Secondary Logo

Journal Logo


Location, location, location

Utilizing Needs-Based Assessment of Trauma Systems-2 in trauma system planning

Dooley, Jennings H. BS; Ozdenerol, Esra PhD; Sharpe, John P. MD, MS; Magnotti, Louis J. MD; Croce, Martin A. MD; Fischer, Peter E. MD, MS

Author Information
Journal of Trauma and Acute Care Surgery: January 2020 - Volume 88 - Issue 1 - p 94–100
doi: 10.1097/TA.0000000000002463
  • Free
  • Editor's Choice


In 2015, the American College of Surgeons Committee on Trauma created the Needs Based Assessment of Trauma Systems (NBATS) tool. The goal was to create a “practical tool, based on the data that is currently available, which can be used to assist regions currently struggling with this issue of new trauma center designation.” Furthermore, all parties that helped create the tool agreed that trauma center designation should be based on the need of the regional population. This original instrument assigned points based on population, EMS transport time, community support, and the number of severely injured patients discharged from nontrauma centers. The points were then adjusted based on the number of already existing (legacy) centers in the region. The final point total was used to determine the number of trauma centers needed.1

Original tests of the NBATS tool revealed multiple flaws. The tool was felt to underestimate the number of centers in urban areas but overestimate the number of centers in rural areas. Furthermore, the results could be widely different and difficult to apply depending on the trauma service area (TSA) and was felt to be too definitive.2–4 In response to these findings, the American College of Surgeons Committee on Trauma revised NBATS to NBATS-2 with the goal to allow for a system to analyze overall impact of additional trauma center designation on access and volume in specific geographic zones (Table 1).

Version 2 of the NBATS-2

Needs Based Assessment of Trauma Systems-2 was recently applied to the urban area surrounding Rochester, NY within a 50-mile radius. The authors used NBATS-2 to assess the impact of adding a trauma center on access to care and on patient volume at each facility.5 The purpose of our study was to expand on the original NBATS-2 application and use NBATS-2 in a large trauma system consisting of both urban and rural areas. Access and volume determinations were made to the existing legacy trauma center with the addition of either a suburban center or two more rural centers. Payer mix of the patients was also analyzed to evaluate potential financial implications.


The Elvis Presley Trauma Center (EPTC) in Memphis, TN is the highest level trauma center for three trauma service areas (West Tennessee, Delta Region of Mississippi, Northeast Arkansas). The three regions consist of a total of 54 counties (425 zip codes) and were chosen as the TSA. Boundary data by county and zip code (ZCTA) were found as shapefiles in the US Census Bureau Database.6 Regional hospitals within the TSA that have facility and subspecialty support to potentially become Level I/II trauma centers were chosen as candidate hospitals. This determination was based on local knowledge. Location data was collected from ArcGIS Online.7 Population data by ZCTA were also collected from the US Census. Injury data were collected from the local trauma registry (NTRACS V5, Digital Innovations) over the 2-year period of 2016 and 2017. Variables collected included demographics, injury location, direct versus transfer, Injury Severity Score (ISS), and payer status. Only patients with ISS greater than 15 were used in analysis.

The TSA was mapped as county and ZCTA polygon shapefiles using ArcMap 10.5. To minimize map distortion, a customized Lambert Conformal Conic WGS 1984 projected coordinate system centered around West Tennessee was used. Three candidate hospitals and EPTC were mapped as points. Injury locations and population data were mapped as ZCTA polygon centroids. For transfers, the hospital from which the patient arrived was recorded as the location of injury.

The ArcMap Service Areas tool was used to calculate 45-minute drive time zones around each of the four hospitals. Three models were created to compare injury distributions and demographic data for candidate hospitals. The Closest Facility tool, which uses ArcGIS Online roads data to measure drive time between incidents and facilities and determines which are nearest to one another, was used in each of the three models to determine the nearest candidate hospital to each injury location (ZCTA centroid). Model 1 includes the EPTC and injuries within a 45-minute drive time of that facility. Model 2 includes EPTC and an additional suburban center in Memphis, TN, and injuries within a 45-minute drive time of each facility. Model 3 includes the EPTC and two rural candidate trauma centers located in Jonesboro, AR and Jackson, TN, and injuries within a 45-minute drive time of each facility.

Statistical analysis of the population within a 45-minute drive and the distribution of injured patients based on closest facility were analyzed in SAS version 9.5 utilizing a McNemar's test for paired populations. This study was approved by the University of Tennessee Health Science Center Institutional Review Board.


