J Neurosurg Spine 13:61–66, 2010

Health care burden of cervical spine fractures in the

United States: analysis of a nationwide database over
a 10-year period

Presented at the 2009 Joint AANS/CNS Spine Section Meeting

Clinical article
Ali A. Baaj, M.D.,1,2 Juan S. Uribe, M.D.,1 Tann A. Nichols, M.D., 3
Nicholas Theodore, M.D., 2 Neil R. Crawford, Ph.D., 2 Volker K. H. Sonntag, M.D., 2
and Fernando L. Vale, M.D.1
1
Department of Neurosurgery, College of Medicine, University of South Florida, Tampa, Florida; 2Barrow
Neurological Institute, St. Joseph’s Hospital and Medical Center, Phoenix, Arizona; and 3Department of Neu-
rosurgery, University of Cincinnati College of Medicine, Mayfield Clinic, Cincinnati, Ohio

Object. The objective of this work was to search a national health care database of patients diagnosed with cervi-
cal spine fractures in the US to analyze discharge, demographic, and hospital charge trends over a 10-year period.
Methods. Clinical data were derived from the Nationwide Inpatient Sample (NIS) for the years 1997 through
2006. The NIS is maintained by the Agency for Healthcare Research and Quality and represents a 20% random strati-
fied sample of all discharges from nonfederal hospitals within the US. Patients with cervical spine fractures with and
without spinal cord injury (SCI) were identified using the appropriate ICD-9-CM codes. The volume of discharges,
length of stay (LOS), hospital charges, total national charges, discharge pattern, age, and sex were analyzed. National
estimates were calculated using the HCUPnet tool.
Results. Approximately 200,000 hospitalizations were identified. In the non-SCI group, there was a 74% in-
crease in hospitalizations and charges between 1997 and 2006, but LOS changed minimally. There was no appre-
ciable change in the rate of in-hospital mortality (< 3%), but discharges home with home health care and to skilled
rehabilitation or nursing facilities increased slightly. In the SCI group, hospitalizations and charges increased by 29
and 38%, respectively. There were no significant changes in LOS or discharge status in this group. Spinal cord injury
was associated with increases in LOS, charges, and adverse outcomes compared with fractures without SCI. Total
national charges associated with both groups combined exceeded $1.3 billion US in 2006.
Conclusions. During the studied period, increases in hospitalizations and charges were observed in both the SCI
and non-SCI groups. The percentage increase was higher in the non-SCI group. Although SCI was associated with
higher adverse outcomes, there were no significant improvements in immediate discharge status in either group dur-
ing the 10 years analyzed. (DOI: 10.3171/2010.3.SPINE09530)

Key Words      •      cervical spine      •      fractures      •      Nationwide Inpatient Sample      •     

outcomes

S
pinal trauma is a major source of morbidity and
mortality in the US. Cervical spine trauma, in par-
ticular, is associated with high costs, LOS, likeli-
hood of SCI, and death.13,19 Conservative and surgical
treatments for cervical fractures have improved,9,17 and
our understanding of the pathophysiology of traumatic
injuries, particularly of the cervical spinal cord, has in-
creased.7,8,19 Surgical advances in the treatment of cervi-
cal spine fractures include an ever-expanding array of
Abbreviations used in this paper: LOS = length of stay; NIS =
Nationwide Inpatient Sample; SCI = spinal cord injury.
complex implants and techniques that use new (and of-
ten expensive) technologies, including fusion biologics,
polymer grafts, and instrumentation.
A limited number of studies, however, have ad-
dressed recent socioeconomic trends associated with the
treatment of traumatic cervical disease.4,5,13 The goals of
this study were to analyze trends related to cervical spine
fractures based on a nationwide database and to elucidate
the health care burden of this major entity. Analysis fo-
cused on trends in hospital charges associated with treat-
ment of cervical spine fractures and status on discharge.

