The Society of Black Academic Surgeons
Insurance status is a potent predictor of outcomes in
both blunt and penetrating trauma
Wendy R. Greene, M.D.a,*, Tolulope A. Oyetunji, M.D., M.P.H.a, Umar Bowers, B.S.a,
Adil H. Haider, M.D., M.P.H.b, Thomas A. Mellman, M.D.a,
Edward E. Cornwell, M.D.a, Suryanarayana M. Siram, M.D.a,
David C. Chang, Ph.D., M.P.H., M.B.A.a,b
aDepartment of Surgery, Howard University College of Medicine, Washington, DC;bDepartment of Surgery, Johns
Hopkins School of Medicine, Baltimore, MD, USA
BACKGROUND: Patients with penetrating injuries are known to have worse outcomes than those
with blunt trauma. We hypothesize that within each injury mechanism there should be no outcome
difference between insured and uninsured patients.
METHODS: The National Trauma Data Bank version 7 was analyzed. Patients aged 65 years and
older and burn patients were excluded. The insurance status was categorized as insured (private,
government/military, or Medicaid) and uninsured. Multivariate analysis adjusted for insurance status,
mechanism of injury, age, race, sex, injury severity score, shock, head injury, extremity injury, teaching
hospital status, and year.
significantly higher in penetrating trauma patients versus blunt trauma patients (7.9% vs 3.0%; P ? .001),
and higher in the uninsured (5.3% vs 3.2%; P ? .001). On multivariate analysis, uninsured patients had an
increased odds of death than insured patients, in both penetrating and blunt trauma patients. Penetrating
trauma patients with insurance still had a greater risk of death than blunt trauma patients without insurance.
CONCLUSIONS: Insurance status is a potent predictor of outcome in both penetrating and blunt
© 2010 Elsevier Inc. All rights reserved.
Mechanism of injury;
National trauma data
According to the Centers for Disease Control, trauma is
the leading cause of death in the age groups from 1 to 44
years.1Minorities and the uninsured are represented dispro-
portionately in these statistics. The health disparity litera-
ture has yet to clearly define why these groups have an
increased risk of mortality. Morris et al2suggested the
presence of underlying disease affects mortality, which was
evidence to support triage decisions to trauma centers.
Haider et al3described the effect of ethnic and insurance
status in relation to trauma mortality; however, the effect of
mechanism of injury was not well defined. Minorities and
uninsured patients are more likely to be involved in pene-
trating trauma, which is well documented to be a more lethal
mechanism of injury.4To ameliorate ethnic and socioeco-
nomic disparities in trauma mortality it is important to
Presented at the 19th Annual Scientific Session of the Society of Black
Academic Surgeons, April 2–4, 2009, Seattle, WA.
* Corresponding author. Tel.: ?1-202-865-1285; fax: ?1-202-865-
E-mail address: email@example.com
Manuscript received June 29, 2009; revised manuscript September 28,
0002-9610/$ - see front matter © 2010 Elsevier Inc. All rights reserved.
The American Journal of Surgery (2010) 199, 554–557
understand the relationships between risk factors. The pri-
mary objective of this study was to evaluate whether insur-
ance status had an effect on trauma mortality that was
independent of the mechanism of blunt versus penetrating
This study involved a retrospective analysis of the Na-
tional Trauma Data Bank version 7.0 from 2002 to 2006.
This voluntary database, which is maintained by the Amer-
ican College of Surgeons, has more than 700 hospital con-
tributors and more than 1.8 million trauma incidents re-
corded. Patients were excluded if they were a burn victim or
older than age 65 years. Patients older than 65 years were
excluded because these patients are eligible for Medicare
and as such most of this patient population will be insured.
In addition, patients were excluded if they had incomplete
information on insurance status or mechanism of injury.
