Work package 2

Baseline characteristics of intra-arterial thrombectomy service provision in England

A full report has been published and parts of this section have been reproduced from the report.58 Flynn D, Ford GA, McMeekin PJ, White P. International Journal of Stroke (volume 11, issue 8), pp. NP85–85, copyright © 2016 by World Stroke Organization. Reprinted by permission of SAGE Publications.

Results from randomised trials demonstrating that rapid IAT treatment reduced dependency for selected stroke patients became available in 201523,24 and led to interventional procedure guidance being issued by NICE in March 2016,59 but very few UK hospitals were able to provide this service. Most treatments were opportunistic in the absence of clearly defined clinical pathways, and there was no information about how interventional neuroradiologists (INRs) performing the procedure were interpreting the evidence. Understanding baseline service provision and patient selection processes was an essential first step towards modelling optimal NHS service provision.

Methods

In November/December 2014, a survey was sent to clinical leads in all 24 regional INR services in England to obtain data on current provision of IAT, IAT patient selection and diagnostic imaging criteria, and centre opinions on future IAT service provision.

Results

Eighteen centres (75%) responded that provided INR services to a population of ≈ 43 million. No centres reported delivery by non-INRs. Ten centres (56%) had formal IAT protocols and six (33%) had protocols for interhospital transfers. A median of 10 [interquartile range (IQR) 16] stroke patients underwent IAT per centre during the previous year. One centre had 24 hours per day, 7 days per week (24/7) IAT provision, two centres had 7-day provision during normal hours, 12 delivered IAT on weekdays and three had no regular provision at all. There was substantial variation in the patient selection criteria and protocols for provision of IAT (Table 3). For future IAT services, there was clear support for centralisation of provision into large HASUs at neuroscience centres (89%) with ‘drip and ship’ interhospital transfers across a formal network (94%).

Conclusions

Most centres in England in 2014 were limited to weekday ad hoc provision of IAT. There was considerable variation across centres in imaging and patient selection for IAT, but the responses to questions about the organisation of future service provision for IAT in England showed a degree of consensus.

Updated estimates of certainty for intra-arterial thrombectomy effectiveness and safety

The full systematic review has been published and parts of this section have been reproduced from the review.60 Flynn D, Francis R, Halvorsrud K, Gonzalo-Almorox E, Craig D, Robalino S, et al. European Stroke Journal (volume 2, issue 4), pp. 308–18, copyright © 2017 by European Stroke Organization. Reprinted by permission of SAGE Publications.

The PRISMA-compliant protocol61 was registered online as PROSPERO CRD42015016649.

Soon after the programme started, several meta-analyses of IAT RCTs were published,23,24,62,63 each of which had taken a slightly different approach, but all found that IAT is an effective treatment. The most robust of these was the academic collaborative triallists (the HERMES group) individual patient record meta-analysis.23 However, the evidence base to define the safety and effectiveness of IAT had expanded by ≈ 45%, with THERAPY (The Randomized, Concurrent Controlled Trial to Assess the Penumbra System’s Safety and Effectiveness in the Treatment of Acute Stroke),1 THRACE (The Contribution of Intra-arterial Thrombectomy in Acute Ischemic Stroke in Patients Treated With Intravenous Thrombolysis)2 and PISTE (Pragmatic Ischaemic Stroke Thrombectomy Evaluation)3 trials all reporting later in 2016. Therefore, an updated evidence synthesis was warranted. A systematic review and meta-analysis with trial sequential analysis (TSA) was conducted to understand the impact of trials reporting in 2016 on the magnitude/certainty of the estimates for effectiveness and safety of IAT.

In parallel with this systematic review looking at efficacy and safety, a detailed narrative review was undertaken of IAT complications.64

Methods

The search strategy is shown in Appendix 5. Random-effects models were conducted of RCTs comparing IAT with/without adjuvant thrombolysis against thrombolysis and other forms of best medical care in the treatment of acute ischaemic stroke. The TSA established the strength of the evidence derived from the meta-analyses.

