1740
A Demographic Profile of New Traumatic Spinal Cord
Injuries: Change and Stability Over 30 Years
Amie B. Jackson, MD, Marcel Dijkers, PhD, Michael J. DeVivo, DrPH, Robert B. Poczatek, MD
ABSTRACT. Jackson AB, Dijkers M, DeVivo M, Poczatek
RB. A demographic profile of new traumatic spinal cord injuries:
change and stability over 30 years. Arch Phys Med Rehabil 2004;
85:1740-8.
Objective: To evaluate epidemiologic trends in new spinal
cord injuries (SCIs) in the United States over 3 decades.
Design: Consecutive case series.
Setting: Model Spinal Cord Injury Systems (MSCIS)
facilities.
Participants: Persons (N⫽30,532) admitted to MSCIS fa-
cilities within 365 days of injury between 1973 and 2003, and
enrolled in the National Spinal Cord Injury Database.
Interventions: Not applicable.
Main Outcome Measures: Data were collected at MSCIS
admission and rehabilitation discharge. Variables included age,
gender, race and ethnic group, year of injury, and level and
extent of injury. Specific etiologies were grouped as motor
vehicle collisions (MVCs), violence, falls, sports, and other.
Demographic and injury severity trends were analyzed by year
of injury groupings according to decades (1973–1979, 1980 –
1989, 1990 –1999, 2000 –2003.) Chi-square tests assessed sta-
tistical significance. One-way analysis of variance compared
mean ages.
Results: The male/female ratio remained fairly stable at 4:1,
but the percentage of women increased slightly over time,
especially from MVC etiologies (P⬍.001). Over time, the
mean age at injury increased significantly (P⬍.001); it was
37.7⫾17.5 years in 2000 –2003. The majority of cases were
white (66.1%). Tetraplegia (54.1%) and complete injuries
(55.6%) occurred more than paraplegia and incomplete inju-
ries, respectively. MVCs (45.6%) remained the most common
etiology; falls (19.6%) held the second position over violence
(17.8%), except for the 1990 –1999 period when the positions
were reversed. Significantly increasing percentages of new
injuries were seen for SCI due to automobile, motorcycle,
bicycle, and all-terrain vehicle crashes, blunt object attacks,
snow skiing, and medical and surgical mishaps.
Conclusions: Many previously seen SCI demographic
trends continued into the 2000 decade.
Key Words: Demography; Epidemiology; Rehabilitation;
Spinal cord injuries.
© 2004 by the American Congress of Rehabilitation Medi-
cine and the American Academy of Physical Medicine and
Rehabilitation
From the University of Alabama at Birmingham, Birmingham, AL (Jackson,
DeVivo, Poczatek); and Mount Sinai School of Medicine, New York, NY (Dijkers).
Supported by the National Institute on Disability and Rehabilitation Research,
Office of Special Education and Rehabilitation Services, US Department of Education
(grant nos. H133N000016, H133N000027, H133A011201).
No commercial party having a direct interest in the results of the research support-
ing this article has or will confer a benefit on the author(s) or on any organization with
which the author(s) is/are associated.
Reprint requests to Amie B. Jackson, MD, 1717 6th Ave S, Spain Rehabilitation
Center, University of Alabama, Birmingham, AL 35233, e-mail: Jacksona@uab.edu.
0003-9993/04/8511-8893$30.00/0
doi:10.1016/j.apmr.2004.04.035
Arch Phys Med Rehabil Vol 85, November 2004
ince its inception in 1973, the facilities participating in the
S Model Spinal Cord Injury Systems (MSCIS) program have
collected valuable data that have been used by professionals in
the clinical, research, engineering, and legal fields to both
improve the quality of life in people with spinal cord injury
(SCI) and to aid in the prevention of SCI in those populations
most at risk. Collectively, the epidemiologic and other data
from 24 designated Model Systems are housed in the National
Spinal Cord Injury Database (NSCID).1 Sixteen of these sys-
tems are currently funded by the National Institute on Disabil-
ity and Rehabilitation Research. They submit data on newly
injured persons who present for admission to the MSCIS, and
they provide follow-up data on cases submitted in prior years.
In 1976, the following eligibility criteria were established for
inclusion of people with SCI in the NSCID: (1) admission to an
MSCIS of care within 365 days of injury; (2) place of injury
and residence within the identified catchment area for the
MSCIS; (3) clinically discernible degree of spinal cord neuro-
logic impairment on admission; (4) continual hospitalization
from injury to MSCIS admission, except for brief periods no
longer than normally accepted as a therapeutic leave of ab-
sence; (5) discharge from the MSCIS as either neurologically
recovered, completed rehabilitation, or died; and (6) signed
informed consent.
Throughout the years, the database has evolved into 2 spe-
cific data sets. The Form I data set collects information on all
people admitted within 60 days of injury who also satisfy the
eligibility criteria. Demographic data, rehabilitation complica-
tions, and status at discharge are prospectively recorded for all
Form I entries. For those people admitted to the MSCIS within
24 hours, more extensive acute-care variables are collected.
These data include factors such as neurologic status on admis-
sion, length of stay, medical complications, and surgical
events. The Registry data set was implemented in 1987 to
include people entering the MSCIS between 61 and 365 days
after injury. Only limited demographic and rehabilitation data
were collected and no long-term follow-up. Many analyses and
publications of these data from MSCIS participating in the
NSCID have been recognized both nationally and internation-
ally for providing demographic data on people with SCI.2
The incidence rate of SCI has been estimated to be from 15
to 40 cases per million population worldwide, with the United
States at the high end of those estimates.3 Given the current
size of the US population, this translates to approximately
11,000 new cases of SCI each year. Prevalence of SCI in the
United States is now approaching 250,000 people and is in-
creasing as a result of increasing survival rates.4,5
As demonstrated in the most recent comprehensive analysis
of demographic and injury trends in persons treated in the
MSCIS conducted by Nobunaga et al6 in 1999, the face of SCI
is changing. Based on available data from 1973 through 1998,
significant trends toward older age at the time of injury, in-
creasing proportions of injuries occurring in racial and ethnic
minority populations, as well as increasing proportions of in-
juries caused by acts of violence and falls, with concomitant
decreasing proportions of injuries because of motor vehicle
crashes (MVCs) and sports, were observed. A slight trend
 EPIDEMIOLOGY OF SPINAL CORD INJURY, Jackson
toward an increasing proportion of women with new SCI was
also observed.5
Since 1998, the NSCID has grown by approximately 10%.
Considering the pace of previously documented changes in the
profile of new cases of SCI and the need to base management
and prevention programs on the most accurate information, it is
necessary to reassess the epidemiology of SCI at timely inter-