Mortality after trauma was the outcome measure. Com-
parisons were made according to insured versus uninsured
status, and mechanism of injury. The insured group in-
cluded private, government/military, and Medicaid recipi-
ents. The Medicaid recipients were grouped with the insured
patients because its coverage, however limited, is still con-
sidered a valid form of insurance. The mechanisms of injury
were categorized as blunt versus penetrating trauma. To
evaluate the interaction of these variables, patients were
categorized as blunt insured (BI), blunt uninsured (BU),
penetrating insured (PI), and penetrating uninsured (PU).
Unadjusted comparison between groups was made using
t tests for continuous variables, or with chi-square tests for
categoric variables. Multiple logistic regression analysis
included the following covariates: insurance status, type
of injury, age, self-reported ethnicity, sex, injury severity
score (ISS), presence of shock, head injury, extremity in-
jury, teaching hospital status, and year of injury. Interaction
between insurance status and mechanism of injury was
examined. Subset analysis then was performed in patients
between ages 18 and 30 years. Statistical analysis was
performed in Stata/MP version 10 (Stata, College Station,
TX). Statistical significance was defined as a P value of less
Of the 1,862,348 patients in the National Trauma Data
Bank during the years studied (2002–2006), 920,269 pa-
tients met the inclusion criteria. Insured patients accounted
for 70% of the study population. Most of the patients
(86.3%) had blunt trauma, whereas penetrating injuries ac-
counted for the remaining 13.7%. Of those with information
on ethnicity, 549,327 (62.5%) were white, 152,866 (17.4%)
were black, 115,420 (13.1%) were Hispanic, and other eth-
nicities made up the remaining 7%. The population was
predominantly male (69.4%), with a mean age of 31 years
(median, 30 y). The median length of stay was 2 days and
the median ISS was 9. Fifty-four percent of patients were
treated at a teaching hospital and 96% of the patients sur-
vived their injuries.
Table 1 presents demographic data of the 4 comparison
groups. There were more males than females in all groups,
with a higher percentage of males in the penetrating groups
than in the blunt injury groups. The mean (median) age for
BI and PI was 31 years (median, 30 y) and 29 years (me-
dian, 26 y), respectively. The mean age for BU and PU was
32 (31) and 30 (27) years, respectively. There were a higher
proportion of minorities in the penetrating trauma groups. In
addition, in the penetrating trauma group, there were a
higher proportion of uninsured patients with gunshot
wounds than insured patients (51.3% vs 42.1%; P ? .001).
All groups had similar median ISS and lengths of stay.
There was a difference by mechanism of injury in terms of
head trauma, with blunt trauma patients more likely to
sustain a head injury than penetrating trauma patients; how-
ever, no difference was noted in head injury by insurance
type. Most patients were transported to a teaching hospital.
The mortality for BI and PI was 14,232 (2%) and 2,869
(5%), whereas mortality for both BU and PU was higher, at
8,058 (4%) and 6,585 (11%), respectively.
On multivariate analysis, uninsured patients had greater
odds of death than insured patients, within both penetrating
and blunt trauma patients. Penetrating trauma patients with
insurance still have a greater risk of death than blunt trauma
patients without insurance (Fig. 1). The complete data from
the multivariate analysis are presented in Table 2. It is noted
that all minority groups have higher risks of death when
compared with whites. There is a steady increase in mor-
tality risk associated with age. As noted in the literature,
women are less likely to die than men. As expected, patients
with a higher ISS, in shock, and with a head injury are more
likely to die. Interestingly, university hospital patients are
slightly less likely to die than community hospital patients.
Furthermore, on interaction analysis, we found that there
this study: BI, BU, PI, and PU
Demographic data of the 4 comparison groups in
Median age, y
Male sex, %
Head injury, %
Median length of stay
Transported to teaching
578,012 215,646 62,942 63,117
52 5657 63
555W.R. Greene et al. Insurance potent predictor trauma outcomes
was a significant synergistic interaction between insurance
and mechanism of injury, in which the influence of pene-
trating injury on mortality outcome in the uninsured group
is greater than in the insured group (odds ratio [OR], 1.22;
P ? .001).
