Hospital Surge Capacity Management through Simulation Mr Nick Howden; Dr Julie Trpkovski and Mr Mark Grebler CAE Professional Services nickh@cae.com.au; Julie.Trpkovski@ontario.ca; markg@cae.com.au Abstract. Canada s battle with SARS revealed significant weaknesses in the Ontario healthcare system, including a limited ability to manage critical care resources across hospitals in response to a sudden spike in demand. In response to this, Ontario is running a new surge management program to help hospitals better manage spikes in demand for critical care services without affecting day to day hospital services. As part of this program, Ontario has engaged CAE to build a simulation capability to exercise surge management plans within and between hospitals. The main objective of this project is to test and exercise implementation of a principled approach to manage surge capacity and leverage critical care resources across the hospital network to ensure patients have access to care. Through participation in this program, each participating hospital will strengthen communication, improve partnerships and ensure access to critical care resources in a timely manner. In order to effectively exercise surge management plans within and between hospitals, CAE is building a simulation environment that will provide the capability to run through a range of surge scenarios at the minor, moderate and major levels. The simulation capability will be able to create exercise scenarios based around a wide range of surge events, from disease outbreaks to mass casualty events and natural disasters. The system will also support testing and analysis of hospital processes and potential future changes and enhancements. This paper describes the development and application of the simulation system. 1. INTRODUCTION In 2003, Canada s battle with SARS revealed significant weaknesses in the Ontario healthcare system, including a limited ability to manage critical care resources across hospitals in response to a sudden spike in demand. As a result, Ontario MOH launched the Surge Management Program to help Ontario hospitals better manage spikes in demand for critical care services without affecting other hospital services. As part of this Program, Ontario has commissioned CAE to produce a Hospital Surge Management Trainer (HSMT). This trainer will provide modelling and simulation capabilities that will enable visualisation of key impacts on hospital resources while different elements of patient flow are challenged. Currently, CAE is developing hospital simulation system with two distinct purposes in mind: 1. The Training Tool allows hospital personnel the opportunity to practice and rehearse surge capacity management procedures according to standardised checklists and procedures. This tool is being developed with the Ontario Ministry of Health (MOH). In the future, with a with the fully developed hospital simulator described in this paper it will be possible to: Create and test evidence-based health policy and processes. By pre-testing in simulation, policies can be confidently implemented, knowing that it will improve KPIs for hospitals and across the local hospital network. At the government level, test what the relative effectiveness is for funding within and across networks. This will help answer the question of how budgets should be allocated. Providing training and decision support for patient flow and surge scenarios through hospital modelling and simulation. Evaluate different strategies for dealing with sudden surges of patients, and then allow hospital managers to rehearse these new strategies before the emergencies occur. Discover inefficiencies and bottlenecks with the general day-to-day hospital process. Feed live hospital data into the simulation to provide an operational decision support tool. 2. The Planning Tool will be used as an operational analyses and planning tool. Currently CAE is approaching hospitals and health decision makers in Victoria to try to extend the Training Tool to be used as this Planning Tool.
2. BACKGROUND The goal for the surge capacity program is to ensure organisations have the methodology to manage surge events at all of the following levels of responses: minor if they can be managed within a single hospital; moderate if they require the cooperation of several hospitals across a Local Hospital Integrated Network (LHIN); and major if the response requires the combined critical care resources of several LHINs or the entire province. The key to this process is developing management behaviours that are utilized on a daily basis to manage minor surges, so that when crisis comes in the form of a larger surge, each organization is prepared to manage the event. Quality indicators aren t enough to reflect preparedness the true test comes from a live event or a simulated training scenario. Putting surge plans in place without training and testing is not effective; for behaviour to change most effectively, people must experience the plans and problems first hand, ideally in a simulated scenario. This process ensures a low risk environment for evaluation with limited risk to patients, staff and organizations. It further provides opportunity to evaluate and concentrate on process improvements, and to create systems that will manage surge events whether they occur at the hospital or LHIN levels. Lastly it allows for quality and patient safety to be maintained. Figure 1: Training Tool Architecture With this design, it is simple to remove or replace the HTML server or even the database with a different one to provide completely different functionality with minimal changes to the actual simulation functionality. 3.2 The Simulation Process Model At a high level, the hospital is modelled as a simple process flow. Patients arrive at a hospital, and progress through the various Hospital Units depending on the patient condition. The diagram below shows the current configuration of Hospital Units and possible patient flows through the hospital. The infrastructure automatically adjusts to account for any modification to the links between the Hospital Units, or even the addition or removal of the units without the need to change any of the functionality. This provides the ability to answer such questions as: what would happen if a second ward was added to the hospital?. 