Results

Patients treated with IAT were significantly more likely to be functionally independent (mRS score of 0–2) at 90 days’ follow-up (OR 2.39, 95% CI 1.88 to 3.04). As shown in Figure 7, the impact of the three 2016 trials1–3 was a slightly decreased pooled effect size, but increased certainty of the mid-point estimate (OR 2.07, 95% CI 1.70 to 2.51).

FIGURE 7 Forest plot for functional independence. M–H, Mantel–Haenszel. Reproduced with permission from Flynn et al.60 Flynn D, Francis R, Halvorsrud K, Gonzalo-Almoroz E, Craig D, Robalino S, et al., European Stroke Journal (Volume 2, Issue 4), pp. 308–18, copyright © 2017 by SAGE Publications Ltd. Reprinted by permission of SAGE Publications Ltd. (Figure not reproduced here; see the original article.)

Intra-arterial thrombectomy, compared with best medical care, did not show any effect on mortality or symptomatic intracerebral haemorrhage at 90-days’ follow-up. Results of the TSA satisfied the criterion for ‘sufficient evidence’ on effectiveness; however, uncertainty remains as to whether IAT is associated with lower mortality or increased risk of symptomatic intracerebral haemorrhage.

Conclusions

The expanded evidence base for IAT yielded a more precise assessment of effectiveness, but uncertainty remained as to the net effect of IAT on mortality and symptomatic intracerebral haemorrhage.

Expert consensus on preferred implementation option for intra-arterial thrombectomy services

A full description is available in Halvorsrud et al.70 Parts of this text have been reproduced with permission from Halvorsrud et al.70 This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (https://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. The text below includes minor additions and formatting changes to the original text.

Although there was evidence supporting IAT therapeutically, there was no consensus on how it should be made available within clinical services. The aim was to establish consensus on options for future organisation of IAT services, based on agreement among physicians with clinical expertise in LAO stroke management.

Methods

A survey questionnaire was developed with 12 options (propositions) for future organisation of thrombectomy services in England (see Appendix 6, Box 6). The British Association of Stroke Physicians (BASP) facilitated recruitment of panellists to provide representative ratings of options (by location and experience). Consensus was defined as ≥ 75% of ratings for each option falling within three categories (i.e. approve, quite strongly approve or very strongly approve) on a seven-point Likert scale. Wider BASP membership and members of the British Society of Neuroradiologists (BSNR) then ranked those propositions on a seven-point Likert scale, reaching consensus following two initial assessment rounds. Data was collected from November 2015 to March 2016, and participation was pseudo-anonymous.

Results

Eleven respondents completed two rounds.71 Three options achieved consensus:

1. selective transfer to nearest neuroscience centre for INR-delivered IAT (100% approve)
2. local imaging then transfer to nearest neuroscience centre for INR-delivered IAT (91% approve)
3. local imaging then transfer to nearest neuroscience centre for advanced imaging and INR-delivered IAT (82% approve).

Subsequently, the wider group of BASP and BSNR members (n = 64) assigned the highest approval ranking for transferring LAO stroke patients to the nearest neuroscience centre for thrombectomy based on the results of local CT/computed tomography angiography (CTA) (option 2).

Conclusions

The Delphi exercise by clinical and imaging stroke experts in England established consensus on a ‘simple’ imaging-driven option, which advocates the secondary transfer of patients (‘drip and ship’) with LAO stroke for thrombectomy based on local CT/CTA alone.

Patient and public preferences on attributes of intra-arterial thrombectomy service organisation

To integrate the opinions of stroke survivors/relatives/carers and the public into the outputs of a health economic model for IAT in England, their preferences were elicited regarding attributes of IAT service provision, including thresholds for additional travel time in the context of this time-critical emergency.

