- Research and Projects
- Get Involved
Following PenCHORD's presentation of the report from this work, Professor Sir Roger Boyle CBE, former National Director for Heart Disease and Stroke at the Department of Health, said:
"The modelling exercise undertaken by PenCHORD at Exeter University Medical School exploring the possible configuration of acute stroke services in Wales was a difficult task given the scattered nature of the Welsh population. However the report demonstrated very clearly how the difficult issues of travel time could be addressed in order to optimise outcomes. In my opinion, the PenCHORD modelling team are the best that we have in this country. Well done."
The Royal College of Physicians of London (RCP) and PenCHORD were tasked by the Stroke Implementation Group (SIG) of the Welsh Government to coordinate the development of service maps and profiles for hyperacute and acute stroke services across Wales. The aim of this work was to provide evidence to support discussions regarding service reconfiguration across the country.
Factors considered in the modelling of the benefits and risks of reconfiguration included:
Maximising the clinical benefit to the people of Wales from the delivery of high-quality hyperacute and acute stroke services, including but not confined to the delivery of thrombolysis and mechanical thrombectomy;
The provision of a sustainable senior medical workforce to support the delivery of these hyperacute and acute stroke services;
The impact of reconfiguration on co-dependent services, both those that rely on co-location with stroke services (particularly cardiac services and vascular surgery) and those services that rely on a medical workforce that covers both stroke and other specialties (e.g. geriatric medicine and acute general medical on-call rotas);
The area of Wales and the Health Boards (HBs) within its boundary but also considering the potential impact of any reconfiguration on bordering acute stroke services in England and of any reconfiguration in England that may impact on Welsh residents;
All patients who live within Wales and those patients whose closest acute/general hospital is within the country.
Stroke services were modelled to incorporate the Welsh population, whilst taking full account of boundary issues. The model included hyperacute and acute stroke services, accounting for essential co-dependencies such as vascular surgery, cardiac services and interventional radiology.
The modelling included the following options:
The status quo;
Configurations meeting best/expert practice, e.g. RCP guidelines*;
Co-location of complex cardiac, vascular and stroke services;
Consideration of co-dependency with neuroradiology and possibly neurosurgery;
The requirement for clot retrieval and its provision;
A range of intermediate geographical configurations up to and including the status quo.
|*RCP Stroke Guidelines|
The model was populated using data from the largest time period available, obtained from (i) the Patient Episode Database for Wales (number of patients with primary diagnosis of stroke, along with home location); (ii) data from relevant national audits (SSNAP and other ‘Intelligent Targets’ data, e.g. thrombolysis treatment times (onset-to-treatment, door-to-treatment) for hyperacute stroke), and; (iii) pre-hospital data from the Welsh Ambulance Services NHS Trust (12 month ambulance data for patients attending with suspected stroke, identified either by phone triage or paramedic).
The modelling was evidence based, with results interpreted and supported by an expert panel drawn up by the RCP Stroke Programme. The panel's objective was to prioritise and balance the competing factors that affect reconfiguration, especially where best practice guidance cannot be explicitly met, as well as the production of an options appraisal for the SIG.
For the selection of location of acute stroke units, the model reported on and looked for solutions which:
Minimise average transport time from home to acute stroke unit;
Minimise the maximum transport time (for any single patient) from home to acute stroke unit;
Maximise the number of patients within 45 minutes transport time;
Maximise the estimated clinical benefit of thrombolysis (this will be based on estimated onset-to-ambulance times, modelled transfer times, and estimated arrival-to-thrombolysis times). For this analysis it will be assumed that all hospitals achieve the same arrival-to-thrombolysis times and all will have a standard thrombolysis rate;
Maximise the number of patients attending a unit that receives at least 600 stroke patients per year.
Also reported in the model were:
The number of patients whose closest hospital is within the study region (Wales), and the number of patients who would leave the study region if they attend their closest hospital (which may be in England);
The number of admissions to each acute stroke hospital;
Expected total ambulance on-scene and travel times (to estimate any impact on the ambulance service).
The model reported the optimal solutions and near-optimal solutions for the selection of locations for any given number of hospitals. For option analysis, maps were generated that show expected transfer times for each geographical area and change in transfer times from current service configuration.
The project produced a range of options intended to provide the greatest health benefit from interventions for acute stroke to the maximum number of people in Wales. These options take account of anticipated changes in demography and disease incidence over the next ten years and also provided an options appraisal for the Welsh SIG, providers, HBs and other key stakeholders. These options were presented in a final report and at an Expert Panel meeting with stakeholders.
Allen M, Pearn K, Monks T, Bray BD, Everson R, Salmon A, James M, Stein K. Can clinical audits be enhanced by pathway simulation and machine learning? An example from the acute stroke pathway BMJ Open 2019;9:e028296.
The Royal College of Physicians (London)
Dr Martin James – Lead Clinician for Stroke, Royal Devon & Exeter NHS Foundation Trust