Optimizing Prehospital Acute Transfer of Patients With Presumed Stroke Given Economic Constraints

ABSTRACT

Background

Treatment with mechanical thrombectomy (MT) remains inaccessible for many patients with acute ischemic stroke (AIS) due to large vessel occlusion (LVO) and under‐utilization prevails across healthcare systems. Increasing the number of thrombectomy centers and ambulance helicopters may alleviate these issues.

Aim

This study aims to determine the most effective combination of optimally located ambulance helicopters and thrombectomy centers for the economically constrained healthcare system.

Methods

This nation‐wide, observational study analyses anonymized patient‐level registry data stretching over a 6‐year study period in Sweden. It combines optimization modeling with cost‐effectiveness analysis to generate combinations of optimally located thrombectomy centers and ambulance helicopters to compare with the current eight locations of thrombectomy centers in Sweden and no ambulance helicopters. The analysis extends to evaluate the cost‐effectiveness of increasing the number of thrombectomy centers and ambulance helicopters when the current eight locations remain fixed.

Results

The most cost‐effective solution comprises 11 thrombectomy centers and 14 ambulance helicopters, corresponding to densities of 1.05 and 1.34 per one million inhabitants, respectively. It yields an estimated annual incremental net monetary benefit (INMB) close to €13.6 million. In the extended scenario analysis, the most cost‐effective solution comprised nine thrombectomy centers and 13 ambulance helicopters, with an estimated annual INMB of €3.8 million.

Conclusions

The most cost‐effective combination of optimally located thrombectomy centers and ambulance helicopters brings about substantial health gains for patients with AIS due to LVO, compared with the current eight locations of thrombectomy centers in Sweden and ambulance helicopters.

1. Introduction

Reperfusion therapy in the form of intravenous thrombolysis (IVT) and mechanical thrombectomy (MT) has become the standard of care in eligible patients with acute ischemic stroke (AIS) due to anterior circulation large vessel occlusion (LVO). Concurrently with technological advances in medical devices and further refinement of endovascular techniques, the indication for EVT keeps expanding; it encompasses select patients in the late time window beyond 6 h and up to 24 h from last seen well, patients with vertebrobasilar occlusions or large core infarcts [1, 2, 3, 4, 5]. To reap the benefits of these advancements, patients need access to specialized acute stroke care, and measures that reduce the time from symptom onset to treatment start (OTT) still entail substantial improvements of functional outcomes in the majority of patients treated with acute reperfusion therapies [6, 7]. The lingering inaccessibility to and under‐utilization of MT across stroke systems of care prompt further implementation of thrombectomy services [8]. To overcome the prevailing shortcomings, it is necessary to either shorten the geographical distance or increase the speed of travel between patients and thrombectomy services, preferably in the most cost‐effective possible way. Currently, seven comprehensive stroke centers (CSC) and one thrombectomy‐capable stroke center (TSC) serve a population of 10.5 million inhabitants in Sweden. Approximately 60% of patients treated with MT arrive at a thrombectomy center by interhospital transfer from the first admitting IVT‐ready hospital [9]. Due to the current geographical distribution of ambulance helicopters and in the void of a nationally coordinated system for airborne transportation of patients with presumed stroke due to LVO, very few patients arrive at thrombectomy centers by ambulance helicopter [10]. The cost‐effectiveness of increasing the number of optimally located CSCs or TSCs has been evaluated previously [11]. The alternative implementation strategy of increasing the number of optimally located ambulance helicopters while keeping thrombectomy centers fixed has been evaluated within the modeling framework of cost‐effectiveness analysis too [12]. Indeed, these studies have demonstrated how to attain the most cost‐effective number and locations of thrombectomy centers and ambulance helicopters, respectively, in separate analysis. Although methodologically and computationally challenging, the ability to conjunctively model the optimal number and locations of thrombectomy centers and ambulance helicopters would capacitate searching for even more cost‐effective implementation strategies by delineating all possible combinations within the probable range of cost‐effectiveness on the vast solution space. It may offer insights into the interaction effect of thrombectomy centers and ambulance helicopters on patient outcomes and costs too. This study aims to determine the most cost‐effective combination of optimally located ambulance helicopters and thrombectomy centers compared with the current eight thrombectomy centers in Sweden and no ambulance helicopters in patients with presumed AIS due to LVO.