J Neurosurg: Spine / Volume 13 / July 2010 61

 A. A. Baaj et al.

Methods

831,590,240
Clinical data were extracted from the NIS for the

35,984
23,110

8.64
2.69

27.77
2006

7.74

5.8

57.1
years 1997 through 2006. The NIS, which is maintained
by the Agency for Healthcare Research and Quality, rep-
resents a 20% random stratified sample of all discharges
from nonfederal hospitals within the US.23 It is the larg-

663,549,168
est all-payer inpatient care database in the US and con-

35,088
18,911

2.65

7.61
57.61
27.78
2005

6.39

5.7
tains data from about 8 million hospital stays from 1000
hospitals each year. The NIS is the only national hospital
database containing charge information on all patients,
regardless of payer, including persons covered by Medi-

691,554,969
care, Medicaid, private insurance, and the uninsured.

33,569
20,601

2.86

7.85
59.87
25.87
2004

7.03
Patients with cervical spine fractures with and with-

6.1
out SCI were identified using the appropriate ICD-9-CM
codes (805.00–805.08, 806.00–806.09). The HCUPnet
tool (http://hcupnet.ahrq.gov/) was used to extract data

639,656,668
on discharges, LOS, hospital charges, national bill, dis-

36,508
17,521

59.64
24.56
3.33

6.67
2003

6.03
charge pattern, age, and sex. When accessing the NIS

6.3
data set through the HCUPnet tool, a weighted factor is
already applied and national estimates are generated. A
Bureau of Labor statistics tool (http://www.bls.gov/data/

461,032,528
inflation_calculator.htm) was used to adjust hospital and

30,536
15,098

25.29
2.98
60.39

6.47
2002
national charges for inflation. Population-adjusted rates

5.24

6.4
(discharges per measure of population) were calculated
using population estimates generated by the US Census
Year
Bureau.22 The t-test (SPSS v 17.0, SPSS Inc.) was em-

313,617,348
ployed to determine statistical significance between SCI
12,983

24,156

7.34
2.83
60.18
and non-SCI groups. A probability value less than 0.05
2001

4.55

24.5
6.1
TABLE 1: Hospitalizations associated with cervical spine fractures without SCI from 1997 to 2006*

was considered significant.

Results

323,316,781
13,067

24,743

61.99
25.21
6.37
1.75
2000

We identified approximately 200,000 hospitalizations

4.63

6.1

with a cervical spine fracture constituting the principal

diagnosis. Most cases (160,455) were cervical spine frac-
tures not associated with SCI. In this group, hospitaliza-
tions increased from 1997 (11,902) to 2006 (23,110), but
307,717,464
22,952
13,407

23.59
63.03
2.52

the change in LOS (6.3 days in 1997 to 5.8 days in 2006)

1999

4.92

6.2
6.3

was minimal. Population-adjusted rates (per 100,000)

*  D/C = discharge; rehab = rehabilitation; SNF = skilled nursing facility.

demonstrated an increase of 74% in hospitalizations be-
tween 1997 and 2006 (Table 1). Inflation-adjusted hos-
pital charges per hospitalization also increased dramati-
315,603,045
13,855

22,779

cally from $20,701 in 1997 to $35,984 in 2006 (+74%).

5.54
2.34

23.31
1998

5.13

64.2
6.6

Based on discharge status, there was no appreciable

change in the percentage of in-hospital deaths (< 3%) dur-
ing the period. However, discharges to skilled rehabilita-
tion/nursing facility and to home with home health care
246,383,302

increased slightly from 22.46 and 5.9%, respectively, in

20,701
11,902

22.46
2.19
64.11
1997

4.44

5.9
6.3

1997 to 27.77 and 8.64%, respectively, in 2006 (Fig. 1 up-

per).
Table 2 provides a summary of the hospitalizations,
charges, and discharge statuses associated with cervical
population-adjusted rate (per

spine fractures in patients with SCI from 1997 to 2006.