On subset analysis, the impact of insurance status per-
sists even when analyzing young patients (age, 18–30 y)
who presumably have little comorbidity. The OR of death in
these young patients, relevant to insured blunt patients, are
BU, 1.75 (95% confidence interval [CI] 1.62–1.89), PI, 2.36
(95% CI, 2.09–2.65), uninsured penetrating, 5.94 (95% CI,
5.42–6.51). These ORs, although they are different from
each other, are not statistically different from the ORs for
the overall population as presented in Fig. 1.
Our results highlight the importance of insurance status
to trauma-related mortality. Those patients who have insur-
ance are less likely to die. The magnitude of this problem is
overshadowed only by the complexities of the contributing
factors. Many studies show the ethnic disparity in health
care and now have attributed this to the care of the trauma
patient. But why would patients who sustain trauma have
different care? Trauma care is one of the most protocolized
and standardized systems of care with direction provided
by the American College of Surgeons–sponsored advanced
trauma life support education. The confounding variables,
which influence outcomes, are multifactorial and all aspects
should be addressed. The mechanism of injury should be
addressed by speed reduction strategies and ethnic dispari-
ties should be addressed by ensuring equitable care for all.
In a similar fashion, health care insurance disparities should
be addressed by providing health care options for all.
In this study, we found that insurance status exerts a
strong impact on mortality in both penetrating and blunt
trauma patients, even among young adults. This article
continues a series of publications that suggest, surprisingly,
that the effect of insurance on outcomes exist even in
trauma care. Our primary aim in this study was to investi-
gate the relationship between the insurance status and the
mechanism of injury: blunt versus penetrating. Our analysis
indicated that the increased number of penetrating trauma
incidents in the uninsured group does not fully account for
the disparity in trauma mortality related to insurance status.
The health disparity literature has proposed additional vari-
ables that influence trauma mortality. They include comor-
bidities, socioeconomic status (SES), weaponry, bias, and
The current literature suggests that there are outcome
disparities not only by race but also by insurance status. The
study by Haider et al3emphasized that insurance status has
a stronger affect on outcomes than ethnicity.3The insured
Blacks were less likely to die than the uninsured Whites.
This disparity was not seen in the Latin American group and
confounding biases and language barriers may have been
Multivariate analysis for in-hospital mortality
Mechanism of Injury
Odds Ratio of Death
status and mechanism of injury.
Multivariate analysis of OR of death by insurance
556The American Journal of Surgery, Vol 199, No 4, April 2010
contributing factors. In addition, Haas and Goldman5found Download full-text
that adult trauma patients without health insurance were
more likely to die in a hospital than were patients with
insurance, after adjustment for ISS. The race and insurance
status are therefore independent risk factors for trauma
mortality. Although the disparities are not explained com-
pletely by ethnicity, we believe that this area can be im-
proved through culturally competent care and equal use of
We acknowledge that there are alternative explanations
for our findings. It is possible that the difference between
outcomes for insured and uninsured patients could be
caused by differences in pre-existing disease and comorbid-
ity. Health insurance status plays a key role in access to
health care services. Recent studies have found that unin-
sured Americans have less access to medical care6and are
less likely to receive appropriate care once they have gained
access to the health care system.7The injured uninsured
patient may have worse outcomes because of undiagnosed
and untreated comorbidities. Milzman et al8found that even
when controlling for ISS and Glasgow Coma Scale score,
patients with pre-existing disease and older than age 55 had
a higher mortality rate compared with those without pre-
existing disease (P ? .001). This explanation seems less
likely in view of our having found similar differences be-
tween insured and uninsured patients in the 18- to 30-year-
old age group.