Methods

The research programme patient and public involvement (PPI) representative facilitated recruitment of 14 stroke survivors and their relatives/carers from a stroke PPI panel to engage in an iterative process to develop an inclusive and accessible ‘aphasia-friendly’ survey. Interactive meetings were convened to obtain feedback on an (1) initial draft of the survey and (2) updated version with graphics and textual presentation that adhered to guidance on developing resources for people with aphasia.72 Subsequent testing was undertaken with 10 stroke survivors/carers and local PPI panels. The resulting anonymous survey was hosted and advertised on the Stroke Association website73 (from January 2017 to May 2017) with information on IAT, including the time sensitive nature of outcomes, limited number of specialist facilities and delays that can occur during secondary transfer from a local stroke unit.

Results

Responses were received from 147 individuals [mean age 49 (SD 16) years; 61% female]: 27 stroke survivors (18%), 51 relatives/carers of stroke survivors (35%) and 69 other members of the public (47%). The majority of stroke survivors were male and the majority of the other groups were female. Respondents were spread across England (50% were resident in north-east England). Full details of respondents are in Appendix 7, Table 17. The survey results are show in Table 4. Differences in proportions for each survey item as a function of participant type were not statistically significant (p > 0.05), indicating that experience of stroke did not influence responses.

Conclusions

In agreement with expert views favouring centralised provision of IAT with secondary transfer as appropriate, 97% of respondents would accept hospital transfer and 75% would be prepared to travel up to 30 miles to access IAT.

Establishing the number of stroke patients eligible for intra-arterial thrombectomy

A full report has been published.74

Material throughout this section has been reproduced with permission from McMeekin et al.74 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The text below includes minor additions and formatting changes to the original text.

To model optimal national service configurations, it was necessary to first estimate the proportion of UK stroke patients eligible for IAT, as this had not previously been established at a population level.

Methods

Using national registry data from the SSNAP for England, Wales and Northern Ireland,6 adjusted for Scotland using data from the Scottish Stroke Care Audit,75 a decision tree with 14 nodes (A–N) was constructed from published trials74 depicting eligibility for IAT, regardless of geographical or service constraints. An updated version in 201876 included two new reports describing IAT efficacy among late presenters.77,78

Results

The updated decision tree is presented in Figure 8.74 The eligible population includes (1) the total UK population, even if currently geographically inaccessible; (2) patients with a confirmed infarct, excluding patients with unconfirmed status (≈ 2%); and (3) patients with basilar artery occlusions eligible for treatment. Patients in the large, lower blue-shaded box are selected by advanced imaging. It was originally estimated that between 9620 and 10,920 UK stroke patients would be eligible for treatment. The revised estimate based on new evidence was that an additional 490 early-presenting and 205 late-presenting patients would be eligible for IAT (i.e. between 10,140 and 11,530 patients). The largest increase was a result of new evidence of IAT benefit for patients with Alberta Stroke Program Early Computed Tomography Score (ASPECTS) of ≥ 5,80 with a smaller contribution from the treatment of patients presenting between 12 and 24 hours.77,78

FIGURE 8 Updated estimate of the number of stroke patients eligible for IAT. ASPECTS,79 Alberta Stroke Program Early Computed Tomography Score; CTP, computed tomography perfusion; CTACS, computed tomography angiography collateral scoring; DAWN,78 DWI or CTP Assessment with Clinical Mismatch in the Triage of Wake-Up and Late Presenting Strokes Undergoing Neutrointervention with Trevo; IVT, intravenous thrombolysis; MR, magnetic resonance; MT, mechanical thrombectomy. a, Total UK population, including those deemed to be geographically inaccessible; b, confirmed infarcts, excluding ≈ 2% of patients whose status is unconfirmed. Note that totals between levels in the tree may differ slightly due to decimal rounding. Reproduced with permission from McMeekin et al.74 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The figure includes minor additions and formatting changes to the original figure. (Figure not reproduced here; see the original article.)

