Search

Background: A trauma system is a clinical network for the delivery of specialist trauma care. Such networks comprise designated trauma centres, with stratified capability, supported by emergency medical services, and aim to match patients' needs with hospitals' resources. Treatment within a trauma system has been shown to be associated with improved survival and functional outcomes. However, the configuration of trauma systems, in terms of the designation of trauma centres, is often contentious. Aim: Scotland is in the process of establishing a trauma system. The aim of this thesis was primarily to determine the feasibility and delivery of a geospatially optimised trauma system configuration for Scotland, and then to test the transferability of the methodology to other settings. Objectives: The objectives were: 1) To establish a mechanism for the prehospital triage of trauma patients 2) To characterise Scotland's geospatial injury profile 3) To determine geospatially optimised system configurations 4) To determine the transferability of the method to other settings. Establishing a mechanism for the prehospital triage of trauma patients: Prehospital triage and tasking are critical to the functioning of a trauma network, and the objective of this part of the study was to establish a mechanism for the prehospital triage of trauma patients, in order to evaluate prehospital decision making and patient flow. Scottish Ambulance Service paramedics were asked to complete the Field Triage Decision Scheme, which was added to the vehicles' onboard computers, for all injured patients whom they had attended. The Field Triage Decision Scheme is a widely used triage protocol, which sequentially evaluates physiological status, critical injuries, mechanism of injury, and special considerations. In total, 80 391 patients attended to by the Scottish Ambulance Service, between 1 July 2013 and 30 July 2014, were included in the study, and notionally triaged. Patients who met the criteria for physiological abnormalities, or had critical injuries, were notionally triaged to major trauma centre care (9%). Patients who met the criteria for a significant mechanism of injury, or special considerations, were notionally triaged to trauma unit care (42%). All remaining patients were notionally triaged to local emergency hospital care (49%). This component of the study demonstrated the feasibility of introducing a national mechanism for the triage of trauma patients by prehospital care providers, and that such a triage, based on Field Triage Decision Scheme, can separate patients into three distinct groups. Scotland's geospatial injury profile: The geographical distribution of injured populations is of key importance to planning services because it is a prime determinant of accessibility. The objective of this part of the study was to characterise Scotland's geospatial injury profile. The analysis was conducted using a combination of Kernel density estimate mapping and hierarchical cluster and nearest neighbour analysis. Scotland has an identifiable geospatial injury profile. The incident distribution was highly clustered. Most injuries occured in the Central Belt of Scotland, with a lesser number along the East coast, broadly following the distribution of the general population, with some variation between triage categories. Injury was a rare event in many remote and rural parts of Scotland. Performance of the current system of care: Characterising the performance of the current system of care is important to understanding the need for change. The objectives of this analysis were to evaluate the performance of the current system, in terms of the types of hospitals which patients in each of the triage categories were taken to. The analysis was conducted by evaluating the level of the hospitals which patients were actually taken to, compared with the triage category which they had been notionally allocated to. As Scotland does not have a trauma system yet – and therefore has not yet designated hospitals as major trauma centres, trauma units, or local emergency hospitals – hospitals were classified as university hospitals (some of which could become major trauma centres), large district general hospitals, general hospitals, and other facilities (including community hospitals, children's hospitals, and maternity hospitals). 53% of injured patients were admitted to a large general hospital, 39% to a university hospital, 5% to a general hospital, and the remainder (3%) to another type of facility. Subgroup analyses by triage category showed similar results. The proportion of patients admitted to each type of hospital varied with the rurality of the incident location. This analysis showed that most trauma patients in Scotland were taken to a large general hospital, or a university hospital, irrespective of the severity of their injuries (as determined by triage). Although this policy ensures rapid access to a hospital, it does not necessarily ensure access to specialist trauma care. Configuration modelling and multi-objective optimisation: Optimising the geographical configuration of trauma systems has the potential to benefit both patients and institutions, by facilitating prompt access to specialist care, while also promoting the efficient use of resources. A systematic approach to trauma system design could furthermore help to reassure stakeholders that the best configuration has been chosen. The objective of this part of the study was to perform a combined network analysis and multi-objective optimisation of all mathematically possible trauma system configurations, in order to determine geospatially optimised system configurations. Drive-times and flight-times from every incident location, to every hospital in Scotland which could become a major trauma centre or trauma unit, were calculated using geographical information system software. All mathematically possible trauma system configurations were then modelled, and evaluated using multi-objective optimisation, with the objectives of a) minimising total system travel time, and b) minimising the number of patients who could not be taken directly to the level care assigned by the triage process. The results were further constrained by setting limits for helicopter usage and minimum volumes of major trauma centres. The different constraints were sequentially analysed, reflecting priorities in system design, in an attempt to maximise the case volume of major trauma centres, using the smallest number of helicopters. Network analysis and multi-objective optimisation revealed that, if a network based on high volume (>650 severely injured patients per year) major trauma centres was to be developed, there would be 21 possible "optimised" configurations. Nine of these configurations would require a fleet of five helicopters, the remainder requiring six aircraft. All of these configurations would have a single major trauma centre, at the Queen Elizabeth University Hospital in Glasgow, as well as 10-17 trauma units (for configurations relying on five aircraft). If a network based on low volume (>240 severely injured patient per year) was deemed acceptable, 11 Pareto-optimised configurations were observed. Two of these could be realised with four helicopters, whereas the remainder would require at least five aircraft. All of the configurations with four helicopters would require two major trauma centres, in Glasgow and Edinburgh, and 15 or 16 trauma units. This analysis demonstrated the feasibility of mathematically modelling complex care systems, and identifying optimally performing configurations. The results indicated that establishing a trauma network for Scotland is feasible, and that such a network – depending on the major trauma centre volume thresholds chosen – could be optimised with a single major trauma centre, in Glasgow, or two major trauma centres, in Glasgow and Edinburgh. Discussion: This research demonstrated that network analysis and multi-objective optimisation, in conjunction with a population-based notional triage, can be used to inform the planning of a major national care system. Scotland's trauma network would be optimally configured with one major trauma centre in Glasgow, or two major trauma centres, in Glasgow and Edinburgh. These results are explained, in a great part, by the spatial distribution of the incidents. The methodology described here is not only applicable to Scotland, as demonstrated by the analysis of the Colorado trauma system. It is also not only applicable to trauma care. Whether explicitly considered or not, there is a geographic dimension to the design of almost any clinical network, and particularly those caring for patients with highly acute conditions, which need to balance accessibility and specialist care.