2. Methods

This study optimizes prehospital acute transfer of patients with presumed stroke by combining data from national quality registers in geographic information system network analyses and in an economic modeling framework for decision analysis.

2.1. Data

The study material derives from a consolidated dataset of anonymized, individual patient‐level registry data on acute stroke patients spanning over a 6‐year period between 2012 and 2017, that includes emergency medical services (EMS) call‐out data from emergency call operator companies, inpatient healthcare episodes, and eventual cause of death data from the Swedish National Board of Health and Welfare, and stroke care data from the Swedish stroke registry (RIKSSTROKE) [13, 14, 15, 16]. It contains 220,267 EMS records on call‐outs to patients with suspected stroke, and 124,484 case records of patients discharged from hospital with confirmed stroke diagnosis encoded with the tenth revision of the International Classification of Diseases (ICD‐10). A description of patient characteristics and selection criteria has been detailed previously [6].

The study population for analysis consists of 18,793 cases of patients with suspected AIS and potential eligibility for MT treatment. It encompasses all patients with a hospital discharge diagnosis code for stroke due to cerebral infarction (I63) presenting with a National Institute of Health Stroke Scale (NIHSS) score greater than or equal to six (NIHSS ⩾ 6) at hospital admission (n = 13,355), and all patients presenting with an NIHSS score less than six (NIHSS < 6) at hospital admission who received treatment with MT (n = 164).

The study population for analysis additionally encompasses 5247 patient cases of false positives, constituted by patients with intracerebral hemorrhage (I61) (n = 2945) and stroke mimics (n = 2302) to reflect the proportion of false positives among patients assessed with the prehospital stroke triage system termed the A2L2 test in Sweden [17].

2.2. Modeling

2.2.1. Geographic Network Analysis

Geographic network analyses are conducted within the software environment of ESRI ArcGIS Desktop 10.6.1. and employ the national road network from the Swedish national road database and a created network of geodesic distances. The built‐in solvers of ArcGIS employ a range of heuristics to find good solutions; this includes semi‐randomized initial solutions, a vertex substitution heuristic, and a metaheuristic. The number of candidate facilities for locating thrombectomy centers includes the current seven CSCs and one TSC in Sweden and four CSCs in neighboring countries, namely in Copenhagen (Denmark), Oulu (Finland), Oslo (Norway), and Trondheim (Norway). It also includes 55 IVT‐ready hospitals in Sweden, making the total number of candidate facilities 67. The number of candidate heliports for locating ambulance helicopters is 35. On a catchment area spanning over 530,000 km² over land and water, with an estimated population of 10.5 million inhabitants, each patient case is represented by the pick‐up point of location in network analyses [18].

The strategic decision problems for locating thrombectomy centers and ambulance helicopters translate into p‐median facility location–allocation problems. The network analyses for EMS patient transportations include providing solutions for both the Drip‐and‐Ship (DS) and Mothership (MS) organizational paradigms, respectively. The model estimates EMS travel times using the quickest path in the road network calculated with road length and maximum allowed speed for each road link. The network analyses for locating ambulance helicopters provide solutions for the MS paradigm and estimate HEMS travel times using the shortest path in the Euclidean network of geodesic distances.

Additionally, the model applies a maximum constraint on travel time and distance for patient transportation with EMS and helicopter emergency medical services (HEMS), respectively. Moreover, the model requires the allocation of at least 50 MT interventions per year for a candidate facility to qualify as a potential facility for locating thrombectomy centers [19].