vals. It is to this end that the current study was undertaken.
METHODS
At the time of this study, the NSCID housed records for
30,532 persons admitted to a MSCIS within 365 days of injury
between 1973 and March 2003. Eligibility criteria and changes
in those criteria over time have been described previously.1,7
The NSCID contains information on new injuries in 2 sep-
arate data sets. The Registry data set contains very limited
demographic and injury severity information on persons treated
at the MSCIS for whom follow-up data will not be collected.
Persons enrolled in the Registry data set typically meet some
but not all of the eligibility criteria for enrollment in the full
database.1,7 The Form I data set contains a much larger set of
admission and discharge medical, functional, and social vari-
ables for persons who will be followed for additional data
collection after MSCIS discharge. Procedures to ensure the
quality and consistency of data collected by the MSCIS have
been documented in detail previously.1,7 Because all the data
necessary to conduct this study are contained in both the
Registry and Form I data sets, we created a combined data set
for analysis, consistent with the approach taken by Nobunaga
et al.6 Combining data sets not only enhances sample size, but
also enhances the representativeness of the study population.
Neurologic level and extent of injury were defined using the
international standards set forth by the American Spinal Injury
Association (ASIA).8 The neurologic level of injury was de-
fined as the most caudal segment of the spinal cord with normal
sensory and motor function on both sides of the body. Tetra-
plegia was defined as impairment or loss of motor and/or
sensory function in a cervical segment of the spinal cord caused
by damage of neural elements within the spinal canal. Paraple-
gia was defined as impairment or loss of motor and/or sensory
function in a thoracic, lumbar, or sacral segment of the spinal
cord following damage to the spinal cord.8
Prior to 1992, completeness of injury was documented in the
database using Frankel grades.9 Since that time, the ASIA
Impairment Scale has been used to document completeness of
injury.8 For purposes of the present study, neurologic extent of
injury was categorized as either complete or incomplete. A
complete injury was defined as absence of sensory and motor
function in the lowest sacral segments.10 An incomplete injury
was defined as an SCI in which there is partial preservation of
sensory and/or motor function (ⱖ3 segments) below the level
of injury and including the lowest sacral segments.8
Ventilator dependency was defined as requiring ventilatory
assistance on a daily basis but not necessarily 24 hours a day.
Persons with phrenic pacers are included as ventilator depen-
dent, but inspiratory positive pressure breathing, continuous
positive airway pressure, and bilevel positive airway pressure
users are not considered to be ventilator dependent.
Etiology of injury information was categorized into 5
groups. MVCs included crashes of autos, busses, jeeps, trucks,
motorcycles, boats, fixed and rotating wing aircraft, snowmo-
biles, bicycles, all terrain vehicles (ATVs), tractors, bulldozers,
go-carts, steamrollers, trains, and forklifts, as well as pedestri-
ans hit by vehicles. Acts of violence included gunshot wounds
(self-or other-inflicted, accidental), stab wounds, being hit by
blunt objects such as baseball bats, falls resulting from inten-
1741
tional pushes, and explosions. Falls included falls from heights
as well as falls on level ground, such as tripping over an object
or slipping on a wet surface. Sports included diving, football,
trampoline, snow skiing, water skiing, wrestling, baseball, bas-
ketball, surfing, horseback riding, gymnastics, rodeo, track,
field sports, hang gliding, air sports, skateboarding, ice hockey,
snow boarding, and other organized sporting events. The final
etiology category included all other causes of SCI, consisting
mostly of being hit by a falling or flying object, and medical or
surgical complications.
To evaluate trends, all cases were grouped by decade of
injury (1973–1979, 1980 –1989, 1990 –1999, 2000 –2003).
Mean age at injury was determined for each time period and
compared by 1-way analysis of variance. For categoric data
such as gender, race, level of injury, extent of injury, and
ventilator dependency, frequencies and percentages were ex-
pressed in contingency tables and differences were assessed for
statistical significance by the chi-square test. Probability values
less than .05 were considered statistically significant.
RESULTS
Over the last 30 years, the MSCIS facilities have admitted
30,532 people who sustained a traumatic SCI and otherwise
satisfied Registry or Form I dataset criteria (table 1). The
average was over 1100 cases per year for the period from 1980
onward. The average was lower in the 1970s when the number
of MSCIS facilities was small but expanding. Because data for
the years following 2000 are limited to 4.75 years (a cutoff of
data entry was made after 9mo into 2003), it is too early to
determine whether this average will continue.
General Demographics
Gender. Over the period studied, about 4 times as many
men sustained an SCI as women. The yearly trends revealed a
statistically significant (P⬍.000) increase of the observed pro-
portion of women (table 1). In the 1970s, women made up
18.2% of the injuries, whereas, in the 2000s, they comprise
21.8%. However, as suggested by figure 1, this increase is
limited to a slight increase in the percentage of women injured
in traffic crashes (the largest etiologic group) and sports and a
strong increase in the percentage of women injured because of
“other” causes.
Age. For all patients enrolled, the average age at injury was
just over 33 years (table 1). Examination of trends from 1973
to 2003 demonstrate a significantly (P⬍.000) higher mean age
at time of injury for each successive decade. The gradual
increase in mean age at injury was found within all etiology
groups, except violence.
Race and ethnicity. Whites make up the largest percentage
of persons enrolled in the database, followed by African Amer-
icans and Hispanics (table 1). Over the last 30 years, the
percentage of whites has declined, while the percentages in all
other racial and ethnic groups has increased, at least until 2003.
Level of injury. Based on the level of injury at time of
rehabilitation discharge, more cervical injuries (54.1%) were
observed to occur throughout all time periods than thoraco-
lumbo-sacral injuries (45.2%). Moreover, there has been a
small but statistically significant trend toward an increasing
percentage of cervical injuries in recent years, ranging from
53.5% in the 1970s to 56.5% since 2000 (P⬍.000).
ASIA grade. ASIA grade A (complete) injuries (55.6%)
continued to occur more frequently than incomplete injuries
(44.4%) (table 2). Among the latter group, 9.1% were ASIA
grade B, 9.8% were ASIA grade C, 24.8% were ASIA grade D,
and 0.7% were ASIA grade E. Figure 2 shows the trends of
Arch Phys Med Rehabil Vol 85, November 2004
1742 EPIDEMIOLOGY OF SPINAL CORD INJURY, Jackson