Conclusion

Up to 12% (11,530/95,500) of UK stroke admissions are eligible for IAT based on treatment criteria from randomised clinical trials.

Helicopter Emergency Medical Services survey

While undertaking the survey of INR centres (see Baseline characteristics of intra-arterial thrombectomy service provision in England), it became apparent that some parts of the population live remotely from a hospital able to provide IAT. However, many hospitals are served by the air ambulance network, which could provide an approach to improve access to treatment. A survey of Helicopter Emergency Medical Services (HEMS) was undertaken to parameterise a health economic model for the cost-effectiveness of HEMS compared with ground-based ambulances providing secondary transfer of stroke patients for IAT.

Methods

An online survey was sent to the clinical leads of nine HEMS serving ‘unavoidably small and remote’ hospitals (NHS England definition: < 200,000 population and > 1 hour of travel from nearest major hospital)81 (Figure 9).

FIGURE 9 Unavoidably small and remote hospitals in England with HEMS. Reproduced with permission from Google Maps (Google Inc., Mountain View, CA, USA). Map data © 2021 Google GeoBasis-DE/BKG (© 2009). (Figure not reproduced here; see the original article.)

The survey gathered data on the number of helicopters, time in operation, number of hospitals served, average travel times by air, fit for purpose helipads, availability (09.00 to 17.00, 24/7 or other) and conditions under which flight is permitted.

Results

Responses were received from all nine HEMS (Table 5).

All HEMS were willing to provide secondary transfers for IAT. HEMS operated a median of 14 hours per day, with extensions to operational hours being planned in two HEMS. The median response time from notification to take-off was 4 minutes and the cost per mission was £2750 (range £2500–3500). Most HEMS (eight of the nine) were Instrument Flight Rules-approved82 for flying in low-visibility conditions. To deliver transfer for IAT robustly, three of the nine HEMS indicated additional funding and/or organisational changes would be required.

Microcosting study to establish true cost of intra-arterial thrombectomy

The full report has been published and parts of this section have been reproduced from this report.83 Republished with permission of Royal College of Physicians, from The cost of providing mechanical thrombectomy in the UK NHS: a micro-costing study, Balami JS, Coughlan D, White PM, McMeekin P, Flynn D, Roffe C, et al., volume 20, edition 3, 2020; permission conveyed through Copyright Clearance Center, Inc.

To estimate cost-effectiveness of national implementation, it was first necessary to establish the cost of providing IAT within the first 72 hours since stroke onset and to explore resource and costs variations across UK centres.

Methods

Microcosting methods were used to enable a precise assessment of the costs of IAT from an NHS perspective. Data were derived from the five nearest neuroscience centres from 2015 to 2018. The resources used and the costs were collected on patients admitted by secondary transfer (i.e. ‘drip and ship’ patients) and directly to centres (i.e. mothership patients).

Results

Data were abstracted directly from clinical records of 310 patients treated with IAT. The mean total per-patient cost of providing IAT and inpatient care within 24 hours of stroke onset was £10,846. The main driver of cost was IAT procedure costs, accounting for 73% (i.e. £7930) of the total 24-hour cost. Total mean cost within 72 hours of stroke onset was £12,440. Costs were higher for patients treated under general anaesthesia than for those treated under local anaesthesia, with a mean difference of £1070 (95% CI £381 to £1759; p = 0.003); the admission to an intensive care unit (ICU)/high-dependency unit (HDU) mean difference was £8210 (95% CI £4833 to £11,588; p < 0001). There was no statistically significant difference in 24-hour costs between ‘mothership’ and secondary ‘drip and ship’ patients: mean difference –£368 (95% CI –£1016 to £279; p = 0.26).

Conclusions

The major factors contributing to the costs of IAT for stroke include consumables and staff for the intervention, the use of general anaesthesia and admission of patients to ICU/HDU, and inter-hospital transport and repatriation. These findings may inform the reimbursement, provision and strategic planning of stroke services.