Over the two-year time period, 15,795 patients out of the regional population of over 2.5 million were injured. There were 1,795 (11%) patients that met the inclusion criteria. The median ISS was 22, and the mean age was 44 years of age. There were 1,312 (73%) patients transported directly to EPTC, and 483 (27%) patients arrived via transfers from nearby facilities. The majority of injured patients (66%) arrived from locations in TN, but 17% came from AR and 17% from MS. Forty-one percent of the patients went to the intensive care unit after arrival, and 31% went to the OR. Mortality rate was 18% (Table 2).

Characteristics of the Study Population, Which Included 1,795 Injuries From a Region of 54 Counties and 425 Zip Codes

In model 1, where EPTC is the only facility, 48% of the total population and 58% of the injured population were within a 45-minute drive time of legacy trauma center. The self-pay rate was 25% (Fig. 1, Table 3).

Figure 1
Figure 1:
Model 1 - Injuries and population within 45-minute drive time of EPTC, Memphis TN.
Predicted Injury Distributions for Models 1, 2, and 3

Model 2 includes EPTC and an additional suburban trauma center (hospital 2). The injured and total population coverage increased by only 1% to a total of 49% and 59% respectively. This small increase was due to a 96% overlap in area served. The volume to the EPTC decreased by over 40%. Self-pay rate at EPTC increased slightly to 27% with a 25% self-pay at Hospital 2 (Fig. 2, Table 3).

Figure 2
Figure 2:
Model 2 - Injuries and population within 45-minute drive time with the addition of one suburban trauma center.

In model 3, with the addition of two rural candidate trauma centers (hospitals 3 and 4), both injured and total population coverage increased significantly to 71% and 62% (p < 0.001). The areas served by the individual hospitals overlapped by less than 1%. Self-pay rate at EPTC also increased significantly to 29% (p = 0.006), while self-pay at the rural candidate trauma centers was low at only 20% and 8% (Fig. 3, Table 3).

Figure 3
Figure 3:
Model 3 - Injuries and population within 45-minute drive time with the addition of two rural trauma centers.


The purpose of a trauma system is to improve patient outcomes via an organized approach to trauma care.8 This includes prehospital management, in-hospital care, and posthospital rehabilitation within a defined geographic region.9 Since the development of the first statewide trauma system in the United States in the 1970s, it has been well established that trauma systems significantly reduce morbidity and mortality of trauma patients.10 A systematic review of 40 years of trauma outcomes data concluded that the development of North American trauma systems reduced mortality by 15%, and most patients had no lasting physical limitations 1 year after injury.11 Although trauma systems are essential for improving patient outcomes, the lack of standardization of trauma systems in the United States has led to significant gaps in access to trauma care.10

In 2015, the American College of Surgeons Committee on Trauma developed the Needs Based Assessment of Trauma Systems Tool (NBATS-1) to begin addressing access. The tool assesses regional trauma care needs by assigning point values to six categories within a TSA: population, median transport time, community support for a trauma center, number of severely injured patients (ISS > 15) discharged from nontrauma center acute care facilities, number of Level I trauma centers, and number of severely injured patients (ISS > 15) seen in trauma centers already in the TSA. The resulting point value corresponds to the estimated number of trauma centers needed in the TSA.1 When applied to TSAs in the State of California and the trauma system in the State of Georgia, the tool recommended more trauma centers in rural areas where patients have longer transport times to reach the nearest trauma center.2–4 However, in both studies, the estimated number of trauma centers needed in urban areas was lower than the number of those that currently exist. This suggests several possibilities: the tool underestimates the number of trauma centers needed in densely populated areas, the tool overestimates the number of trauma centers needed in less densely populated areas, excess trauma centers exist in urban areas, or rural areas need more trauma centers. Without a gold standard trauma system for comparison or the ability to create predictive models with NBATS-1, it is difficult to interpret the results, and the tool's practicality is thus rather limited.

In 2018, a workgroup created version 2 of the NBATS tool (NBATS-2) to incorporate predictive modeling of trauma systems based on Geographic Info Science (GIS) methodology (Table 1). Predictive modeling of trauma systems is essential for effective trauma system planning. Simply increasing the number of trauma centers within a TSA does not guarantee that more patients receive better care, as evidenced by the increased trauma mortality rate within a community in the Bronx, New York after the designation of a new trauma center in 2001.12 With the addition of a new trauma center, the existing trauma center will experience decreased patient volume and changes to its payer mix, both of which have significant associations with patient outcome.13,14 GIS-based methodology is what makes predictive modeling possible in trauma system planning. It enables objective assessment of initial geographic trauma risk and accessibility of trauma care within a TSA.5,15 By aggregating demographic data of injured patients with geographic data, GIS methodology can be used to predict population coverage, trauma center volume, and payer mix after the addition of a trauma center. In our study, we demonstrated the addition of a suburban trauma center had little impact on access to the trauma centers while negatively impacting the volume of the legacy center. However, the addition of two more rural centers significantly increased access. This ability to customize your system based on geography is an advantage of NBATS-2.

