatment decisions for TBI are highly dependent on the location and sever- ity of the hemorrhage and the identified options are detailed in Section 1.1.5. Test accuracy The next topic addresses the test accuracy and contains the following issues. ID Issue D1001 What is the accuracy of the test against the reference standard? D1002 How does the test compare to other optional tests in terms of accuracy measures? D1003 What is the reference standard and how likely does it classify the target condition correctly? D1004 What are the requirements for accuracy in the context the tech- nology will be used? D1006 Does the test reliably rule in or rule out the target condition? D1007 How does the test accuracy vary in different settings? D1008 What is known about the intra- and interobserver variation in test interpretation? D1019 Is there evidence that the replacing test is more specific or safer than the previous one? An article was published in 2024 showing that the sensitivity and specificity of GCS for TBI for moderate to severe brain injury (GCS score ≤ 12) was 83.1% and 93.1% respectively [89]. For stroke, another study was performed in 2018 where GCS was compared to Glasgow Coma Scale motor component (GCS-M), also known as the Simplified Motor Score (SMS). Sensitivity and specificity of the SMS was compared during different times and for different localisation of haemorrhage in the brain. 33 3. Results Depending on these different factors, the sensitivity for GCS-M varied from 66.7 % for when death occurred after 72h from haemorrhage in the left hemisphere to 100 % when death occurred after 0h in the right hemisphere. The corresponding specificities for these scenarios were 82.4 % and 85.7 % [90]. In 2021, a study was done on the performance of SATS in the Western Cape Gov- ernment Emergency Medical Services (WCG EMS) [91]. For trauma, 28.8 % of patients were under-triaged corresponding to a sensitivity of 71.2 % and 13.7% were over-triaged corresponding to a specificity of 86.3% [91]. No performance metrics could be found on the SATS used on stroke patients. Tables 3.2 and 3.3 shows an overview of the tools and their corresponding specificities and sensitivities. Tool Sensitivity Specificity Setting Intended use Reference FAST 77% * 60% * Pre-hospital Initial identification [92] GCS-M 66.7%– 100% ** 82.7%– 85.7% ** Pre-hospital In-hospital Triage [90] SATS N/A N/A Pre-hospital In-hospital Triage - Table 3.2: Sensitivity and specificity of triage tools for stroke. * For acute ischaemic stroke. ** Dependent on time of stroke onset and localisation of the haemorrhage. For stroke detection, MD100 has demonstrated a sensitivity of 97% and a specificity of 48% [93]. Compared to other triage tools, the MD100 exhibits high sensitivity, which is critical for both the initial identification and appropriate triage of patients with suspected stroke. While the GCS-M has reported a sensitivity of up to 100 %, it is important to note that this figure, as previously discussed, refers to a case where the patient died within 0 hours, limiting its generalisability. Although the specificity of the MD100 is currently lower than that of the other tools assessed, this limitation must be balanced against its high sensitivity. A lower specificity implies an increased risk of over-triage, potentially leading to greater strain on healthcare resources. On the other hand, the benefit of a high sensitivity is that it minimizes the risk of missed stroke cases, which is a critical consideration in stroke care management. Additionally, the MD100 is not supposed to replace the triage assessment scales currently used. It should instead act as a complement, which in combination could increase both the current total sensitivity and specificity of triaging stroke and TBI. 34 3. Results Tool Sensitivity Specificity Setting Intended use Reference GCS 83.1% * 93.1% * Pre-hospital In-hospital Triage [89] SATS 71.2% 86.3% Pre-hospital In-hospital Triage [91] Table 3.3: Sensitivity and specificity of triage tools for trauma * For GCS score ≤ 12 in prehospital setting For the detection of chronic subdural haematomas, MD100 demonstrated a sensi- tivity of 100 % and a specificity of 75% [84]. In this context, the MD100 showed the highest sensitivity among the triage tools evaluated for TBI, indicating strong potential for accurate identification of affected patients. While the specificity of the MD100 remains lower than that of some existing tools, this should be interpreted in light of its superior sensitivity. The trade-off between sensitivity and specificity is particularly relevant in emergency and pre-hospital set- tings, where the risk of under-triage may have severe clinical consequences. Thus, the higher sensitivity of the MD100 may justify its relatively lower specificity, espe- cially in scenarios where early detection is critical to patient outcomes. The potential impact on the accuracy across different clinical settings has not yet been thoroughly investigated. A local clinical trial is therefore required to validate the performance of the device within the specific healthcare context and ensure compliance with local conditions and practices [46], [73]. Change-in management The last topic in regards to the clinical effectiveness domain answers the listed issues in regards to the change-in management. ID Issue D0010 How does the technology modify the need for hospitalisation? D0020 Does the use of the test lead to improved detection of the condi- tion? D0021 How does use of the test change physicians’ management deci- sions? D0022 Does the test detect other potential health conditions that can impact the subsequent management decisions? The MD100 is expected to contribute to reduced time-to-treatment, which may also shorten the need for in-hospital care during