Branas CC, MacKenzie EJ, Williams JC, Schwab CW, Teter HM, Flanigan MC, Blatt AJ, ReVelle CS. Access to Trauma Centers in the United States. JAMA. 2005;293(21):2626-2633. doi:10.1001/jama.293.21.2626
Author Affiliations: Center for Clinical Epidemiology
and Biostatistics, Leonard Davis Institute of Health Economics (Dr Branas)
and Division of Traumatology and Surgical Critical Care (Drs Branas and Schwab),
University of Pennsylvania School of Medicine, Philadelphia; Johns Hopkins
Bloomberg School of Public Health, Center for Injury Research and Policy,
Baltimore, Md (Dr MacKenzie); Johns Hopkins Whiting School of Engineering,
Department of Geography and Environmental Engineering, Baltimore, Md (Drs
Williams and ReVelle); American Trauma Society, Upper Marlboro, Md (Mr Teter);
and Calspan-University at Buffalo Research Center, Center for Transportation
Injury Research, Buffalo, NY (Dr Flanigan and Mr Blatt).
Context Previous studies have reported that the number and distribution of trauma
centers are uneven across states, suggesting large differences in access to
trauma center care.
Objective To estimate the proportion of US residents having access to trauma centers
within 45 and 60 minutes.
Design and Setting Cross-sectional study using data from 2 national databases as part of
the Trauma Resource Allocation Model for Ambulances and Hospitals (TRAMAH)
project. Trauma centers, base helipads, and block group population were counted
for all 50 states and the District of Columbia as of January 2005.
Main Outcome Measures Percentages of national, regional, and state populations having access
to all 703 level I, II, and III trauma centers in the United States by either
ground ambulance or helicopter within 45 and 60 minutes.
Results An estimated 69.2% and 84.1% of all US residents had access to a level
I or II trauma center within 45 and 60 minutes, respectively. The 46.7 million
Americans who had no access within an hour lived mostly in rural areas, whereas
the 42.8 million Americans who had access to 20 or more level I or II trauma
centers within an hour lived mostly in urban areas. Within 45 and 60 minutes,
respectively, 26.7% and 27.7% of US residents had access to level I or II
trauma centers by helicopter only and 1.9% and 3.1% of US residents had access
to level I or II centers only from trauma centers or base helipads outside
their home states.
Conclusion Selecting trauma centers based on geographic need, appropriately locating
medical helicopter bases, and establishing formal agreements for sharing trauma
care resources across states should be considered to improve access to trauma
care in the United States.
The systems approach to the delivery of trauma care is widely accepted
as an effective strategy for reducing death due to injury.1- 5 A
critical component of this systems approach is the designation or verification
of trauma center hospitals equipped to treat more severely injured patients.
Although the overall number of trauma centers has increased over the last
decade, recent studies have shown that their geographic distribution varies
widely across states. These studies suggest that in many areas of the country
residents are without timely access to trauma centers that could save their
lives. In other areas, there may be too many trauma centers, possibly leading
to inefficiencies, lower patient volumes per center, and reduced quality of
We used 2 new national databases to assess the US trauma system safety-net—where
it is absent and where it is well developed. We also assessed the role of
helicopters and the sharing of trauma care resources across states.
We defined access as the percentage of the population that could reach
a trauma center within a certain time (population access) and the percentage
of land area from which a trauma center was reachable within a certain time
(land area access). Access to level I or II trauma centers and access to level
I, II, or III trauma centers was calculated. Although the precise criteria
used in differentiating levels of trauma care varies by state, most are based
on the guidelines published by American College of Surgeons Committee on Trauma.12 Level I and II trauma centers provide comprehensive
care for the most critically injured patients and have immediate availability
of trauma surgeons, anesthesiologists, and certain other physician specialists.
Level III centers provide prompt assessment, resuscitation, surgery, and stabilization,
with transfer to a level I or II center when indicated.