There was only a moderate (29%) increase in hospitaliza-
tions during this period for this group. Inflation-adjust-
Variable

home health (%)

ed hospital charges per hospitalization increased from

hospitalizations

routine D/C (%)

rehab/SNF (%)
national bill ($)

$84,722 in 1997 to $117,081 in 2006 (+38%). More than

100,000)

charges ($)
LOS (days)

death (%)

50% of the patients were discharged to a skilled nursing

facility. The in-hospital mortality rate was stable at about
12% (Fig. 1 lower).

62 J Neurosurg: Spine / Volume 13 / July 2010

Cervical spine fractures in US

519,137,154
117,081

26.66

4.04
12.49

50.51
2006
4434
1.48

17.6

446,257,395
119,865

23.33
12.65

51.73
2005
3723

4.12
1.26

16.4

591,782,232
115,854

26.28
52.48
11.76
2004

3.19
5108
1.74

18

479,237,700
115,340

47.44
24.48

3.48
2003
4155

12.11
1.43

17.3

334,183,356
98,492

3.26
25.85
49.35
2002
3393

12.16
17.8
1.18
Year

281,838,504
88,796

52.04
22.27

3.65
2001

12.19
3174

19.5
1.11
TABLE 2: Hospitalizations associated with cervical spine fractures with SCI from 1997 to 2006

Fig. 1.  Graphs. Discharge status from 1997 to 2006 in patients with
cervical spine fractures without SCI (upper) and with SCI (lower).

266,331,282
80,682

13.22

48.96
3.23
25.01
2000
3301

17.9
1.17

The total “national bill,” which is the product of

hospitalizations and charges per hospitalization associ-
ated with both non-SCI and SCI cervical spine fractures
295,345,116

jumped from about $500 million in 1997 to $1.3 billion in

81,972

8.22
24.27
54.03
3.01
3603
1999

1.32

17.8

2006 (+160%). Figures 2 through 7 show the differences

in hospitalizations, LOS, charges, and discharge statuses
between the SCI and non-SCI groups. The 10-year aver-
ages in LOS, charges and discharge status between the
278,427,864

SCI and non-SCI groups (Table 3) were all significantly

81,127

24.96

3.24
48.31
3432
1998

12.16
18.2
1.27

different (p < 0.001).

The cervical level most involved in the non-SCI
group was C-2 (38%), and most SCIs occurred at C5–7
(Table 4). In the SCI group, males constituted more than
261,706,258

*  LOS = length of stay, D/C = discharge.

75% of admissions, and the 18 to 44–year-old age group

84,722

4.08
9.05
26.92
50.24
3089
1997

20.2
1.15

composed 40% of those admissions. Federal programs

(Medicare/Medicaid) funded 34% of the hospitalizations
in the SCI group, and teaching hospitals provided 73% of
the care.
population-adjusted rate

rehab nursing/home (%)

Discussion
Variable

home health (%)

(per 100,000)
hospitalizations

routine D/C (%)

national bill ($)

The results of our study highlight the continuous and

charges ($)
LOS (days)

significant health care burden of cervical spine fractures

death (%)