The given set of candidate facilities for locating thrombectomy centers and ambulance helicopters makes the number of possible combinations large for some solutions (Appendix S1). Thus, results from previous studies motivate limiting the solution space to solving the location‐allocation problem for n = 8, …, 12 thrombectomy centers using the road network, and for n = 5, …, 16 ambulance helicopters using the geodesic network, to find the most cost‐effective combination of optimally located thrombectomy centers and ambulance helicopters [11, 20]. Additionally, the analysis provides solutions to n = 8, …, 12 thrombectomy centers while holding fixed the current 8 thrombectomy centers in Sweden.

The model sets the upper limit of the OTT window to 270 min for IVT and 360 min for MT. Furthermore, the model assumes fixed time lapses for some actions and processes in the acute stroke care management (Table S1). Hence, the maximum allowed travel time for EMS transportation of a patient from the pick‐up point of location to the nearest IVT‐ready hospital is 170 min. The corresponding upper limits to the nearest thrombectomy center under the Drip‐and‐Ship (DS) and Mothership (MS) paradigms are 185 and 240 min, respectively. The maximum allowed travel time at disposal for HEMS from heliport to the patient pick‐up point of location, and then onwards to the nearest thrombectomy center is 227.6 min. This translates into a travel distance of 1024 km at the average cruising speed of 270 km/h.

2.2.2. Costs and Health Effects

The patient‐level cost consists of the staffing cost of running thrombectomy center services and medical equipment costs associated with the respective treatment modality in addition to the individually estimated distance‐based cost for patient transportation with EMS by the DS and MS paradigm, respectively, and with HEMS. A full breakdown of cost items related to the operability of thrombectomy centers and the medical equipment costs associated with the respective treatment modality has been delineated previously [11].

The estimated patient‐level costs for the 1st and 2nd year post‐stroke according to mRS category are obtained from literature and take on a societal perspective [21]. Costs are converted into 2021 euro, using the average exchange rate for the year 2021 between Swedish krona and euro: €1 = SEK10.515.

Patient age and admission NIHSS score remain fixed in each patient case, while the calculated OTTs for IVT and MT vary across solutions. By applying predictive generalized linear models (GLMs) (one for each treatment modality), the model estimates the mRS‐90d score in each patient case and for all available combinations of mode of transportation, organizational paradigm, and treatment modality in a solution. Moreover, the model selects the mode of transportation, organizational paradigm, and treatment modality that minimize the expected mRS‐90d score in each individual patient case. Thus, the preferred mode of transportation, organizational paradigm, and the availability and clinical effectiveness of different treatment modalities in individual patients hinge upon the proximity to ambulance helicopters and thrombectomy centers that vary across solutions.

The selected mRS‐90d scores are then converted into utility weights obtained from literature [22]. Three‐year survival rates are calculated using study population data and according to mRS categories 0–2, 3, 4, and 5. The age‐adjusted, annual survival rate trend in patients with ischemic stroke aligns with that of the Swedish reference population at 3 years post‐stroke. Therefore, the age‐adjusted survival rate trend in the Swedish reference population serves as the basis for calculating survival rates from year four and onward (Table S2) [23, 24].

Outcome distributions of the remaining quality‐adjusted life years (QALY) for patients of each mRS category are obtained with a time‐inhomogeneous, discrete‐time Markov chain (DTMC) [11, 25].

2.2.3. Measures of Cost‐Effectiveness Within the Decision‐Analytical Framework

The selected cost‐effectiveness measures derive from patient‐level costs and QALYs and consist of the net health benefit (NHB), the net monetary benefit (NMB), and the incremental NMB (INMB). These metrics are innately suitable in cost‐effectiveness analysis (CEA) with more than two comparators; fixed quantity measures of cost‐effectiveness make cross comparisons and ranking of comparators seamlessly easy to undertake. Nonetheless, the main focus of the CEA is to compare solutions to various combinations of optimally located thrombectomy centers and ambulance helicopters, with the status quo of thrombectomy centers in Sweden. The solution that attains the highest expected INMB is selected as the most cost‐effective combination of optimally located thrombectomy centers and ambulance helicopters in the prehospital acute stroke care system for triage‐positive patients due to suspected LVO AIS.