Estimating the effectiveness and cost-effectiveness of establishing intra-arterial thrombectomy: a discrete-event simulation

The full report for modelling study one has been published,84 as has the full report for modelling study two.85

Material throughout this section has been reproduced with permission from McMeekin et al.84 This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (https://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. The text below includes minor additions and formatting changes to the original text.

Material throughout this section has been reproduced with permission from Coughlan et al.85 This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: https://creativecommons.org/licenses/by/4.0/. The text below includes minor additions and formatting changes to the original text.

The evidence that IAT for acute ischaemic stroke is both a highly effective and cost-effective treatment is unequivocal. Less certain is how NHS providers might maximise the benefits of IAT and how services should be extended across the NHS beyond its currently low level of provision. To support decision-makers, we developed an economic modelling framework to estimate the health and financial consequences of alternative models of service provision. In keeping with best practice,86 the model included uncertainty and discounting, but capital expenditure and other costs associated with establishing new health-care infrastructure were not included.

Model structure

There were two stages incorporated in to the model: (1) short-term predictions of the mRS at 90 days based on characteristics of stroke patients at presentation and (2) long-term, lifetime projections based on 90-day mRS patient characteristics. Resource consequences (financial costs) and health outcomes were derived from morbidity (mRS) and mortality. The ‘code base’ of the discrete-event simulation (DES) originates from a previous Programme Grants for Applied Research (PGfAR) [Development and Assessment of Services for Hyper-acute Stroke (DASH); RP-PG-0606-1241], which involved building a model to examine the consequences of pre-hospital redirection of patients eligible for thrombolysis.87 It was extended to include the secondary transfer of patients eligible for mechanical thrombectomy.

The DES was iterated 2000 times for each person in the population who would be affected by the change in service configuration. The mean outcomes were aggregated to estimate the marginal effects before and after proposed service change.

Key assumptions

The model assumes that the patients included in our simulation had the same properties as those included in the HERMES analysis of time to treatment.23 Although we varied the age of patients within our simulation, this had an effect on post 90-day survival and deterioration only.

Data sources

Short-term outcomes (up to 270 minutes) for patients treated with IAT were defined by mRS using results of the HERMES meta-analysis,23 which identified the relationship between time to treatment and groin puncture.

Long-term mortality in the model was derived from the Oxford Vascular Study (OxVasc) study,88 data from the Lothian stroke register89 about the increased mortality associated with stroke survivors and repeated random draws from national life tables90 for patients alive at censor. Using a knotted spline regression technique to extrapolate survival in each simulation run generated a different set of extrapolated parametric survival curves for each iteration of the simulation, from which the time to death could be calculated from the uniform randomly drawn probability of death.84,91 To allow for improvements in survival since the Lothian Stroke Register88 data were collected, we applied a reduction in mortality of 25% from year 6 onwards.84,92 The simulation was implemented in the R statistical package (The R Foundation for Statistical Computing, Vienna, Austria).

Results

Two scenarios were modelled with the DES framework, reflecting current issues facing commissioners of IAT services in England. The first considered the effectiveness and cost-effectiveness of increasing current IAT provision from 24 to 30 centres, and the second considered the effectiveness and cost-effectiveness of secondary transfer of eligible patients from remote hospitals to current IAT centres.

Modelling study one: increasing the numbers of centres providing intra-arterial thrombectomy

To achieve the objectives of the National Commissioning Policy for IAT in a geographically equitable way would require the creation of new centres. Collaborating with colleagues from the National Institute for Health and Care Research (NIHR) Collaboration for Leadership in Applied Health Research and Care (CLAHRC) South West Peninsula, it had been previously ascertained that the optimal number of neuroscience/comprehensive stroke centres (CSCs) providing endovascular thrombectomy (IAT) in England would be 30 (net six new centres), subject to geographical and service level constraints.93 The DES was used to estimate the relative effectiveness and cost-effectiveness of increasing the number of centres from current provision at 24 centres to 30 centres. Using the parameter values (see Appendix 8, Table 18, the DES estimated effectiveness and lifetime cost-effectiveness (from a payer perspective) for 1 year of incidence of stroke in England.