Two different prehospital time periods, 45 and 60 minutes, were used
in assessing access. We chose 60 minutes because a prehospital time period
longer than 60 minutes (the so-called “golden hour”13)
has been associated with a significant increase in the risk of death for severely
injured patients.14 We chose 45 minutes for
comparative purposes and because the golden hour cutoff has little scientific
evidence to support it.15 For purposes of this
study, prehospital time is measured from receipt of emergency call to hospital
To estimate the role of helicopter transport, access to trauma centers
was assessed by driving alone and by driving or flying. To estimate the contribution
of shared resources (trauma centers and helipads) across state boundaries,
access to trauma centers based on in-state trauma care resources only was
compared with access based on resources located both in-state and in neighboring
Three data sources were used to estimate access to trauma centers. Data
on trauma centers were obtained from the January 2005 Trauma Center Inventory.10,16 This inventory provides the longitude
and latitude coordinates of all level I, II, and III trauma centers that have
been designated by a state or regional authority, verified by the American
College of Surgeons Committee on Trauma (ACSCOT), or both. No level IV or
higher trauma centers were included in our analysis. We also excluded free-standing
pediatric trauma centers (n = 36) because they would have necessitated
separate, age-specific estimates of access.
Air ambulance data were obtained from the January 2005 version of the
Atlas and Database of Air Medical Services (ADAMS).17 This
database provides the longitude and latitude coordinates of all helipad base
locations operated by air medical service providers that respond to emergency
medical and trauma scenes in the 50 states and the District of Columbia. All
commercial, not-for-profit, public, and selected military air medical service
providers are listed in the database.18,19 We
included only rotorwing air medical services in our analysis (ie, no fixed-wing
services). Base helipads were typically airports, independent hangars, or
designated areas on hospital grounds to which helicopters returned in between
Our main geographic units of analysis were block groups, subdivisions
of census tracts that do not cross state boundaries. Each block group’s
population was used to calculate access by assigning it a point in space,
or centroid, that was nearest to most of its residents. The longitude-latitude
coordinates of these population-weighted centroid points were then compared
with the points representing trauma centers and base helipads. Population
estimates and population-weighted centroids for 208 667 block groups
were calculated for August 2004 (Claritas Inc, Ithaca, NY) based on data and
computational methods from the US Census Bureau,20 local
and state estimates, trends in consumer counts, and trends in deliverable
addresses from the US Postal Service.21 Block
groups were aggregated to compute estimates of access for the entire country,
the 4 Census Bureau regions (Northeast, Midwest, South, and West), all 50
states and the District of Columbia, and areas of the country defined by urbanicity
(urban, suburban, and rural).
The Trauma Resource Allocation Model for Ambulances and Hospitals (TRAMAH)7,22,23 was used as the basis
for calculating access given the locations of existing trauma centers and
base helipads. Access was calculated by summing either the population or land
area of block groups that could reach a trauma center by helicopter or ground
ambulance within the specified prehospital time period. The populations or
land areas of block groups that could reach a trauma center within the time
period specified were never counted more than once in the summation formula
for access. All programming code was written, compiled, and tested using Compaq
Visual Fortran Version 6.6 (2000 Compaq Computer Corporation, Houston, Tex).
Two coders tested separate and independently written versions of the code.
Both versions produced the same results.
To calculate ground ambulance driving times, we used an average urban
driving speed of 20.1 mph, an average suburban driving speed of 47.5 mph,
and an average rural driving speed of 56.4 mph.24 Drives
were classified as urban, suburban, or rural by averaging the population densities
(residents per square mile) of the scene and trauma center block groups and
then determining whether this average population density fell into the highest,
middle, or lowest third among all US block groups. The population densities
of intervening block groups were not considered. We then added 1.4, 1.4, and
2.9 minutes to account for the average time from receipt of emergency call
to departure in urban, suburban, and rural areas, respectively. An additional
13.5, 13.5, and 15.1 minutes in urban, suburban, and rural areas, respectively,
was added to account for the average time spent on the scene.24
Although we directly calculated the driving time from scene block group
to trauma center, we could not explicitly determine the locations of ground
ambulance depots. Therefore, as with the TRAMAH,7,22 we
estimated the time from ground ambulance depot to the scene. To obtain the
total driving time, the time from scene to trauma center was multiplied by
an empirically determined24 constant: 1.6,
1.5, and 1.4 for urban, suburban, and rural drives, respectively. All driving
distances were estimated using previously validated mathematical models of
actual road travel.25,26 All speeds
and times were previously derived as part of a meta-analysis of empirically
determined prehospital care times for trauma.24
To calculate helicopter flying times we used the typical cruise speeds
of the helicopters reported in ADAMS for each base helipad and estimated flying
distances as straight lines. We then added 3.5 minutes to account for the
average time from helicopter receipt of emergency call to take-off. Our models
assumed that helicopters and ground ambulances are simultaneously dispatched
to the scene of a trauma incident (which reportedly occurs a majority of the
time27) as opposed to ground ambulances responding
first and assessing the situation before calling for a helicopter. An additional
21.6 minutes was added to account for the average time spent on the scene
(including landing zone ascertainment). These times were previously derived
as part of a meta-analysis of empirically determined prehospital care times
for trauma.24 As with the TRAMAH,7,22 base
helipads and trauma centers were paired if they were within the specified
prehospital time period of one another. Base helipads and trauma centers that
were not within the specified prehospital time period of one another were
not paired. Paired helipads increased the percentage of residents with access:
they could launch a helicopter, pick up patients, and reach at least one trauma
center within the specified prehospital time period.7,28
We also performed a sensitivity analysis to judge the extent to which
our assumptions about input times would alter estimates of access. Even though
the typical prehospital times we used were derived as part of a meta-analysis
of empirically determined prehospital care times for trauma,24 specific
trauma systems will disagree about what prehospital times are meaningful.