Table 1: Characteristics of Persons With New SCI Over 4 Time Periods

Time Periods

Variables 1973–1979 1980–1989 1990–1999 2000–2003 Total

Total cases 4565 10,270 12,552 3145 30,532

Percentage of total 15.0 33.6 41.1 10.3 100.0
Sex
Male 81.8 82.4 80.2 78.2 81.0
Female 18.2 17.6 19.8 21.8 19.0
Age group (y)
0–15 6.4 3.7 3.1 2.0 3.7
16–30 61.9 57.6 45.7 40.7 51.6
31–45 17.9 22.1 27.3 27.3 24.1
46–60 9.0 9.9 13.5 17.9 12.1
65⫹ 4.8 6.6 10.4 12.1 8.5
Mean age 28.9 31.4 35.3 38.0 33.3
Standard deviation 20.2 17.9 20.7 24.5 20.4
Median 24 26 31 34 28
Racial/ethnic group
White 76.8 68.2 59.9 67.4 66.0
African American 14.2 20.4 24.2 19.4 21.0
Hispanic-Latino 6.0 8.3 12.0 10.1 9.7
Other* 2.9 3.1 4.9 3.2 3.5
Etiology category
MVC 48.7 46.2 42.7 50.4 45.6
Violence 13.3 16.9 21.8 11.2 17.8
Sports 14.4 12.4 8.5 9.0 10.7
Falls 16.5 18.5 20.7 23.8 19.6
Other 7.2 6.1 6.4 5.5 6.3

NOTE. Values may not equal 100% because of rounding.

*Includes Native American, Eskimo, Asian-Pacific islander, and unknown.

complete versus incomplete SCI for the entire time period. The
percentage of complete injuries rose substantially during the
1990s but has since dropped below pre-1990 levels and now
constitutes slightly less than half (48.8%) of all new injures
(P⬍.000).
Neurologic classification. Combining ASIA grade and
level of injury, we defined 4 groups: paraplegia incomplete,
paraplegia complete, tetraplegia incomplete, and tetraplegia
complete. The percentage in each category has changed some-
what over the years (fig 3), and decade-to-decade differences
are statistically significant (P⬍.000). However, there was no
Fig 1. Percentage of women by time period and etiology.
trend toward systematic increase or decrease of any category.
For all years combined (table 2), the tetraplegia incomplete
group has the greatest percentage of subjects (30.6%), followed
by paraplegia complete (26.1%), tetraplegia complete (23.4%),
paraplegia incomplete (19.2%), and normal (0.7%).
Ventilator dependency. In the 1973–1979 period, only
2.3% of people with SCI were discharged requiring mechanical
ventilation, whereas 6.8% required ventilation on discharge
between 2000 and 2003. This trend was statistically significant
(P⬍.000).
Etiology of Injury
Over the 30 years of data collection, MVC has been the most
common cause of SCI. Falls (19.6%) have been the second
most common cause, followed by violence, sports, and other
etiologies, respectively. These rankings have been fairly con-
stant over the decades, with the exception of the 1990 –1999
period when the second most common cause of SCI was
violence (table 1).
Analyses of etiology of injury by demographic and neuro-
logic factors were also conducted (table 2). Men have a 4:1
representation in all etiology groups with the exception of
MVC where the male to female ratio is 3:1. In this database,
African Americans make up the largest (52.2%) and Hispanics
the third largest (21.1%) percentage of people with violence
etiologies. Whites are the largest group for all other etiologies:
MVC (74.8%), sports (87.2%), falls (69.8%), and other
(75.6%). MVCs, sports, and falls were significantly more likely
to result in injuries in the cervical region of the spinal cord,
whereas violence and other etiologies were significantly more
likely to result in injuries to the thoracic, lumbar, and sacral
regions of the spinal cord (P⬍.000). As stated above, more