in the US. The number of hospitalizations associated with

cervical spine fractures continues to increase with a con-

J Neurosurg: Spine / Volume 13 / July 2010 63


Fig. 2.  Hospitalizations in SCI and non-SCI groups from 1997 to
2006.
comitant increase in charges, but outcomes at discharge
remain unchanged. Three findings, however, deserve spe-
cial attention.
First, the SCI and non-SCI populations are vastly dif-
ferent in terms of epidemiology, treatment, and outcomes.
It is well established that cases of SCI are associated with
increased costs, longer hospitalizations, and worse out-
comes than cases not involving SCI.19 The data retrieved
from the NIS corroborate these facts. Compared with the
non-SCI group, the SCI group was associated with about
a 3-fold increase in hospital charges, LOS, and adverse
discharge outcomes (adverse outcome defined as non-
routine discharge). The 10-year average rate of in-house
mortality in the SCI group was 12% while published in-
hospital mortality rates (for SCI in general) have ranged
from 6%18 to 34%.16
Second, we noted a dramatic and disproportionate in-
crease in hospitalizations and charges associated with the
care of patients with non-SCI fractures compared with
those with SCIs. Are the increases in these cases due to
improved diagnosis? The increased use of CT scanning in
the emergency department setting, particularly in trauma,
is likely one factor leading to increases in the detection
of spinal fractures.2,14 Another hypothesis could be that
because more spinal fractures are being detected in the
Fig. 3.  Length of stay in SCI and non-SCI groups from 1997 to
2006.
64
A. A. Baaj et al.
Fig. 4.  Hospital charges in SCI and non-SCI groups from 1997 to
2006.
emergency department, more are admitted for observa-
tion, follow-up imaging, or surgical intervention, thus
increasing costs. In a landmark article on utilization of
spinal fusion in the US, one involving the use of NIS data,
Cowan et al.4 reported a 31% increase in spinal fusion for
traumatic cervical spine injuries between 1993 and 2001.
Do the increased charges associated with cervical spine
fracture hospitalizations stem from an increase in surgi-
cal interventions, or are there other causes? Further stud-
ies are needed to determine the answer. What is clear,
however, is that the increased rate of surgical intervention
in the cervical spine is consistent with published reports
on trends of spinal surgery and fusion in general.6
Third, we observed the apparent lack of improvement
in status at discharge in both groups. During the 10 years
examined, no significant changes were observed in LOS,
in in-hospital mortality rate, or in the need for skilled
nursing facility after discharge despite considerable prog-
ress in understanding SCI and advances in surgical treat-
ments. Across the entire period analyzed, routine dis-
charge was consistently about 60% for the non-SCI group
and less than 30% for the SCI group. Approximately
50% of patients with SCI still needed care at a skilled
nursing facility after discharge. One explanation is that
Fig. 5.  In-house mortality rates in SCI and non-SCI groups from
1997 to 2006.
J Neurosurg: Spine / Volume 13 / July 2010
Cervical spine fractures in US
Table 3: Ten-year averages between SCI and non-SCI groups

Patient Group
Variable w/o SCI w/ SCI
no. of hospitalizations 16045.5 3741.2
LOS (days) 6.2 18.1*
charges ($) 28,702 98,393*
national bill ($) 479,402,151 375,424,686
routine D/C (%) 60.8 25.0*
rehab/SNF (%) 25.0 50.5*
in-house death (%) 2.6 11.6*
home health (%) 6.9% 3.5*

Fig. 6.  Discharges to skilled nursing facility (SNF) in SCI and non- *  Statistically significant (p ≤ 0.001).
SCI groups from 1997 to 2006.

the cause of fractures cannot be ascertained. Although

comorbidities may have been associated with these hos-
pitalizations, including multiple traumatic injuries, and
contributed to the overall rate of morbidity and mortality.
Another explanation is that improvements in neurological
status may not be realized for several weeks if not months
after injury. Such improvements would not be detected in
this type of in-patient database analysis. Clinical studies
with longer follow-up are needed to assess the long-term
outcomes in this subset of patients.
The finding that C-2 was the most frequently in-
volved level in the non-SCI group is consistent with other
reports.5,25 The high incidence of C-2 involvement is like-
ly to reflect the high frequency of dens fractures in the
elderly. That young males account for most of the SCI
admissions is also well established.13,16,19 The consistency
between such findings in the database and published re-
ports lends support to the utility and validity of national
estimates from the HCUPnet database.
The limitations of the NIS and similar databases
are well documented and have been addressed previous-
ly.1,3,11,21 The main disadvantages of such analyses are the
lack of postdischarge follow-up and the lack of inpatient
outcome metrics more detailed than discharge status.
Coding accuracy also may be a limitation. Furthermore,
most injuries are likely traumatic, some pathological frac-
tures may be coded in a similar fashion. Therefore, there
is no way to distinguish among similar diagnoses that
share the same generic code. Finally, the results of the
statistical analysis must be interpreted with caution and
considered only as an evaluation of trends. It is possible
that the trends observed at the summary level would dif-
fer in some way if individual data were examined.
The advantages of using large-scale databases in-
clude avoiding selection bias and providing a means to
study trends on a national level. Single-institution studies
may provide long-term and postdischarge follow-up data,
but state and national databases afford the opportunity to
study large samples of patients and are particularly useful
for trend analyses. As of the completion of this study, at
least 40 published studies that used NIS data for analy-
sis in neurosurgery were identified on PubMed, and at
least 8 of those studies were spine related.4,6,10,12,15,20,21,24
Familiarity with databases like the NIS is critical because
they are often used by policy-makers to review the cost
effectiveness of interventions with implications for reim-
bursement.1
Table 4: Frequency of fractured levels in SCI and non-SCI
groups