2.2.4. Modeling Assumptions and Scenarios

The modeling scenarios assume 1229 eligible candidates for treatment with MT per year among an estimated 1708 triage‐positive patient cases and reflect the national MT rate at 7% of all confirmed cases of patients with AIS in Sweden during the year 2021 [26]. The maximum willingness‐to‐pay (WTP) per QALY gained set at €80,000 represents the lowest cost per QALY gained among declined reimbursements of treatments in severe health conditions by the Swedish Dental and Pharmaceutical Benefits Agency [27].

The base‐case solution assumes the current number and locations of thrombectomy centers in Sweden as per the year 2023, comprising seven comprehensive stroke centers (CSC) and one thrombectomy‐capable stroke center (TSC). It corresponds to a thrombectomy center density of 0.77 per one million inhabitants. It has recently been suggested that the most cost‐effective number of optimally located TSCs to complement the CSCs in Sweden is four, setting the said density to 1.05 per one million inhabitants [11].

The main scenario assumes no predetermined locations of thrombectomy centers and is tasked with determining the most cost‐effective combination of freely located thrombectomy centers and ambulance helicopters. In the secondary scenario, the analysis sets out to determine the most cost‐effective combination of optimally located ambulance helicopters and the 9th, …, 12th optimally located thrombectomy centers, respectively, to complement the current 8 thrombectomy centers.

2.2.5. Deterministic Sensitivity Analysis

The deterministic sensitivity analysis (DSA) examines the sensitivity in results by varying the maximum WTP per QALY gained in the range between €0 and €200,000.

3. Results

3.1. Accessibility to MT, Organizational Paradigms, and Utilization of HEMS

With the current eight thrombectomy centers and no operational ambulance helicopters, it is estimated that 98.5% of all patients with presumed acute stroke have access to treatment with MT within 360 min from symptom onset. In terms of modeled patient outcomes, the DS pathway is preferred over the MS pathway in 29.9% of all patient cases. In comparison, the modeled solution with eight optimally located thrombectomy centers and no ambulance helicopters provides access to MT for 99.8% of all patients with presumed acute stroke, and prefers the DS pathway in 24.5% of all patient cases. With the introduction of five optimally located ambulance helicopters, the preference for the DS pathway decreases to 20.8%, and the ambulance helicopter becomes the preferred mode of transportation in 9.5% of all patient cases. With 12 optimally located thrombectomy centers and 16 optimally located ambulance helicopters, the preference for the DS pathway plunges to 15.9%, and ambulance helicopters handle 11.5% of all patient transportations.

3.2. OTTs, Costs, and QALYs

The lowest achievable mean OTT to IVT with the fewest number of located thrombectomy centers and ambulance helicopters was 127 min in the solution with 10 thrombectomy centers and 10 ambulance helicopters. The corresponding solution for achieving the lowest mean OTT to MT comprised 12 thrombectomy centers and 12 ambulance helicopters. The maximum QALY production per year (3100 QALYs) was attained by locating 12 thrombectomy centers and 16 ambulance helicopters (Table 1).

TABLE 1.

Modeling outcomes for all subset combinations of 8–12 thrombectomy centers and 5–16 ambulance helicopters.