Arch Phys Med Rehabil Vol 85, November 2004

 EPIDEMIOLOGY OF SPINAL CORD INJURY, Jackson 1743

Table 2: All Time Periods Combined: Percentages Within Demographic Variables of Persons With New SCI, by Etiology Category
Etiology Category

MVC Violence Sports Falls Other Total

Variables (%) (%) (%) (%) (%) (%)

Sex*
Male 75.9 86.9 89.5 82.0 82.7 80.9
Female 24.1 13.1 10.5 18.0 17.3 19.1
Race/ethnic group*
White 74.8 23.9 87.2 69.8 75.6 66.1
African American 13.6 52.2 6.5 19.6 15.0 21.0
Hispanic 7.8 21.1 4.8 7.8 7.1 9.8
Other 3.8 2.8 1.5 2.8 2.3 3.1
Level of injury*
Tetraplegia 57.0 31.6 88.3 54.8 36.1 54.1
Paraplegia 42.3 68.1 11.2 44.1 63.3 45.2
Normal 0.7 0.3 0.6 1.1 0.6 0.7
Completeness of injury (ASIA grade)*
A 56.5 66.1 54.6 45.7 49.8 55.6
B 9.4 6.2 14.1 8.1 9.1 9.1
C 9.0 9.1 7.5 12.0 15.0 9.8
D 24.4 18.4 23.3 32.9 25.4 24.8
E 0.7 0.3 0.6 1.2 0.7 0.7
Neurologic class*
PI 16.7 25.6 5.7 22.0 33.2 19.2
PC 25.7 42.5 5.7 22.1 29.9 26.1
TI 32.3 13.7 44.5 37.2 22.9 30.6
TC 24.7 18.0 43.5 17.7 13.3 23.4
Normal 0.7 0.2 0.6 1.1 0.6 0.7

NOTE. Total number of cases N⫽30,496. Values may not equal 100% due to rounding.
Abbreviations: PC, paraplegia incomplete; PI, paraplegia incomplete; TC, tetraplegia complete; TI, tetraplegia incomplete.
*Significant change observed within categories over time (P⬍.001).

than half (55.6%) of SCIs are ASIA grade A (complete). Each
etiology category, however, has varying percentages of ASIA
grades. For example, approximately 60% of violence, MVC,
and sports etiologies resulted in complete injuries, whereas
slightly less than half of injuries in the falls and other groups
were complete injuries. The second most common ASIA grade
for all etiology groups was ASIA grade D (24.8% of total).
Finally, of the 4 neurologic classification groups, tetraplegia
incomplete was the most common type to occur as a result of
MVC, sports, and falls (P⬍.000). Sports injuries also produced
a very high percentage of tetraplegia complete cases, mostly
because of diving accidents. Paraplegia complete occurred
significantly (P⬍.000) more in the violence and in the other
group.
The mean age at time of injury varied greatly by etiology
groups (fig 4). People who sustained an SCI from sports were
the youngest, followed (in increasing order) by those who
sustained their injury from violence, MVC, other causes, and
falls. This rank order has remained stable over the last 30 years.
The distribution of etiology of injury by age was analyzed
further. Table 3 reconfirms the previously reported differences

Fig 3. Percentage of neurologic classification groups at MSCIS dis-

charge for each decade and total. Abbreviations: PC, paraplegia
Fig 2. Percentage of complete and incomplete SCI for each de- complete; PI, paraplegia incomplete; TC, tetraplegia complete; TI,
cade. tetraplegia incomplete.

Arch Phys Med Rehabil Vol 85, November 2004

1744 EPIDEMIOLOGY OF SPINAL CORD INJURY, Jackson

Fig 4. Mean age by decade and etiology.

Fig 5. Trends of MVC frequency among age groups.

