Sunday, May 8, 2011

Rural EMS
Emergency Response Mapping

Statement of the Problem
                In the Emergency Medical Services community, time is always considered the most important factor in providing quality care and increasing survivability statistics. It has been accepted as general knowledge that increases in time to initial care, access to advanced life support, arrival at appropriate medical facilities and finally to advanced interventions are all key to positive outcomes to emergency situations. This time saving is always the top priority within a community’s Emergency Medical Service’s plans for improvements. When viewing rural area emergency systems, however, this time saving may be compromised because distances from Rescue Squads and Fire Departments to individual residences can be greater than 20 miles. Often, the ambulance being dispatched to a particular incident is not the closest unit to the scene. This increases response times, making the time factors discussed above nearly impossible to meet. This is the problem which needs evaluation and remediation.

Theoretical Framework
                Beginning in the early 1980’s, EMS providers began looking at the effects of response times on their patients. With advances in the 1970’s in Advanced Life Services, especially in cardiology and traumatic injury, the difference in outcomes became obvious for those who received the same treatments at different time intervals. One of the first EMS programs to evaluate time as a real variable in outcomes in emergency situations was in Seattle. The article by Jonathan Mayer discusses how data collected across the entire city showed certain areas had longer response times than other areas. It also showed that the supply and demand model of EMS did not necessarily apply. Areas that were assumed to be high volume areas still saw better outcomes because of quicker response times, despite being given a heavier volume. Those areas that were deemed not as busy and therefore given less resources (ambulances and fire trucks) saw worse outcomes even though there was a unit available to the area far more often, simply because the time to respond was greater due to the longer distance. (Mayer, 71) This geographically disaggregated response area was found to be less effective. Even when applying the outliers and variables to accommodate population density, income and health indices, the results were still statistically significant. (Mayer 73) This is the beginning of the theory for the changes proposed below.
                Along with the potential outcomes of individuals facing emergencies, local governments are tasked with finding a way to make providing services to the community more efficient and less costly. Ambulances cost well over $100k, and staffing full time employees with the proper training can cost another $200k a year for each ambulance in service. Since more local governments are feeling the economic strain of the current financial crisis, models and algorithms are entering the conversation that find ways to maximize efficiency and minimize cost. Ideas, especially the ones discussed by Paul Sorensen and Richard Church include shift hours and rotations, locations of resources, dispatching models, types of staffed ambulance (Basic vs. Advanced Life Support) and even the backfilling of available units into areas where units are currently on calls. (Sorensen, 145) All these issues play a part in determining how a system should adjust and change structures to meet the demands and individual natures of the community it serves.
                Providing emergency services in the rural setting complicates the equation. By expanding the distance a single unit has to travel, both to the call location as well as the hospital and back makes for extended transit times. As was seen in the Mayer article, times do play a critical role. At the same time, those distances equal increased costs in vehicle maintenance and gas consumption. Compound this with a rural community lacking the tax base or population to support the system fully and a community or local government has a potential crisis in waiting. To complicate things even more, most rural settings do not have the appropriate resources to train and hire their own staffs, so contracts have to be made with private companies, which often ends up costing the locality even more money. It becomes obvious that the localities in rural settings should make every effort to establish cost saving practices.  This becomes the framework for the proposed dialogue presented below.

Thesis Statement
When measuring response times in rural Emergency Medical Services, combining traditional county dispatch systems into a single system will provide for faster and more efficient response times than the current system, resulting in quicker care and better patient outcomes.

Measurement/Methodology
Factors making the feasibility of combining systems are the first issue that needs evaluation. In this proposed case study, the counties of Southampton, Surry and Sussex Virginia will be combined. These three counties are contiguous, connected by multiple roads. They are also currently employing contractual Emergency Medical Technicians from the same company, Medical Transport, Inc., a division of Sentara Enterprises, as their primary Emergency Medical Service providers. Because of the familiarity and continuity of business practices and Operational Medical Direction, the combination of these three counties EMS dispatch systems is more feasible than a region of the state in which no connections exist between county EMS systems.
To provide analysis of the current system, maps of the three counties will be provided. Overlayed onto these basic maps will be the current 1st due run areas of the 7 Rescue Squads servicing the 3 counties. A 1st due run area is the area in which a call received in the dispatch center would be given to a particular station. A network analysis tool, providing times to individual locations within the county will be presented. The analysis will be to measure the current times to locations throughout the three counties. The proposed maps will use the current network analysis of times to locations to redraw 1st due run areas, based on time and disregarding traditional county lines. The same data can be overlayed to present improvements in potential health outcomes, based on the time to critical care.