Fracture (level) Frequency (%)

w/o SCI
C-1 15
C-2 38
C-3 4
C-4 5
C-5 10
C-6 13
C-7 10
multilevel 6
w/ SCI
C1–4 40
Fig. 7.  Routine discharges in SCI and non-SCI groups from 1997 C5–7 60
to 2006.

J Neurosurg: Spine / Volume 13 / July 2010 65


Conclusions
During the study period, increases in hospitalizations
and charges were observed in both the SCI and non-SCI
groups. The percentage increase was higher for the non-
SCI group. The SCI group was associated with a 3-fold in-
crease in adverse discharge outcomes. Despite increased
charges, there were no significant improvements in im-
mediate discharge status in the two groups during the 10
years analyzed. Further studies are warranted to deter-
mine the cause of increased hospitalizations (particularly
with non-SCI fractures) and ways to improve immediate
discharge outcomes.
Disclosure
The authors report no conflict of interest concerning the mate-
rials or methods used in this study or the findings specified in this
paper.
Acknowledgments
The authors thank Ms. Katheryne Downes from the University
of South Florida for assistance with the statistical analysis. They also
thank the Neuroscience Publications Office at Barrow Neurological
Institute for editorial assistance.
References
  1.  Andaluz N, Zuccarello M: Recent trends in the treatment of
spontaneous intracerebral hemorrhage: analysis of a nation-
wide inpatient database. J Neurosurg 110:403–410, 2009
  2.  Antevil JL, Sise MJ, Sack DI, Kidder B, Hopper A, Brown
CV: Spiral computed tomography for the initial evaluation of
spine trauma: a new standard of care? J Trauma 61:382–387,
2006
  3.  Baaj AA, Benbadis SR, Tatum WO, Vale FL: Trends in the use
of vagus nerve stimulation for epilepsy: analysis of a nation-
wide database. Neurosurg Focus 25(3):E10, 2008
  4.  Cowan JA Jr, Dimick JB, Wainess R, Upchurch GR Jr, Chan-
dler WF, La Marca F: Changes in the utilization of spinal fu-
sion in the United States. Neurosurgery 59:15–20, 2006
  5.  Damadi AA, Saxe AW, Fath JJ, Apelgren KN: Cervical spine
fractures in patients 65 years or older: a 3-year experience at a
level I trauma center. J Trauma 64:745–748, 2008
  6. Deyo RA, Gray DT, Kreuter W, Mirza S, Martin BI: United
States trends in lumbar fusion surgery for degenerative condi-
tions. Spine 30:1441–1447, 2005
  7.  Fehlings MG, Sekhon LH: Acute interventions in spinal cord
injury: what do we know, what should we do? Clin Neuro-
surg 48:226–242, 2001
  8.  Fehlings MG, Sekhon LH, Tator C: The role and timing of
decompression in acute spinal cord injury: what do we know?
What should we do? Spine 26 (24 Suppl):S101–S110, 2001
  9.  Frangen TM, Zilkens C, Muhr G, Schinkel C: Odontoid frac-
tures in the elderly: dorsal C1/C2 fusion is superior to halo-
vest immobilization. J Trauma 63:83–89, 2007
10.  Gray DT, Deyo RA, Kreuter W, Mirza SK, Heagerty PJ, Com-
stock BA, et al: Population-based trends in volumes and rates
of ambulatory lumbar spine surgery. Spine 31:1957–1964,
2006
66