in etiologies of injury for specific age groups.6 MVCs were the

most common etiology for those age 60 years or younger,
whereas falls were the most common cause of injury for ages
over 60 years. The second most common etiology category
varied greatly among the age groups: among those 0 to 15
years, it was sports; among those 16 to 30 years, it was
violence; and among those from 31 to 45 and 46 to 60 years, it
was falls. The second ranked etiology within the oldest group
was MVCs.
A statistically significant increase existed in the percentage
of cases because of MVCs, and there was a decrease in the
percentage of cases of violent etiology over time among the age
groups 0 to 15 (P⫽.010), 31 to 45 (P⬍.000), and 46 to 60
(P⬍.000) years. Among people between the ages of 16 and 30
years, the percentage of cases because of violence peaked in
the 1990s at 32.7% and has decreased to 19.5% since 2000
(P⬍.000). In the oldest age group, the percentage of cases
because of violence has declined steadily from approximately
6% in the 1970s to less than 1% since 2000, whereas the
percentage of cases because of falls has increased concomi-
tantly, reaching 58% among those over the age of 61 years (figs
5, 6).
Injuries From Specific Etiologies
Specific etiologies of SCI are in table 4. Overall, there were
several statistically significant trends in etiologies over the last
30 years.
Motor vehicle collisions. Table 4 provides the percentage
of people who incurred an SCI because of each specific type of
traffic accidents from 1973 to 2003. These data revealed trends
of increasing percentages of MVCs from crashes involving
automobiles, motorcycles, bicycles, and ATVs relative to all
other causes of injury (P⬍.000). The ATV category was added
to the database in 1986; before that time, these cases were
classified as other vehicle.
Violence. The data show a statistically significant
(P⬍.000) increase in the percentage of SCIs from gunshot
wounds, stabbings, and blunt object attacks through the 1990 –
1999 period, followed by a decrease. The number of persons
injured because of explosions was too small to show trends.
Sports. Within the category of sports, diving was the most
common etiology over all decades, but the percentage of all
injuries that resulted from diving has steadily decreased
(P⬍.000). Football also saw a decrease (except for the most
recent years), but snow skiing saw an increase. As a group, the
percentage of cases because of all sports (except diving) com-
bined has remained fairly stable over time.
Falls. The database does not distinguish between different
types of falls such as those from a height versus level ground.
The percentage of injuries resulting from falls increased with
each time period (P⬍.000).
Other. The major subcategories in the other etiology group
are injuries from medical or surgical complications or from
unclassified causes. The percentage of SCIs from these causes
has significantly increased over time (P⬍.001).
DISCUSSION
With an estimated incidence around 40 cases per million,
SCI is a fairly uncommon cause of disability. However, its
costs, both economic and social, are out of proportion to its
frequency.11-14 It is thus important to review the database
trends to gain understanding of the ramifications of this trau-
matic event. This study examines the significant trends over the
past 30 years and suggests how these correspond with the
trends in the general population.

Table 3: Etiology of Injury by Age Group

Age Group

0–15 16–30 31–45 46–60 61–75 76–99

Etiology Years Years Years Years Years Years Total

MVC 44.0 47.8 47.4 41.1 35.1 29.8 45.6

Violence 21.5 23.0 16.3 7.7 3.0 1.1 17.8
Sports 23.3 14.6 7.3 3.8 1.9 0.5 10.7
Falls 8.3 10.8 21.9 35.5 46.4 59.7 19.6
Other 2.9 3.9 7.1 11.8 13.5 8.9 6.3
Total 100.0 100.1 100.0 99.9 99.9 100.0 100.0
Total cases 1136 15,743 7357 3684 1947 629 30,496

NOTE. Values are in percentage; total cases are N. Values may not equal 100% because of rounding.

Arch Phys Med Rehabil Vol 85, November 2004

 EPIDEMIOLOGY OF SPINAL CORD INJURY, Jackson 1745

rather than behavioral change would seem to underlie an in-

Fig 6. Trends of violence frequency among age groups.
One significant trend is the change in the percentage of
patients who are women. When the database was first set up,
the “women’s revolution” was first starting to affect large
numbers of women, but most women still fulfilled “traditional”
roles. In the years since, they have been employed in high-risk
jobs (eg, construction), participated in high-risk sports (eg,
stock car racing), and started other endeavors like gang mem-
bership that bring the threat of injury, including SCI. These role
changes have had a minor impact on the make-up of the SCI
patient population (see table 1), and figure 1 confirms that. The
only etiology category with a major change is other. Biologic
crease in more falls and medical and surgical mishaps, because
these are typical of older individuals, and women are a high
percentage of the elderly population. It may be that the move
by women into high-risk activities is still too limited to have
more than a marginal impact on statistics; alternatively, the
women engaged in these activities may perform them more
safely than men, thus exposing themselves to lower risk of
injury.
From 1973 through 1999, the percentage of SCI cases who
were members of a racial or ethnic minority expanded contin-
uously, paralleling the composition of the United States pop-
ulation. The apparent reversal of the trend in 2000 and later
years may be the effect of the end of contributions to the
database by 2 fairly large MSCIS with a large minority clien-
tele. Alternatively, it may reflect the end of the “epidemic” of
violence-onset SCI; as is clear from table 2, this epidemic
mostly affected African Americans and Hispanics. With a
sharp reduction in SCI because of violence, the “normal”
distribution over the ethnic and racial groups may have been
restored. Further analysis, involving more complete data for the
most recent years, is needed to determine the causes of this
change. However, given the US Census Bureau data on the
continued growth of minorities, especially Hispanics (it has
been predicted that in 2050 whites will be in the minority), an
ever-increasing minority SCI clientele should be expected by

Table 4: Time Trends for Most Frequent Specific Etiologies Within Major Etiology Groups
Time Period