Number of ambulance helicopters
MTCs	Outcomes	Comparator	0	5	6	7	8	9	10	c10	11	12	13	14	15	16
8	OTTIVT	126	122	124	124	124	124	124	124	123	124	124	124	124	125	125
8	OTTMT	170	144	143	143	143	143	142	142	139	142	142	142	142	141	141
8	QALYS	2936	3048	3064	3066	3066	3067	3068	3069	3079	3069	3070	3072	3073	3074	3074
8	Cost	177.5	175.5	175.6	175.6	175.7	175.7	175.7	175.7	176.1	175.7	175.6	175.6	175.7	175.7	175.9
8	INMB per patient	0	8947	9902	10,022	9974	10,054	10,084	10,097	10,408	10,206	10,271	10,468	10,448	10,397	10,326
9	OTTIVT	NA	122	123	123	123	123	124	124	123	124	124	124	124	124	124
9	OTTMT	NA	140	140	140	140	139	139	138	136	139	139	139	139	138	138
9	QALYS	NA	3059	3074	3075	3077	3078	3078	3086	3092	3080	3081	3083	3084	3084	3085
9	Cost	NA	176.0	176.1	176.1	176.1	176.1	176.1	176.4	176.7	176.0	176.0	176.0	176.0	176.2	176.3
9	INMB per patient	NA	9286	10,148	10,236	10,327	10,389	10,398	10,648	10,784	10,601	10,660	10,800	10,807	10,760	10,700
10	OTTIVT	NA	122	123	123	123	123	123	123	123	123	123	124	124	124	124
10	OTTMT	NA	139	139	139	139	139	138	135	144	138	138	138	138	137	138
10	QALYS	NA	3067	3081	3082	3084	3085	3085	3097	3062	3088	3088	3089	3090	3091	3091
10	Cost	NA	176.2	176.3	176.4	176.3	176.4	176.4	177.1	175.6	176.3	176.3	176.3	176.4	176.5	176.6
10	INMB per patient	NA	9635	10,420	10,464	10,562	10,608	10,624	10,766	9725	10,855	10,901	10,936	10,951	10,913	10,860
11	OTTIVT	NA	121	122	122	122	122	122	123	122	123	123	123	123	123	123
11	OTTMT	NA	136	136	136	136	136	136	140	139	136	136	135	135	135	135
11	QALYS	NA	3074	3088	3088	3090	3090	3091	3074	3082	3093	3094	3095	3096	3096	3096
11	Cost	NA	176.5	176.6	176.6	176.6	176.7	176.7	176.0	176.2	176.6	176.6	176.6	176.7	176.8	176.9
11	INMB per patient	NA	9826	10,588	10,632	10,720	10,741	10,772	10,248	10,572	10,972	11,007	11,048	11,050	11,006	10,966
12	OTTIVT	NA	121	122	122	122	122	122	122	122	122	122	122	123	123	123
12	OTTMT	NA	135	135	135	135	135	135	136	135	135	135	135	135	135	135
12	QALYS	NA	3079	3092	3093	3094	3095	3096	3088	3092	3097	3097	3098	3099	3099	3100
12	Cost	NA	177.0	177.1	177.1	177.1	177.1	177.1	176.6	177.0	177.1	177.2	177.1	177.2	177.3	177.4
12	INMB per patient	NA	9791	10,545	10,588	10,674	10,725	10,755	10,695	10,584	10,770	10,806	10,847	10,852	10,812	10,776

3.3. Cost‐Effective Combinations of Thrombectomy Centers and Ambulance Helicopters

Among all studied combinations of optimally located thrombectomy centers and ambulance helicopters, the solution with 11 thrombectomy centers and 14 ambulance helicopters was the most cost‐effective in comparison with the current eight thrombectomy centers and no ambulance helicopter operability (Figure 1). It corresponds to a thrombectomy center density of 1.05 per one million inhabitants and an ambulance helicopter density of 1.34 per one million inhabitants. This translates into a ratio of circa 4:5 between thrombectomy centers and ambulance helicopters. The most cost‐effective solution has an estimated annual INMB close to €13.6 million, which translates into an average INMB per patient of €11,050 (Figure 2).

FIGURE 1.

The most cost‐effective solution comprises 11 optimally located C‐/TSCs and 14 optimally located ambulance helicopter.

FIGURE 2.

Incremental net monetary benefit per year (€) for combinations of optimally located thrombectomy centers and ambulance helicopters, with maximum WTP per QALY gained set at €80,000. (The oversized data point represents the most cost‐effective solution. Solutions with the current 10 locations for ambulance helicopters are highlighted in black).