Of the estimated 80,800 patients admitted to hospital in England with acute stroke per year, 21,740 were modelled to have their acute care affected by the proposed reconfiguration. The median time to treatment for eligible early presenters (< 270 minutes since onset) would reduce from a median of 195 (IQR 155–249) minutes to 165 (IQR 105–224) minutes. The model predicted that the reconfiguration would mean an additional 33 independent patients (mRS 0–1), and 30 fewer dependent patients dying (mRS 3–6) per year. The addition of six centres generates 190 QALYs (95% CI –6 to 399 QALYs) and results in savings to the health-care system of £1,864,000 per year (95% CI –£1,204,000 to £5,017,000 per year). The estimated budget impact was a saving of £980,000 in year 1 and £7.07M across years 2–5. We concluded that changes in acute stroke service configuration would produce clinical and cost impacts that were highly likely to result in cost saving. Over 5 years, there would be a return on capital investment of £8M, but it is important to acknowledge that capital expenditure, workforce expansion and other infrastructure costs are not included and are likely to be substantial.

Modelling study two: using helicopters to transfer eligible patients from geographically remote hospitals

Although England faces fewer geographical challenges to the provision of acute and emergency care than other parts of the UK, there are populations for whom transfer times from their local hospital to a major centre may compromise care. NHS England defines an ‘unavoidably small and remote’ hospital as one serving a population of < 200,000 people who are domiciled > 60 minutes’ travel by road from the nearest (major acute) hospital.81 We aimed to estimate the marginal cost-effectiveness of secondary transfer by air ambulance compared with ground-based ambulances for patients attending such a hospital.

Ten hospitals in England were identified that served a combined population of two million, including 512,875 domiciled in remote locations. It was calculated that up to 501 early-presenting stroke patients per annum would benefit from secondary transfer to IAT using an air ambulance compared with using ground-based ambulances. The mean probability of living independently at 90 days increased when using an air ambulance (0.57) compared with using ground-based ambulances (0.53). Using an air ambulance as a secondary transportation strategy that enabled patients to receive IAT 60 minutes earlier resulted in greater QALYs (0.14) over a lifetime horizon, but was more costly (£3785). The incremental cost-effectiveness ratio was £28,027 per QALY gained.

Discussion

As with all models reporting ordinal outcome measures, the DES under-reports gains because small improvements as a result of speedier treatment are not counted. However, the DES was used successfully to model two important issues facing commissioners considering how to meet the goals of the National Commissioning Policy for IAT. Both scenarios modelled showed a small benefit from increasing provision and speeding up access to treatment.

Patient, carer and public preferences

It is important that public views are included during modelling and commissioning decisions for the provision of a new service. A survey approach established the preferences and trade-offs related to localised versus centralised IAT services from the perspective of stroke survivors, their relatives/carers and the public.

Methods

Best–worst scaling (BWS) is a preference elicitation technique where respondents select their ‘best’ (most preferred) and ‘worst’ (least preferred) items across a range of subsets. A co-design process involving stroke survivors and their relatives/carers established the form and content of the BWS survey (maximising readability and accessibility), informed by best evidence on presentation of probabilities (numerical and graphical)94 and design of information for people with aphasia.72 Online testing was conducted with stroke survivors/carers and researchers with expertise in choice experiments.

The BWS survey was hosted on the Stroke Association website, and a link to the survey was distributed to all Healthwatch services in England and in the June 2019 NHS ‘In Touch’ newsletter.