Etiology 1973–1979 1980–1989 1990–1999 2000–2003 Total

MVC
Auto 38.1 35.9 33.6 37.2 35.4
Motorcycle 6.5 5.8 4.3 6.9 5.4
Pedestrian 1.7 1.7 1.8 1.3 1.7
Bicycle 0.8 1.0 1.1 2.1 1.1
Other vehicle 0.5 0.8 0.5 0.6 0.6
ATV 0.0 0.3 0.6 1.4 0.5
Fix-wing aircraft 0.6 0.4 0.2 0.3 0.3
Snowmobile 0.1 0.1 0.3 0.4 0.2
Other* ⱕ0.2 ⱕ0.2 ⱕ0.2 ⱕ0.2 ⱕ0.2
Violence
Gunshot 12.0 14.8 19.8 9.8 15.9
Stab 0.6 1.2 0.7 0.4 0.8
Blunt object 0.6 0.9 1.2 0.9 1.0
Sports
Diving 9.5 8.2 5.0 3.9 6.6
Snow ski 0.3 0.6 0.7 0.9 0.6
Football 1.1 0.6 0.3 0.5 0.5
Surfing 0.3 0.5 0.4 0.5 0.4
Horse 0.2 0.4 0.4 0.5 0.4
Wrestling 0.4 0.3 0.3 0.3 0.3
Trampoline 0.6 0.2 0.1 0.3 0.2
Gymnastics 0.3 0.3 0.1 0.1 0.2
Other† ⱕ0.2 ⱕ0.2 ⱕ0.2 ⱕ0.2 ⱕ0.2
Falls 16.5 18.5 20.7 23.8 19.6
Other
Falling object 5.2 3.0 2.6 2.8 3.1
Medical/surgery 1.2 1.9 2.9 2.1 2.2
Unclassified 0.9 1.2 0.9 0.7 1.0
Total cases 4562 10,263 12,563 3140 30,501

NOTE. Values are in percentage; total cases are N. Values may not equal 100% because of rounding.
*Includes boats and rotating wing aircraft.
†
Includes water skiing, baseball, basketball, rodeo, track, field, hang gliding, air sports, and skateboarding.