We changed our assumptions about different ambulance travel times (urban drives,
suburban drives, rural drives, and helicopter flights) in 5-minute increments
to see how they ultimately affected our estimates of access.
As of January 2005, there were 190 level I, 255 level II, and 258 level
III trauma centers located throughout the United States (ie, 1.5 level I and
II and 2.4 level I, II, and III trauma centers per million population). These
trauma centers were served by 571 base helipads and 683 helicopters (ie, 1.9
base helipads and 2.3 helicopters per million population).
An estimated 69.2% of all US residents had access to a level I or II
trauma center within 45 minutes by ground ambulance or helicopter. When this
time parameter was increased to 60 minutes, access increased to 84.1% of the
US population (Table 1). The Northeast
had the greatest access to level I or II centers within both 45 and 60 minutes
(85.8% and 96.9%, respectively) followed by the West (76.5% and 86.0%), the
Midwest (65.7% and 85.7%), and the South (58.9% and 76.1%). Within 45 minutes,
8.4% of rural, 72.7% of suburban, and 89.4% of urban block groups had access
to level I or II trauma centers. These percentages increased to 24.0%, 86.2%,
and 95.3% within 60 minutes.
Across the United States, population access increased by 5.0% for 45
minutes and 4.6% for 60 minutes when level III trauma centers were included.
Level III centers increased access most heavily in the South for both 45 and
60 minutes (8.5% and 8.5%, respectively) followed by the West (4.7% and 3.6%),
Midwest (3.0% and 2.0%), and Northeast (0.8% and 0.7%) (Table 1).
Totals of 9.6% and 25.1% of US land area were located within 45 and
60 minutes, respectively, of a level I or II trauma center (Figure 1). The Northeast had the greatest amount of land area located
within both 45 and 60 minutes (36.6% and 69.9%, respectively) of level I or
II centers, followed by the Midwest (13.1% and 34.5%), the South (12.0% and
35.9%), and the West (4.7% and 12.5%).
Americans had an average of 5.5 level I and II trauma centers and 6.1
level I, II, and III trauma centers accessible by ground ambulance or helicopter
within 45 minutes. These averages increased to 10.0 and 11.5 for 60 minutes.
The population in the Northeast had the highest average number of level I
and II trauma centers accessible within both 45 and 60 minutes (13.0 and 25.1,
respectively), followed by the Midwest (6.2 and 11.0), West (4.2 and 7.4),
and South (2.0 and 3.5). Within 45 minutes and 60 minutes, respectively, 39.1%
and 33.8% of Americans had access to between 1 and 4 trauma centers while
7.7% and 14.6% had access to 20 or more level I or II centers.
Americans were serviced by an average of 5.1 and 9.4 base helipads within
45 and 60 minutes, respectively. The population in the Northeast was serviced
by the highest average number of base helipads within both 45 and 60 minutes
(12.4 and 24.3, respectively), followed by the Midwest (5.8 and 10.4), West
(3.8 and 6.9), and South (1.7 and 3.0). Within 45 and 60 minutes, respectively,
41.0% and 38.7% of Americans were serviced by 1 to 4 base helipads while 7.6%
and 12.7% were serviced by 20 or more base helipads.