A. A. Baaj et al.
11.  Lad SP, Patil CG, Berta S, Santarelli JG, Ho C, Boakye M:
National trends in spinal fusion for cervical spondylotic mye-
lopathy. Surg Neurol 71:66–69, 2009
12.  Lad SP, Patil CG, Lad EM, Boakye M: Trends in pathological
vertebral fractures in the United States: 1993 to 2004. J Neu-
rosurg Spine 7:305–310, 2007
13.  Leucht P, Fischer K, Muhr G, Mueller EJ: Epidemiology of
traumatic spine fractures. Injury 40:166–172, 2009
14.  Mathen R, Inaba K, Munera F, Teixeira PG, Rivas L, McKen-
ney M, et al: Prospective evaluation of multislice computed
tomography versus plain radiographic cervical spine clear-
ance in trauma patients. J Trauma 62:1427–1431, 2007
15.  Patil PG, Turner DA, Pietrobon R: National trends in surgical
procedures for degenerative cervical spine diease: 1990-2000.
Neurosurgery 57:753–758, 2005
16.  Pickett W, Simpson K, Walker J, Brison RJ: Traumatic spi-
nal cord injury in Ontario, Canada. J Trauma 55:1070–1076,
2003
17.  Rath SA, Moszko S, Schäffner PM, Cantone G, Braun V, Rich-
ter HP, et al: Accuracy of pedicle screw insertion in the cer-
vical spine for internal fixation using frameless stereotactic
guidance. J Neurosurg Spine 8:237–245, 2008
18.  Saunders LL, Selassie AW, Hill EG, Nicholas JS, Varma AK,
Lackland DT, et al: Traumatic spinal cord injury mortality,
1981-1998. J Trauma 66:184–190, 2009
19.  Sekhon LH, Fehlings MG: Epidemiology, demographics, and
pathophysiology of acute spinal cord injury. Spine 26 (24
Suppl):S2–S12, 2001
20.  Shamji MF, Cook C, Pietrobon R, Tackett S, Brown C, Isaacs
RE: Impact of surgical approach on complications and re-
source utilization of cervical spine fusion: a nationwide per-
spective to the surgical treatment of diffuse cervical spondy-
losis. Spine J 9:31–38, 2009
21.  Shamji MF, Cook C, Tackett S, Brown C, Isaacs RE: Impact
of preoperative neurological status on perioperative morbidity
associated with anterior and posterior cervical fusion. J Neu-
rosurg Spine 9:10–16, 2008
22. United States Census Bureau: Population Estimates 2009.
http://www.census.gov/popest/national/national.html
23. U.S. Department of Health and Human Services: Heathcare
Cost and Utilization Project (HCUP) Facts and Figures,
2006: Statistics on Hospital-based Care in the United States.
(http://www.hcup-us.ahrq.gov/reports/factsandfigures/facts_
figures_2006.jsp ) [Accessed March 24, 2010]
24.  Wang MC, Chan L, Maiman DJ, Kreuter W, Deyo RA: Com-
plications and mortality associated with cervical spine sur-
gery for degenerative disease in the United States. Spine 32:
342–347, 2007
25.  Weller SJ, Malek AM, Rossitch E Jr: Cervical spine fractures
in the elderly. Surg Neurol 47:274–281, 1997
Manuscript submitted June 15, 2009.
Accepted March 3, 2010.
Address correspondence to: Fernando L. Vale, M.D., Department
of Neurosurgery, University of South Florida, 2 Tampa General
Circle, USF Health, 7th Floor, Tampa, Florida 33606. email: fvale@
health.usf.edu.
J Neurosurg: Spine / Volume 13 / July 2010