rehabilitation. When used in pre- hospital settings, the device can support clinical decision-making and potentially allow certain facilities to be bypassed, thereby reducing the burden on lower-level hospitals and optimizing patient flow within the healthcare system. However, this 35 3. Results may concurrently place additional demand on higher-level facilities, which must be considered in system integration. The high sensitivity of the MD100 is particularly advantageous when compared to existing triage assessment scales, as it demonstrates the highest sensitivity 3.2, 3.3, indicating a reduced risk of missed positive cases. This suggests that the use of MD100 may enhance the detection of positive patients, contributing to improved clinical decision-making. As of today, the MD100 can only differentiate between "stroke" or "no-stroke" but MWT has proven potential for a wider medical usage which potentially could make it applicable in other areas in the future [94]. However, this is up for further investigation. 3.5 Cost and economic evaluation Another critical factor to consider particularly given that implementation is being assessed in a LMIC context, is the economic evaluation of the technology. This domain examines the associated costs and resource utilization, while also exploring variations and uncertainties that may influence the overall cost-effectiveness of the MD100. Examination of costs and resource utilisation First, the examination of costs and resource utilisation are examinated by answering the issues below. ID Issue E0006 What are the estimated differences in costs and outcomes between the technology and its comparator(s) E0001 What types of resources are used when delivering the assessed technology and its comparators (resource-use identification)? E0002 What amounts of resources are used when delivering the assessed technology and its comparators (resource-use measurement)? E0009 What were the measured and/or estimated costs of the assessed technology and its comparator(s) (resource-use valuation)? D0023 How does the technology modify the need for other technologies and use of resources? G0007 What are the likely budget impacts of implementing the tech- nologies being compared? In contrast to current triage tools, the use of the MD100 requires a substantial initial investment. The cost per device has been estimated to €90,000 (R1,850,000) and the total procurement cost depends on the number of devices acquired. For the purpose of illustrating the number of devices that may need to be adopted at each level of care, Table 3.4 shows the total number of units or facilities at each level. 36 3. Results Table 3.4: Number of facilities for each suggested implementation site and the purpose of the MD100 in each instance. Number of Devices Implementation Sites Purpose 250 [49] Public EMS (prehospital units) Identify and transport potential stroke patients directly to treatment-capable facilities 75 [33] Level 1 facilities Triage patients to determine stroke likelihood before referral 8 [33] Level 2 facilities Prioritize suspected stroke patients in the CT queue or use when CT scan is not operational 2 [33] Level 3 facilities Prioritize suspected stroke patients in the CT queue The fixed cost of the device includes the MD100 system itself, the cost of training healthcare professionals in operating the device as well as the cost of integrating the device into existing health systems. The variable cost of the device includes the single-use hygiene covers, software updates and maintenance of the device, electricity for charging it and potential training refreshes of healthcare professionals. The MD100 does not require extra staffing beyond already existing personnel as healthcare workers currently responsible for patient assessment and triage will be trained to operate the device. However, the use of MD100 could potentially in- crease the demand for confirmatory CT-scans and thereby place additional strain on resources at the hospitals. Regarding healthcare costs for the patient, a previous study showed that the cumu- lative costs remained the same for the patients who received treatment compared to the ones who did not [72]. The healthcare costs are rather dispersed throughout the life-time, ultimately resulting in the same total costs but with a higher QoL for the patients who received treatment faster [72]. Measurement and estimation of outcomes Next, measurement and estimation of outcomes of the technology are analysed via the issue below. ID Issue E0005 What is (are) the measured and/or estimated health-related out- come(s) of the assessed technology and its comparator(s) (out- come identification, measurement and valuation)? For stroke, every minute counts. From a study performed on stroke patients suffering LVOs, for every minute reduction in time to receiving a thrombectomy resulted in 1.3 days of additional disability-free life [95]. This shows that even the slightest 37 3. Results reduction in time-to-treatment, results in a gain to be found for the patients. With QALYs gained, patients are additionally able to work and by this contribute to society, which has a positive social impact. Another study showed that every 10-minute reduction in time-to-treatment corre- sponds to an average gain of 39 days of disability-free life as well as an INMB of approximately $10,500 [72]. Both studies draw the conclusion that every minute re- duction in time-to-treatment adds to the patient outcomes as well as the economical societal impact [72], [95]. Appendix A illustrates the amount of time that must be saved over the entire life- time of the MD100 for the application to be considered economically viable. The calculation includes estimated values, or numbers sourced from relevant literature. Here, the previously mentioned stu