In the scenario analysis with the first eight thrombectomy centers locked at the current locations of thrombectomy centers in Sweden, the highest annual NMB was reached with a combination of one additional thrombectomy center and 13 ambulance helicopters, making it the most cost‐effective solution (Figure 3). This is equivalent to densities of 0.86 and 1.24 per one million inhabitants for thrombectomy centers and ambulance helicopters, respectively. The solution would generate an annual INMB of €3.8 million, with an average INMB per patient equal to €3125 (Figure 4).

FIGURE 3.

The most cost‐effective solution when the first eight thrombectomy centers are locked at current locations comprises 9 thrombectomy centers and 13 ambulance helicopters.

FIGURE 4.

Secondary scenario: Incremental net monetary benefit per year (€) for combinations of 8 fixed thrombectomy centers at current locations, the 9th up to the 12th optimally located thrombectomy center and 5 to 16 ambulance helicopters, with the maximum WTP per QALY gained set at €80,000. (The oversized data point represents the most cost‐effective solution. Solutions with the current 10 locations for ambulance helicopters are highlighted in black).

The model did not select the CSCs of neighboring countries in any solution.

3.4. Varying the Maximum WTP per QALY Gained

When solutions were compared in the maximum WTP per QALY gained range between €0 and €200,000, cost‐effectiveness emerged in the solution with 11 C‐/TSCs and 13 ambulance helicopters when the maximum WTP per QALY gained settled at €57,077. This combination prevailed as the most cost‐effective solution until the maximum WTP per QALY gained reached €77,707, when it was overtaken by the combination comprising 11 C‐/TSCs and 14 ambulance helicopters. It remained as the most cost‐effective solution until the maximum WTP per QALY gained hit €195,144, when the combination of 12 C‐/TSCs and 15 ambulance helicopters became the most cost‐effective solution for the remaining range up to €200,000.

4. Discussion

This interdisciplinary study employs applied operational research methodologies and applied health economics to evaluate a wide range of combinations of optimally located thrombectomy centers and ambulance helicopters within the decision‐analytical framework of cost‐effectiveness modeling, using individual, patient‐level registry data and current evidence from the literature. The most cost‐effective solution set the densities of thrombectomy centers and ambulance helicopters to 1.05 and 1.34 per one million inhabitants, respectively. The solution generates substantial health gains in comparison with the current density of 0.77 thrombectomy centers per one million inhabitants in Sweden, and no ambulance helicopter operability. In the scenario analysis with the first eight thrombectomy centers locked at the current locations, the most cost‐effective solution settled the densities of thrombectomy centers and ambulance helicopters at 0.86 and 1.24 per one million inhabitants, respectively. It may be noted that the base‐case scenario that constitutes the comparator in the cost‐effectiveness analysis does not mirror the current ambulance helicopter operability in the Swedish healthcare system with 10 operating ambulance helicopters.

It has previously been shown that both the number of thrombectomy centers and ambulance helicopters are important factors to consider in the further development of acute stroke care systems for patients with presumed stroke due to LVO and potential eligibility for treatment with MT [11, 28]. This study demonstrates that the choice of locations for thrombectomy centers has a great impact on results too. Results show that the optimal number and locations of thrombectomy centers shift with the optimal number and locations of ambulance helicopters. Thus, to design a cost‐effective acute stroke care system for patients with presumed AIS due to LVO requires the capability to evaluate the potential number and locations of thrombectomy centers and ambulance helicopters in conjunction. Partial optimization of thrombectomy center locations has a decisively adverse impact on the cost‐effectiveness of solutions, which the scenario analysis with the first eight thrombectomy centers locked at current locations exemplifies with distinct clarity.

The sensitivity analysis shows that results are sensitive to the maximum WTP per QALY gained. Three different solutions interchanged the position as the most cost‐effective solution over the studied maximum WTP per QALY range.