Based on local injury epidemiologic data, NBATS-2 can estimate volume to the new centers and any potential volume decrease in existing centers. Volume remains a key component in sustaining both clinical excellence and financial stability. A retrospective study of patient records from the National Trauma Data Bank showed that increased patient volume at trauma centers positively correlates with improved mortality outcomes.13 For 31 Level I and Level II trauma centers participating in a trauma benchmark study, high-volume centers had better mortality rates and shorter lengths of stay than low-volume centers for patients at high risk of adverse outcomes.16 This may be due to increased provider competency and resident physician training when treating more patients, leading to better quality of care. Based on this data, if the addition of a new trauma center results in a great enough decrease in patient volume at a legacy trauma center, the community which the trauma system is supposed to serve will be adversely affected. However, when considering the other end of the spectrum, it is intuitive that a trauma center whose patient volume frequently exceeds its capacity would also result in worse patient outcomes. If there are not enough resources at a facility to meet patient needs, the quality of care will be worse than that of a facility where resources and provider skills align with patient volume and need. It therefore follows that there is an ideal range for trauma center patient volume that correlates with improved patient outcomes. That “ideal” number of patients varies locally within any trauma system. The value of NBATS-2 is its ability to predict how adding a trauma center to the system will affect patient volumes at existing centers, which may then be interpreted by local entities to help them determine if an additional center would be detrimental or advantageous to the trauma patient.

A hospital's payer mix is also a key component that influences patient outcomes. An analysis of 265 California hospitals showed that centers with increased percentages of private insurers have better quality of services and patient outcomes. Centers with better overall “financial health” invest more in hospital infrastructure and equipment, and they have higher rates of clinical adherence with fewer annual deaths and readmissions.14 Therefore, if the addition of a new trauma center results in decreased rates of privately insured patients at an existing trauma center, the financial health of the hospital will decrease, and patient outcomes may worsen. The NBATS-2 tool in our study demonstrated the addition of the new trauma centers not only decrease volume of patients to the legacy center but the percentage of those patients who are self-pay would increase assuming a constant distribution of payer mix in the region.

In our study, we used NBATS-2 to create three predictive models of candidate trauma centers within the defined TSA. We expanded on a previous application of NBATS-2 to an urban area surrounding Rochester, NY, by including both rural and urban regions within the defined TSA and by incorporating payer mix into our models, in addition to population coverage, travel time, and trauma center patient volume. In model 2, the self-pay rate at EPTC only slightly increases, but there is a greater than 40% decrease in patient volume (ISS > 15). Such a dramatic drop in volume at EPTC may negatively affect provider competency, resulting in worse patient outcomes. Similarly, despite the significant increase in population coverage in model 3, patient volumes for the two new centers are only 270 and 172 over a 2-year period. Volumes this low may result in poor patient outcomes at the rural centers. Also, in model 3, the 16% increase in self-pay rate at EPTC with the addition of two rural candidate trauma centers represents a significant change in payer mix, but the overall number of patients paying out of pocket decreases. Although increased self-pay rate and decreased “financial health” of a hospital correlate with worse patient outcomes, the effect on patient care in this case is unclear.

While the results of our study are speculations that have not been proven in the real world, our study shows the significant potential of using NBATS-2 and geospatial modeling in trauma system planning. Needs Based Assessment of Trauma Systems-1 was an important first step in defining and establishing objective criteria for a trauma system that meets its regional needs, but the ability to create predictive models using GIS-based methodology makes NBATS-2 a much more practical tool than the original. The efficacy of our study and of NBATS-2 for trauma system planning depends on (1) accurate geospatial modeling via ArcGIS, and (2) evidence-based correlations between patient volume, payer mix, and patient outcomes. Future studies should assess the accuracy of ArcGIS modeling and NBATS-2 by comparing GIS-based predictions with real world results in regions where new trauma centers have been added. Future studies should also use GIS technology to consider the impact of air transport in trauma system planning. Due to the complexities of air transport in the mid-south area, we did not include air transport in this evaluation of NBATS-2, but it is an important aspect of patient access to care and should be considered for future models. We intentionally did not include it in this specific analysis due to the complexities of aeromedical transport that exist in Tennessee. As an example, in our region there are 26 emergency response helicopters operated by seven different companies in this catchment area alone, and the geographic regions covered by each of these helicopter response teams do not always align with patients being flown to the nearest trauma center. While we certainly plan on analyzing aeromedical transport as a component of trauma system planning in future studies, we did not feel that this was appropriate for our first large-scale trial of the NBATS-2 tool. Additionally, more studies should be done to better define the risk-benefit relationships between patient volume, payer mix, and patient outcome.