Other explanations may relate to factors associated with
lower SES (which is associated with being uninsured) in
addition to insurance status. Many studies have shown cor-
relations of injury, mortality rates, and socioeconomic fac-
tors. In the pediatric population, it has been shown that
children from lower SES communities had higher injury
hospitalization and mortality rates, and presented more fre-
quently with more lethal mechanisms of injury (pedestrian,
firearm).9Similarly, studies of the adult population have
suggested that a possible mechanism may be a deficiency in
It is possible that different weaponry may be involved in
insured versus uninsured patients’ injuries. In our study,
uninsured patients were wounded by higher-velocity weap-
ons (gunshot wound) than insured patients (stabbings), and
these differences in weaponry used may account for some of
the differences in outcomes between insured versus unin-
sured penetrating trauma patients. Although we adjusted for
the ISS, we must bear in mind that the ISS is known to
underestimate severity in penetrating trauma patients.11
In addition, we must ponder whether bias is involved.
Although equally unacceptable, it is easier to conceptualize
the existence of disparities in outcomes related to race/
ethnicity because these are characteristics that can be ob-
served by providers. Studies have shown remarkable differ-
ences in outcomes for chronic and even acute illnesses when
races are compared.12,13Language barriers and low health
literacy are other factors that must be considered. Dealing
with patients with low literacy takes time—a commodity
that is in short supply.10
Our study was limited by the absence of data to explore
the alternative explanations listed earlier. However, we be-
lieve that we have addressed comorbidity by performing a
subset analysis on the 18- to 30-year-old age group.
In contrast, the large number of trauma incidents ana-
lyzed and the breadth of coverage strengthen our study
across the country. The broad-based nature of the database
makes the findings more generalizable. Future studies
should examine the course of care in the prehospital, emer-
gency department, and posthospital setting, as well as the
causes of death.
In conclusion, insurance status exerts a strong impact on
mortality in both penetrating and blunt trauma patients, and
this effect is observed even among young adults. The evi-
dence continues to mount in support of the proposition that
a productive first step in eliminating health outcome dispar-
ities would be to address health coverage disparities.
1. Center for Disease Control and Prevention’s National Center for
Health Statistics: National Vital Statistics System. Atlanta: GA; 2004.
2. Morris JA Jr, Mackenzie EJ, Edelstein SL. The effect of preexisting
conditions on mortality in trauma patients. JAMA 1990;263:1942–6.
3. Haider AH, Chang DC, Efron DT, et al. Race and insurance status as
risk factors for trauma mortality. Arch Surg 2008;143:945–9.
4. Brohi K. Peripheral vascular trauma. Available at: http://www.
trauma.org. Accessed May 15, 2009.
5. Haas JS, Goldman L. Acutely injured patients with trauma in Massa-
chusetts: differences in care and mortality, by insurance status. Am J
Public Health 1994;84:1605–8.
6. Ayania JZ, Weissman JS, Schneider EC, et al. Unmet health needs of
uninsured adults in the United States. JAMA 2000;284:2061–9.
7. Richardson LC, Tian L, Voti L, et al. The roles of teaching hospitals,
insurance status, and race/ethnicity in receipt of adjuvant therapy for
regional-stage breast cancer in Florida. Am J Public Health 2006;96:
8. Milzman DP, Boulanger BR, Rodriquez A, et al. Pre-existing disease
in trauma patients: a predictor or fate independent of age and injury
severity score. J Trauma 1992;32:236–244.
9. Marcin JP, Schembri MS, Jinsong H, et al. A population-based anal-
ysis of socioeconomic status and insurance status and their relationship
with pediatric trauma hospitalization and mortality rates. Am J Public
10. Williams MV, Davis T, Parker RM, et al. The role of health literacy in
patient-physician communication. Fam Med 2002;34:383–9.
11. Cornwell EE 3rd, Velmahos GC, Berne TV, et al. Lethal abdominal
gunshot wounds at a level I trauma center: analysis of TRISS (Revised
Trauma Score and Injury Severity Score) fallouts. J Am Coll Surg
12. Ward E, Ahmedin J, Vilma C, et al. Cancer disparities by race/
ethnicity and socioeconomic status. CA Cancer J Clin 2004;54:78–93.
13. Johnson PA, Lee TH, Cook EF, et al. Effect of race on the presentation
and management of patients with acute chest pain. Ann Intern Med
557W.R. Greene et al.Insurance potent predictor trauma outcomes