Expected Results
The expected results will include the discovery of locations within the current county lines that have prolonged response times, due simply to excessive distance from the scene. The proposed map will provide a comparison to show that by ignoring classic boundaries, such as county and municipal lines, a significant savings in time for responding to emergency situations can be made. This provides for better patient outcomes, and increased efficiency and cost saving for all three counties.

Report
                The maps did tend to show the expected results, although in a few places, there were some unexpected results. These will be discussed further below, as well as the methodological process of data collection, integration and analysis.
                The data was collected from two main sources. First, the road map, used to produce times of transit, was provided by ESRI US National Road Map Atlas. This map provided every road in the three counties, as well as their speed limits, which was necessary to establish the length of time it took to travel a certain distance on a certain road. The county maps were collected from the Virginia Department of Transportation. These map layers were what was intersected together to create the areas of interest for data integration and analysis.
                Integrating the data presented the greatest issue. Attribute data had to be added to the road map, to provide actual times, in minutes to travel a section of road. Once this process was completed, however, the network analysis tool in ArcGIS was able to build the appropriate network and set up service areas for all seven of the rescue squads combined, as well as each squad separately. The final step in the integration the data was to intersect the current run areas, which were digitalized from paper maps into ArcGIS, with the time lengths presented in each separate rescue squad run area. This created the current run area maps. The service area function in the network analysis provided the proposed run areas.
                The analysis of the data showed certain trends. There is, in fact, no doubt that the current run areas of rescue squads are not providing the best and quickest care to the citizens of the three counties. Places like Waverly are spending 30 or more minutes driving to locations less than 20 minutes from Stony Creek and Capron. Surry is driving 30 plus minutes to locations within 10 minutes of Waverly. Surprisingly, however, areas of southern Southampton are very well covered, with almost none of the locations between Capron, Courtland and Boykins having more than a 20 minute response time. The area of most concern is by far the area in the middle of the triangle drawn between Capron, Stony Creek and Waverly. This area, seen in red on Map 2, is currently being covered by Waverly. It is interesting to note, that except for Surry, all of the other stations are actually closer than Waverly. It is simply because of the county line of Sussex and Southampton, added to the traditional lines between Stony Creek and Waverly that continues to leave the area underserved and far less than protected. Simply stated, as it is currently drawn, the citizens in this area are basically guaranteed to have worse outcomes in emergency situations than if a different rescue squad responded to their location. This should be changed, and the proposed Map 3 is the best start.
                Understandably, the analysis of time alone cannot be the only deciding factor in changing response models. A deeper study of population densities, call volumes, income and poverty levels, health risk identifiers and other data have to be placed into the conversation. Regression analyses would be the best potential determinant of which combination of variables would be ideal to display graphically on any future analysis of changing response areas and potentially combining EMS systems across county lines. There is also the issue of establishing a dispatch center capable of taking all of the calls, dispatching the appropriate units from the appropriate counties (EMS, Fire and Police). There are significant startup costs associated with combining systems. Finding the best way to pay for the system is complex as well. County tax rates and fees would have to be changed to meet the requirements, political actors would want to t the best deal for their citizens and conflict would arise, resulting in stagnation. It would be no easily established system.

Conclusion
                The reallocation of resources is not a new trend. It has been a studied part of EMS for over 30 years. Providing system analyses is the best way to decide what new squads are needed as well as where they should be located. If this analysis is apparently the best way to determine if large metropolitan areas provide for their needs, the same should be done for rural systems.  Given standards of care and treatment, along with consistent training within the EMS community, response times to the scene of an emergency is the single most important factor in providing quality care and improving outcomes for patients. The opportunity to improve response, without having to add units and increasing costs, should not be ignored.


References

Mayer, Jonathan D. 1981. "A Method for the Geographical Evaluation of Emergency Medical Service
Performance." American Journal of Public Health 71, no. 8: 841. Academic Search Complete,
EBSCOhost (accessed February 8, 2011).

Sorensen, Paul, and Richard Church. 2010. "Integrating expected coverage and local reliability for
emergency medical services location problems." Socio-Economic Planning Sciences 44, no. 1: 8-
18. Academic Search Complete, EBSCOhost(accessed February 8, 2011).

No comments:

Post a Comment