A total of 26.7% and 27.7% of US residents could access a level I or
II trauma center within 45 and 60 minutes, respectively, only if they were
flown by helicopter. Helicopters had the greatest impact on access to level
I or II centers within 45 minutes in the West (34.0%) followed by the Northeast
(28.1%), South (25.8%), and Midwest (22.3%). Helicopters had the greatest
impact on access to level I or II centers within 60 minutes in the South (29.8%)
followed by the Midwest (29.4%), West (27.4%), and Northeast (23.7%) (Table 2).
A total of 1.9% and 3.1% of US residents had access to level I or II
centers within 45 and 60 minutes, respectively, solely due to trauma centers
or base helipads (or both) located outside their home states. Access to level
I or II centers within 45 minutes was increased most by the resources of neighboring
states in the Northeast (3.2%) followed by the Midwest (2.8%), South (2.0%),
and West (0.2%). Access to level I or II centers within 60 minutes was increased
most by the resources of neighboring states in the Midwest (4.6%), followed
by the South (4.4%), Northeast (1.7%), and West (0.8%) (Table 2).
The sensitivity analysis showed that driving times changed the final
45- and 60-minute percentages by less than 1% when up to 10 minutes was added
or subtracted. The final 60-minute access percentage for the United States
was increased by 5.6% when it was assumed that helicopter flying times took
10 minutes less than expected and decreased by 8.4% when it was assumed that
helicopter flying times took 10 minutes more than expected. The final 45-minute
access percentage for the United States was increased by 11.0% when assuming
that helicopter flying times took 10 minutes less than expected and decreased
by 18.8% when assuming that helicopter flying times took 10 minutes more than
expected (Figure 2).
As of January 2005, the US trauma care system as a whole had evolved
to the point where more than four fifths of Americans had access to level
I or II trauma centers within an hour. At the same time, however, 46.7 million
Americans had no access to a level I or II trauma center within an hour. Most
Americans without access lived in rural areas of the country. Given these
gaps, our analyses point to several modifiable aspects of the trauma care
system that could improve access.
First, 13.6 million Americans had access to level III, but not level
I or II, trauma centers within an hour. Level III hospitals provide initial
evaluation and assessment of injured patients and transfer the more severely
injured to a higher level of care.12 Since
the early 1990s the number of level III centers has increased in both urban
and rural areas of the country in an attempt to develop more inclusive systems
of care.10 Although controversy remains as
to the appropriate role of level III trauma centers,29- 33 access
to the trauma care system can be increased by establishing level III centers
for communities where no level I or II center is accessible.
Second, 10.3 million Americans had access within an hour via the trauma
centers and helicopters of neighboring states. This is substantial enough
to warrant the attention of planners in better preparing for day-to-day volumes
of trauma patients as well as mass casualty incidents in which volumes could
overwhelm the resources of any one state.34,35 To
date, 47 states and the District of Columbia have ratified standardized mutual
aid agreements to improve interstate cooperation for mass casualty response.36 Only 31 states, however, have standardized protocols
for prehospital triage, and special permissions or certificates are sometimes
required for out-of-state ambulances.37 Where
appropriate, neighboring states should support border crossing arrangements
for day-to-day trauma care because of the high number of patients who might
benefit and because border crossing for day-to-day volumes of trauma will
ultimately help prepare for the execution of larger, interstate mutual aid
agreements during mass casualty incidents.38
Third, helicopters provided access for 81.4 million Americans who otherwise
would not have been able to reach a trauma center within an hour. This, and
the fact that base helipads are more moveable than trauma centers, makes them
appealing as modifiable components of the trauma system. Because many trauma
centers are in urban areas, the location of new base helipads (preferably
as satellites to trauma centers as opposed to at the trauma centers themselves28) could benefit trauma patients in outlying suburban
and rural areas. Additional medical helicopter flight programs could be an
important, and practical, means of extending trauma center access to populations
that currently have none.39
Despite the need to address areas with limited access, our results also
showed that 42.8 million Americans had access to 20 or more trauma centers
within an hour. These high levels of access were most evident near urban centers
such as those in the Northeast. Although it remains uncertain at what point
access to multiple trauma centers becomes excessive, placing appropriate limits
on the number of trauma centers may produce higher volumes of severely injured
patients per center, enhance the experience of providers, and improve the
overall quality of trauma care.11,40- 42
Our study has several limitations. Our estimates of access were based
on where people lived and not where they were injured. Although people are
certainly injured outside of their residences, no national data exist that
provide the locations of all types of severe injuries at a very small level
of aggregation (such as block groups).