3. THE TRAINING TOOL ARCHITECTURE With the HSMT Training Tool, multiple players can participate in a surge training scenario. They log into a simple web page, are assigned particular roles in the scenario by the Exercise Manager and once the simulation starts, are able to control the allocations of resources in response to the scenario. 3.1 The Technical Architecture The architecture that facilitates this consists of three main components: 1. A simulation engine which contains instances of different hospital process models (left box in diagram below) 2. A HTML web-server running Django which communicates with the users web-browsers (the right box in the diagram below) 3. A database facilitates the communication between the simulation and the web-server. All of the output data from the simulation, and commands from the users to the simulation are stored in this database. In addition to communication, this database also enables easy analysis once the simulation is complete. Figure 2: The hospital process model Each Hospital Unit is defined by a more detailed process model, which contains a procedure which requires resources (personnel, equipment and space). If there are resources available, the patient can progress through the Hospital Unit, otherwise they need to wait until the resources are available. The diagram below shows some of the detail in the process model of each Hospital Unit. Figure 3: The Hospital Unit process model
3.3 Example Uses The HSMT tool can be used for a range of training scenarios that can range from the minor hospital surges to major state or province wide disasters. The goal is to develop behaviours that are utilized on a daily basis to manage minor surges, to ensure when crisis comes in the form of a moderate or major surge, each organization is prepared to manage the event. For minor surges, various hospital unit managers (for example, the Emergency Room or Operating Room managers) could log in individually, and respond to a pre-configured set of events and see what effects the resourcing decisions they made had on bottlenecks for the other hospital units. For a major surge scenario, hospital managers and the LHIN manager can test and refine their communication and management procedures. 3.4 The Player s Interface 3.4.1 Dashboard The players in the scenario are given a fairly simple view of the hospital in the form of a dashboard. It is designed to show them the statistics that they would have access to in the actual hospital. The diagram below shows what the dashboard may look like. Figure 4: A sample dashboard Using the dashboard, players can see when their resources are becoming overloaded, and request transfers from other Hospital Units or other hospitals to correct the situation. The dashboard then enables them to see the effects of their request on other Hospital Units as well as their own. 3.4.2 The Floor-plan It is also possible to allow further visualisation if extra emersion is required. In the future, a floor-plan view can be enabled which will show an alternate representation of the resource usage of the hospital. Figure 5: A sample floor-plan 3.5 Deployment The framework allows the users to interact with the system through any web-browser. This means that a simulation scenario can be executed without any installation required on the client machines, or any
other IT infrastructure setup. The users can also be distributed anywhere around the world, as long as they have an internet connection. 4. THE PLANNING TOOL ARCHITECTURE With the Planning Tool, users can perform different types of studies from Monte Carlo and sensitivity experimentation, to activity base costing analysis, to capability studies to determine the what equipment to acquire. 4.1 The Technical Architecture The architecture for the Planning Tool utilises the development done for the Training Tool. Here, the process model described in section 3.2 communicates through a messaging service directly to a set of analysis tools that run within the simulation instance. Figure 7: Analysis Graphs Similar to the Training Tool, it is possible to perform run-time analysis using the Planning Tool. The user can experiment with the resource allocation and see the resulting effects on the waiting times. Figure 6: The Planning Tool Architecture A single user runs the simulations and is able to change variables and modify the process model to determine inefficiencies with the day-to-day running of the hospital. Using Monte-Carlo and sensitivity analysis, it is possible to determine what resources are the most constrained by the hospital, and as a result, where the hospital should be spending its money to increase patient throughput and decrease waiting times. The statistics generated can also provide robust analytical backing and visualisation to funding bids. 4.2 Analysis Tools Using this architecture, it is very simple to modify the data and resulting graphs that are produced by the simulation. The diagrams below show an example of what can be produced. 5. OTHER POTENTIAL APPLICATION AREAS As the trainees become accustomed to the using the Training Tool, they will learn how to interpret the data available in the dashboard (described in 3.4.1) and what decisions to make to correct any resourcing issues they are faced with. Given that familiarity with the dashboard, the next logical step is to feed in live data into the dashboard so the hospital staff will be able to apply what they learnt during the simulated training directly to real world events. The technical architecture of the Training Tool is compartmentalised enough that it is very simple to change to tool so that instead of having the HTML
server stimulated by a simulation, the real world data would be sent to the server to be seen by the hospital staff. 6. CONCLUSION Currently the Australian Government is launching a reform to tighten up the public healthcare system to reduce patient waiting times. The most cost-effective and efficient ways determining the best way of achieving this goal is through simulation. The tool described in this document will provide a way to create and test evidence-based health policy and processes. By pre-testing in simulation, policies can confidently be implemented and processes changed, knowing that it will improve KPIs for hospitals, across local networks and even for the entire state. 7. ACKNOWLEDGEMENTS The following people were involved in the development of the Hospital Surge Management Trainer: Nima Bahramifarid Jack Gao Mark Grebler Evan Harris Nick Howden Courtney Kersten Patrick Lachance Edward Robinson Andrea Scipione Julie Trpkovski (Ontario Ministry of Health) Dan Zikovitz