Arch Phys Med Rehabil Vol 85, November 2004

1746 EPIDEMIOLOGY OF SPINAL CORD INJURY, Jackson
rehabilitation providers. Where staff members are still mostly
white (and female), a mismatch with male minority patients is
to be expected. Therefore, staff diversification and diversity
training are indicated.
The data presented also indicate that these staffs increasingly
need gerontologic and geriatrics expertise: the mean age of
newly injured patients has continued to increase since the last
report, and it is now 9 years higher than it was in 1973–1979.
Although many rehabilitation texts and providers still seem to
think that the typical SCI client is a young man around 20 years
of age, that is no longer the case. The mean age is increasing
in all etiology categories (fig 4), least of all in violence (which
remains a young minority male’s domain). This rise mirrors an
almost identical increase in the mean age of the general pop-
ulation, although a change in underlying age-specific incidence
rates or referral patterns to MSCIS cannot be ruled out as
contributing to this trend. Other analyses of the NSCID have
indicated that mortality after SCI has been declining, both early
after injury (during acute care) and later, when patients have
returned to a community residence.15-17 Combined, these
trends indicate a need for changes in rehabilitation and fol-
low-up programs. However, considering our still insufficient
understanding of “aging with SCI” and “aging into SCI,” more
research is needed to understand what the needs and problems
of these groups are and how they can best be served.
Changes in etiology of injury are relevant to the develop-
ment of prevention programs and are also important because
etiology has a major effect on level and completeness of
neurologic injury (see table 2). At one extreme, sports-related
(especially, diving) injuries result in tetraplegia in almost all
instances (table 2), whereas interpersonal violence (mostly,
firearms) results in paraplegia in two thirds of cases. The
completeness-of-injury differential is not as great, but, even so,
significant differences exist between falls and violence among
other groups.
There have been distinct changes in the leading causes of
SCI during the past few years. As noted earlier, the proportion
of SCI because of acts of violence has decreased substantially
since peaking in the mid 1990s. This finding is consistent with
the general decline in violent crime rates in the United States
that has also occurred during this time period. Conversely, the
percentage of new cases of SCI that result from falls continues
to increase. The increase in falls as a cause of SCI is likely
directly attributable to the rising mean age at time of injury
because falls continue to be the leading cause of SCI in persons
over the age of 60 years.
Overall, the percentage of SCI because of sports has de-
clined over time, although that trend did not continue in the
latest time period. In particular, the percentage of SCI because
of diving injuries continues to decrease rather dramatically.
Diving-related SCI has been the subject of considerable re-
search concerning the potentially preventable circumstances
surrounding those accidents as well as numerous primary pre-
vention programs such as the Feet First First Time program.18
Newer and safer residential pool designs have also likely
contributed to the apparent success in reducing the frequency
and percentage of diving-related SCI.
The frequency and percentage of SCI because of American
football has also decreased significantly over time, despite a
slight increase during the current decade. This is likely attrib-
utable to changing rules that outlawed “spearing” (driving
head-first into an opposing player) as well as development and
use of better protective equipment and better training of players
in proper tackling techniques. In fact, the occurrence of foot-
ball-related SCI has declined enough so that it has been re-
placed by snow skiing as the second leading cause of sports-
Arch Phys Med Rehabil Vol 85, November 2004
related SCI. As a result, efforts to learn more about the
circumstances surrounding snow skiing accidents and effective
ways to prevent these injuries should be given increased atten-
tion.
It is important to keep in mind that those recreational sports
activities that involve the most SCIs are not necessarily the
most risky activities. Actual risk can only be determined based
on comparison with levels of exposure to that activity. For
example, gymnastics events that cause few SCIs but also have
relatively few participants have higher incidence rates per
athlete than does American football.19
The percentage of neurologically complete injuries rose sub-
stantially in the 1990s but has returned to historic levels of
approximately 50%. Several factors may be contributing to this
favorable recent trend, including the reduction in gunshot in-
juries that typically result in neurologically complete injuries
and the increase in falls among the elderly that often result in
neurologically incomplete injuries. Improved medical and sur-
gical management may also be contributing to this trend. Use
of methylprednisolone in the early management of SCI is not
documented in the NSCID and, therefore, its potential contri-
bution to the recent decrease in neurologically complete inju-
ries cannot be evaluated.
The NSCID has several well-documented strengths and lim-
itations that must be considered when evaluating the results of
this study.1,7 Strengths include the large sample size, geo-
graphic and patient diversity, standardization of data collection
methods and measures, excellent case identification proce-
dures, prospective data collection using both physical exami-
nation and patient interview, and comprehensiveness of the
information in the database.
This database has shortcomings, however, that both limit the
questions it can answer and necessitate careful interpretation of
results. The most important of these shortcomings is the fact
that the database (despite the fact that one of its components is
called the Registry) is not population based: only those people
with SCI who are admitted to an MSCIS facility soon after
injury are represented. Thus, no estimates of the incidence or
prevalence of SCI can be developed from it, and characteriza-
tions of the US SCI population based on its data have to be
interpreted with care.
There have been no recent nationwide studies of the inci-
dence of SCI in the United States. The various state registries,
supported by the US Centers for Disease Control and Preven-
tion, have published limited information, but no attempt has
been made to generalize from their data to the United States as
a whole.20-23 The most recent study of SCI prevalence is now
almost 10 years old.5 The 16 currently funded MSCIS facilities
are located in 14 states; formerly funded systems add 5 states
to that number, but the population of patients served by these
systems may not be representative of all new SCI cases na-
tionwide. In 1990, Gibson24 estimated that the joint MSCIS
saw 15% of all SCI cases in the United States; however, that
number almost certainly has been reduced, with expansion of
the number of hospital-based and freestanding rehabilitation
facilities since that time.
Based on comparisons with population-based state registries
of SCI, it has been previously demonstrated that the NSCID is
representative of all new cases of SCI that occur in the United
States each year, with the exception that neurologically com-
plete lesions (particularly high-level cervical lesions) are some-
what overrepresented. Nonwhites, males, and acts of violence
are also probably overrepresented slightly.25,26
Even an evaluation of changes over time based on NSCID
information has limitations. Any differences in the make-up of
the group of people with new SCI from one decade to the next
 EPIDEMIOLOGY OF SPINAL CORD INJURY, Jackson
as offered in the present study may reflect changes in under-
lying SCI incidence rates in the population nationwide, but
there might be alternative explanations. For instance, changes
in the database may result from the make-up of the MSCIS
group (even though membership has been fairly stable over the
years, especially in a core group of large systems, there have