The comprehensive dataset connecting data from emergency call operator services with data from national quality registries to create individual, patient‐level registry data provides the detailed information on each single patient case required for making real‐world, patient‐level cost‐effectiveness analyses. The interdisciplinary approach of applied health economics and operational research facilitates the comprehensive economic evaluation of acute stroke care systems in regard to thrombectomy centers and ambulance helicopters.

While the solvers provided by ArcGIS do not guarantee mathematically optimal solutions, it is unlikely that it has had any major impact on results; all the numerical analyses are valid for the presented solutions. The reliability of results is limited to the Swedish healthcare setting. The underlying study population data for analysis stems from a 6‐year study period between 2012 and 2017 when the indication for MT was limited to the narrow time window of 360 min from symptom onset. Therefore, results may not reflect the extended indication for thrombectomy and current reperfusion rates and patient outcomes from endovascular therapies. On the basis of the tissue‐clock selection paradigm, the benefit of thrombectomy may extend well beyond the narrow time window for select patients with target mismatch profiles. Indeed, the extension of ischemic injury is not perfectly correlated with the time lapsed from symptom onset, as the prevalence of the “late window paradox” phenomenon, particularly pronounced in patients with large‐core ischemic stroke shows [29]. However, it does not follow that shorter onset‐to‐treatment time has no impact on functional outcomes [30]. Still, most patients benefit from shortened onset‐to‐treatment time to acute reperfusion therapies. Furthermore, whether patients are fast or slow progressors, eligible for treatment with IVT, MT or IVT + MT remains uncertain in the prehospital phase of acute stroke care. Therefore, the most cost‐effective combination of the optimal number and locations of thrombectomy centers remains valid and informatively applicable to real‐world settings for as long as confirmation of diagnosis, occlusion site(s), collateral status, and other characteristics underlying treatment decisions depend on in‐hospital examination. However, and contingent upon availability of adequate individual‐level patient data, it seems reasonable to hypothesize that the incorporation of extended indications for thrombectomy and thrombolysis in future analyses would favor Drip‐and‐Ship over Mothership in more patient cases than the current analysis.

This study paves the way for addressing issues of inaccessibility to and under‐utilization of endovascular reperfusion therapies with a comprehensive take on both prehospital modes of transportation and thrombectomy center density and locations, guided by the decision‐analytical framework of cost‐effectiveness analysis. Efforts to improve patient outcomes following stroke are plenty and conducted across a vast field of disciplines. To keep up with the latest advancements in the field of acute stroke care and in particular with the fast pace in the development of new drugs and medical devices for improved rates of successful reperfusion in patients with LVO AIS, cost‐effectiveness analyses need updating on a regular basis to comprise a reliable source of information to support healthcare decision‐making. Therefore, more frequent and regular reassessments of results from cost‐effectiveness analyses in the field of acute stroke care systems are warranted.

5. Conclusion

Compared with the current eight thrombectomy center locations in Sweden, and assuming no ambulance helicopter operability, the most cost‐effective combination of optimally located thrombectomy centers and ambulance helicopters comprises 11 optimally located thrombectomy centers and 14 optimally located ambulance helicopters. The commensurable densities are 1.05 thrombectomy centers and 1.34 ambulance helicopters per one million inhabitants, respectively. It constitutes a cost‐effective solution that would generate substantial health gains in patients with AIS due to LVO.

Author Contributions

Nicklas Ennab Vogel: conceptualization, investigation, methodology, validation, visualization, writing – review and editing, software, project administration, formal analysis, data curation, resources, writing – original draft. Lars‐Åke Levin: investigation, conceptualization, funding acquisition, supervision, writing – review and editing, resources. Tobias Andersson Granberg: writing – review and editing, validation, supervision, funding acquisition, resources. Per Wester: writing – review and editing, supervision.

Ethics Statement

Ethical approval for this study was obtained from The Swedish Ethical Review Authority with approval numbers/IDs: Dnr 2017/487–31 and Dnr 2019–00721.

Consent

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Appendix S1.

Data Availability Statement

The data that support the findings of this study are available on request from any qualified researcher with the appropriate ethics approval upon reasonable request to the corresponding author.