Nongeographic issues that could have potentially changed access were
not considered.43 These issues included areas
without 911 telephone service, inappropriate prehospital triage, inclement
weather, roadway congestion, and out-of-service times for ambulances and trauma
centers. However, the effect of these factors on our results was probably
minimal: most people in the United States have 911 access,44 relatively
few helicopter flights are precluded by weather,45,46 traffic
conditions reportedly have only minor effects on ground ambulance emergency
response speeds,47 and helicopters are estimated
to be fully out of service only a small percentage of the time.48 Other
process issues, such as inappropriate prehospital triage, are beyond the scope
of this study but have been addressed elsewhere.49,50
A final limitation of our study concerns the precision of our input
assumptions about prehospital times and travel speeds. Changes in estimates
of helicopter flying times produced the largest shifts in the percentages
with access to a trauma center. Because we used individually reported helicopter
flying speeds from each base helipad, most of these shifts would likely be
due to changes in other helicopter travel inputs, such as the length of warm-up
or on-scene time intervals. By comparison, ground ambulance travel inputs
generated very small shifts in our estimates of access.
To our knowledge, this is the first national assessment of trauma center
access that simultaneously considers the locations of trauma centers, ambulances,
and residential populations. Although some work has been completed at the
local, state, and regional levels,7,51,52 our
study has the advantage of estimating access both in terms of prehospital
time, as opposed to distance, and in terms of population, as opposed to land
area. Distance and land area metrics can lead to potentially misleading impressions
of access.53 For instance, on a map, the western
United States seemed to have little access to trauma centers when, in fact,
its population had a high level of access that was second only to the Northeast.
This study demonstrates that should they be seriously injured, 46.7
million Americans would have no access to a level I or II trauma center within
an hour. An approximately equal number of Americans, 42.8 million, would have
access to 20 or more level I or II trauma centers within an hour. Judiciously
selecting trauma centers based on geographic need, appropriately locating
medical helicopter bases, and establishing formal agreements for sharing trauma
care resources across states should be considered to improve access to trauma
care in the United States.
Corresponding Author: Charles C. Branas,
PhD, Center for Clinical Epidemiology and Biostatistics, University of Pennsylvania
School of Medicine, Room 829 Blockley Hall, 423 Guardian Dr, Philadelphia,
PA 19104-6021 (email@example.com).
Author Contributions: As principal investigator,
Dr Branas had full access to all of the data in the study and takes responsibility
for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Branas, MacKenzie,
Williams, Schwab, ReVelle.
Acquisition of data: Branas, MacKenzie, Teter,
Analysis and interpretation of data: Branas,
MacKenzie, Williams, Teter.
Drafting of the manuscript: Branas.
Critical revision of the manuscript for important
intellectual content: Branas, MacKenzie, Williams, Schwab, Teter, Flanigan,
Statistical analysis: Branas, MacKenzie, Williams.
Obtained funding: Branas, MacKenzie, Teter,
Administrative, technical, or material support:
Branas, MacKenzie, Schwab, Flanigan, Blatt, ReVelle.
Study supervision: Branas.
Financial Disclosures: None reported.
Funding/Support: This work was supported in
part by the American Trauma Society Trauma Information and Exchange Program,
grant R01HS010914 from the Agency for Healthcare Research and Quality, and
grant H28CCH319122 from the Centers for Disease Control and Prevention’s
Center for Injury Prevention and Control.
Role of the Sponsors: One or more of these
organizations were separately involved in the design and conduct of the study,
the collection and interpretation of the data, and the review and approval
of the manuscript.
Disclaimer: This article is solely the responsibility
of the authors and does not necessarily represent the official views of the
Center for Transportation Injury Research, the Association of Air Medical
Services, the US Department of Transportation, or the US Department of Health
and Human Services.
Acknowledgment: We are grateful for the contributions
and thoughtful suggestions of Anthony R. Carlini, MS, Johns Hopkins Bloomberg
School of Public Health; Jianguo Li, MSc, MS, University of Pennsylvania School
of Medicine; Louis V. Lombardo, National Highway Traffic Safety Administration;
Ryan P. Radecki, Ohio State University College of Medicine; and Sandra D.
Teitelbaum, MLS, American Trauma Society and Johns Hopkins Bloomberg School
of Public Health.