Version 2 of the NBATS tool adds a new dimension to trauma system planning within a region by utilizing geospatial modeling. Needs Based Assessment of Trauma Systems-2 demonstrates how geospatial modeling in a practical easy-to-use tool can be used to assess changes in population and injury coverage, as well as potential volume and financial implications to a current trauma system.


All authors contributed equally. J.H.D. and P.E.F. had full access to all of the data and take responsibility for the integrity of the data and the accuracy of the data analysis.


The authors declare no funding or conflicts of interest.


1. American College of Surgeons. Needs Based Assessment of Trauma Systems. Preliminary Draft 1. Available at: Accessed May 1, 2018.
2. Fischer PE, Schiro SE, Thomason MH. The performance of the NBATS tool in predicting trauma center need of a mature trauma system [abstract]. AAST. 2017.
3. Uribe-Leitz T, Esquivel MM, Knowlton LM, Ciesla D, Lin F, Hsia RY, Spain DA, Winchell RJ, Staudenmayer KL. The American College of Surgeons Needs-Based Assessment of Trauma Systems: estimates for the State of California. J Trauma Acute Care Surg. 2017;82(5):861–866.
4. Ashley DW, Pracht EE, Garlow LE, Medeiros RS, Atkins EV, Johns TJ, Ferdinand CH, Dente CJ, Dunne JR, Nicholas JM. Evaluation of the Georgia Trauma System using the American College of Surgeons Needs Based Assessment of Trauma Systems Tool. Trauma Surg Acute Care Open. 2018.
5. Winchell RJ, Xu P, Mount L, Huegerich R. Development of a geospatial approach for the quantitative analysis of trauma center access. AAST 2018. J Trauma Acute Care Surg. 2019;86(3):397–405.
6. United States Census Bureau. TIGER/line Shapefiles. Retrieved from: Accessed May 25, 2018.
7. ArcGIS Online. USA hospitals. Retrieved from: Accessed May 25, 2018.
8. Lansink KW, Leenen LP. Do designated trauma systems improve outcome? Curr Opin Crit Care. 2007;13(6):686–690.
9. David JS, Bouzat P, Raux M. Evolution and organisation of trauma systems. Anaesth Crit Care Pain Med. 2019;38(2):2352–5568.
10. Soto JM, Zhang Y, Huang JH, Feng D. An overview of the American trauma system. Chin J Traumatol. 2018;21(2):77–79.
11. Celso B, Tepas J, Langland-Orban B, Pracht E, Papa L, Lottenberg L, Flint L. A systematic review and meta-analysis comparing outcome of severely injured patients treated in trauma centers following the establishment of trauma systems. J Trauma. 2006;60(2):371–378.
12. Simon R, Stone M, Cucuzzo J. The impact of a new trauma center on an existing nearby trauma center. J Trauma. 2009;67(3):645–650.
13. Bell TM, Boustany KC, Jenkins PC, Zarzaur BL. The relationship between trauma center volume and in-hospital outcomes. J Surg Res. 2015;196(2):350–357.
14. Manary M, Staelin R, Boulding W, Glickman SW. Payer mix & financial health drive hospital quality: implications for value-based reimbursement policies. Behavioral Science & Policy Spring. 2015:77–84.
15. Schuurman N, Hameed SM, Fiedler R, Bell N, Simons RK. The spatial epidemiology of trauma: the potential of geographic information science to organize data and reveal patterns of injury and services. Can J Surg. 2008;51(5):389–395.
16. Nathens AB, Jurkovich GJ, Maier RV, Grossman DC, MacKenzie EJ, Moore M, Rivara FP. Relationship between trauma center volume and outcomes. JAMA. 2001;285(9):1164–1171.

Trauma systems; trauma centers; geographic modeling

© 2020 Lippincott Williams & Wilkins, Inc.