Respondents stated their best and worst preferred options for four attributes (with two levels each – derived from the survey in Patient and public preferences on attributes of intra-arterial thrombectomy service organisation) in part 1 (i.e. service organisation), and three attributes (with two levels each) in part 2 (i.e. modelled outcomes for 24 vs. 30 IAT centres). Examples are shown in Appendix 9, Figures 19 and 20. Individual respondents were required to answer 2 sets of 8 (total of 16) BWS questions. The BWS responses were transformed into standardised scores with 95% CIs [ranging from –1 to +1, which represented the salience (best or worst) of attribute levels].

Results

One hundred and five respondents fully completed the survey [mean age 37 (range 18–86) years; 70% female; 47% from north-east England; 18% urban, 56% suburban, 26% rural]. The respondent types were as follows: stroke survivors, 18% (10% with aphasia); relative or carer of a stroke survivor, 32%; and members of the public, 50%.

Standardised scores for service organisation (Figure 10) showed that experienced medical teams (rather than a local service) and local services with less experienced medical teams were the most and least preferred attributes, respectively. Secondary transfer (present or not) was associated with positive (best) preferences. Length of stay of < 48 hours was associated with negative (worst) preferences. Travel times of > 45 minutes were associated with negative preferences.

FIGURE 10 Standardised scores for service design attributes. Note that overlapping 95% CIs indicate that the difference is not significant. LOS, length of stay. (Figure not reproduced here; see the original article.)

Standardised scores for modelled outcomes (Figure 11) showed that maximal effectiveness (30 centres) and costs (both levels) were the most and least preferred attributes, respectively.

FIGURE 11 Standardised scores for outcomes. Note that overlapping 95% CIs indicate that the difference is not significant. MRI, magnetic resonance imaging. (Figure not reproduced here; see the original article.)

Equity was associated with positive preferences, but the differences between levels were marginal. There were no statistically significant differences between standardised scores stratified in terms of stroke survivor/relative status compared with the public for service organisation or outcomes.

Conclusions

Greater expertise/experience of medical teams (compared with local services with less experienced teams) was the most preferred service organisation attribute. Secondary transfers with travel times of up to 45 minutes are acceptable to stroke patients/carers and the public. Respondents preferred effectiveness over equity (and both were valued more than costs).

A commissioning decision support tool: the Interface for Thrombectomy Economic Modelling and outcomeS in stroke

To assist commissioners and other stakeholders in understanding the health, financial and equity (proximity to treatment) implications of alternative geographical locations of CSCs (where IAT is available) and primary stroke centres (thrombolysis but no IAT available), a web platform was created using the model described in Estimating the effectiveness and cost-effectiveness of establishing intra-arterial thrombectomy: a discrete-event simulation.

Methods

The Interface for Thrombectomy Economic Modelling and outcomeS in stroke is a web-based application based on the R statistical package. Like the DES, the ITEMS software repeatedly models individual patient lifetimes based on random draws from the probabilities of relevant events occurring. The conceptual process for running the DES and saving the outputs is shown in Appendix 10, Figure 21.

The input is by a graphical interface, which restricts the parameters that can be altered in any simulation, but allows the user to try a large combination of values for the main factors determining service configuration and performance (Figure 12). Owing to expert and public preferences for secondary transportation to receive IAT if appropriate, ITEMS allows users to specify the initial distribution of the population sampled from ‘urban’, ‘suburban’ or ‘rural’, and considers uncertainty around inputs resulting from changes in journey times.

FIGURE 12 The ITEMS online user interface. (Figure not reproduced here; see the original article.)

The output displays a cost-effectiveness plane, and estimates marginal costs and health outcomes (either as individual mRS scores at 90 days or as lifetime QALYs). In addition to the DES described previously, ITEMS presents outputs that describe how individual patients might be affected by mean changes in outcomes on a cost-effectiveness plane (Figure 13). By trading off complexity for performance, ITEMS enables commissioners and other stakeholders to undertake a rapid review of alternative regional IAT service configurations, including economic impact.

FIGURE 13 The ITEMS interface output display. (Figure not reproduced here; see the original article.)