been changes), or they may result from changes in referral
patterns that affect who gets referred to an MSCIS from a
community or other hospital, either immediately after injury or
later. The effects of such referral changes are difficult to
quantify. For instance, it is often claimed that smaller hospitals,
often staffed by physicians and allied health professionals
trained by the MSCIS, have increasingly tended to keep “sim-
pler” cases, for example, incomplete paraplegics, and differen-
tially to refer complicated ones, for example, ventilator-depen-
dent tetraplegics. That would seem to be supported by the
increasing percentage of cases discharged from the MSCIS on
a ventilator, as reported here.
Because the NSCID is not population-based, data can only
be presented as frequencies and percentages and not as rates.
Trends in percentages do not necessarily imply corresponding
changes in underlying incidence rates. For example, the per-
centage of cases caused by acts of violence would rise if the
underlying incidence rate because of acts of violence rises
relative to that of other causes. However, the percentage of
cases caused by acts of violence would also rise if the under-
lying incidence rate caused by acts of violence were declining
but at a slower rate than that of other causes.
However, changes in the composition of the MSCIS patient
population as derived from the NSCID often are confirmed by
data from other sources, either parallel information on samples
with SCI (eg, the cases in a state registry) or on societal
changes. For instance, Dijkers27 found that an increase in SCI
because of interpersonal violence and a decrease in SCI be-
cause of MVCs (through 1997) was paralleled by such statistics
as the death rate for firearm injury, which increased, and the
crude death rate for MVCs, which declined over the same
years. Thus, provided that they are interpreted with some
understanding of how the data have been collected, the trends
reported here are relevant to service providers, policy makers,
and researchers—as are the lack of trends.
The epidemiologic summary of SCI presented in this study
reflects the characteristics of new injuries that occur each year.
It is important to understand that the epidemiologic profile of
all persons who currently have an SCI differs considerably
from the profile presented herein. Because of differential long-
term survival rates and the fact that most persons who have an
SCI at this time have had it for many years, persons with SCI
who are alive today are more likely to be older, female, and
white and more likely to have paraplegia and incomplete inju-
ries than depicted in this study.
CONCLUSIONS
This new analysis of the latest trends in the epidemiology of
SCI has revealed several new findings relative to the last report
on this topic from the MSCIS.6 These include dramatic de-
clines in the percentages of neurologically complete injuries,
cases caused by acts of violence and certain recreational activ-
ities such as diving, along with a significant increase in the
percentage of ventilator-dependent cases. A continuation in the
previously reported trend toward increasing mean age at injury
of new cases was also documented. The MSCIS and other
providers of care for persons with SCI must take these trends
into account in planning for future primary prevention pro-
grams and the provision of rehabilitation services.
1747
References
1. Stover SL, DeVivo MJ, Go BK. History, implementation, and
current status of the national spinal cord injury database. Arch
Phys Med Rehabil 1999;80:1365-71.
2. Pagliacci M, Celani M, Zampolini M, et al. An Italian survey of
traumatic spinal cord injury. The Gruppo Italiano Studio Epide-
miologico Mielolesioni study. Arch Phys Med Rehabil 2003;84:
1266-75.
3. Sekhon LH, Fehlings MG. Epidemiology, demographics, and
pathophysiology of acute spinal cord injury. Spine 2001;26(Suppl
24):S2-12.
4. Richards JS, Go BK, Rutt RD, Lazarus PB. The national spinal
cord injury collaborative database. In: Stover SL, DeLisa JA,
Whiteneck GG, editors. Spinal cord injury: clinical outcomes from
the model systems. Gaithersburg: Aspen; 1995. p 10-20.
5. Lasfargues JE, Custis D, Morrone F, Carswell J, Nguyen T. A
model for estimating spinal cord injury prevalence in the United
States. Paraplegia 1995;33:62-8.
6. Nobunaga AI, Go BK, Karunas RB. Recent demographic and
injury trends in people served by the Model Spinal Cord Injury
Care Systems. Arch Phys Med Rehabil 1999;80:1372-82.
7. DeVivo MJ, Go BK, Jackson AB. Overview of the national spinal
cord injury statistical center database. J Spinal Cord Med 2002;
24:335-8.
8. American Spinal Injury Association. International standards for
neurological classification of spinal cord injury. Chicago: ASIA;
2000.
9. Frankel HL, Hancock DO, Hyslop G, et al. The value of postural
reduction in the initial management of closed injuries to the spine
with paraplegia and tertaplegia. I. Paraplegia 1969;7:179-92.
10. Waters RL, Adkins RH, Yakura JS. Definition of complete spinal
cord injury. Paraplegia 1991;29:573-81.
11. DeVivo MJ. Causes and costs of spinal cord injury in the United
States. Spinal Cord 1997;35:809-13.
12. DeVivo MJ, Whiteneck GG, Charles ED. The economic impact of
spinal cord injury. In: Stover SL, DeLisa JA, Whiteneck GG,
editors. Spinal cord injury: clinical outcomes from the model
systems. Gaithersburg: Aspen; 1995. p 234-71.
13. Whiteneck GG. The high costs of high-level quadriplegia. In:
Apple DF, Hudson LM, editors. Spinal cord injury: the model.
Proceedings of the national consensus conference on catastrophic
illness and injury. Atlanta: Shepherd Center for Treatment of
Spinal Injuries; 1990. p 114-7.
14. Berkowitz M, O’Leary PK, Kruse DL, Harvey C. Spinal cord
injury: an analysis of medical and social costs. New York: Demos;
1998.
15. DeVivo MJ, Stover SL. Long-term survival and causes of death.
In: Stover SL, DeLisa JA, Whiteneck GG, editors. Spinal cord
injury: clinical outcomes from the model systems. Gaithersburg:
Aspen; 1995. p 289-316.
16. DeVivo MJ, Ivie CS. Life expectancy of ventilator-dependent
persons with spinal cord injuries. Chest 1995;108:226-32.
17. DeVivo MJ, Krause JS, Lammertse DP. Recent trends in mortality
and causes of death among persons with spinal cord injury. Arch
Phys Med Rehabil 1999;80:1411-9.
18. DeVivo MJ, Sekar P. Prevention of spinal cord injuries that occur
in swimming pools. Spinal Cord 1997;35:509-15.
19. Clarke KS. Spinal cord injuries in organized sports. Model Sys-
tems SCI Digest 1980;2:9-17.
20. Johnson RL, Gabella BA, Gerhart KA, McCray J, Menconi JC,
Whiteneck GG. Evaluating sources of traumatic spinal cord injury
surveillance data in Colorado. Am J Epidemiol 1997;146:266-72.
21. Acton PA, Farley T, Freni LW, Ilegbodu VA, Sniezek JE, Wohl-
leb JC. Traumatic spinal cord injury in Arkansas, 1980 to 1989.
Arch Phys Med Rehabil 1993;74:1035-40.
22. Price C, Makintubee S, Herndon W, Istre GR. Epidemiology of
Arch Phys Med Rehabil Vol 85, November 2004
1748 EPIDEMIOLOGY OF SPINAL CORD INJURY, Jackson
traumatic spinal cord injury and acute hospitalization and
rehabilitation charges for spinal cord injuries in Oklahoma,
1988-1990. Am J Epidemiol 1994;139:37-47.
23. Thurman DJ, Burnett CL, Jeppson L, Beaudoin DE, Sniezek JE.
Surveillance of spinal cord injuries in Utah, USA. Paraplegia
1994;32:665-9.
24. Gibson CJ. Criteria for evaluating performance of the system. In:
Apple DF, Hudson LM, editors. Spinal cord injury: the model.
Proceedings of the national consensus conference on catastrophic
illness and injury. Atlanta: Shepherd Center for Treatment of
Spinal Injuries; 1990. p 45-8.
Arch Phys Med Rehabil Vol 85, November 2004
25. Go BK, DeVivo MJ, Richards JS. The epidemiology of spinal
cord injury. In: Stover SL, DeLisa JA, Whiteneck GG, editors.
Spinal cord injury: clinical outcomes from the model systems.
Gaithersburg: Aspen; 1995. p 21-55.
26. DeVivo MJ. Epidemiology of traumatic spinal cord injury. In:
Kirshblum S, Campagnolo DI, DeLisa JA, editors. Spinal cord
medicine. Philadelphia: Lippincott, Williams & Wilkins; 2002. p
69-81.
27. Dijkers M. Spinal cord injury caused by interpersonal violence:
epidemiologic data from the national spinal cord injury database.
Top Spinal Cord Inj Rehabil 1999;4:1-22.