THE JOURNAL OF TRAUMA Vol. 14, No.

3
COpyright © 1974 by The Williams & Wilkins Co. Printed in U.S.A.

THE INJURY SEVERITY SCORE: A METHOD FOR DESCRIBING

( PATIENTS WITH MULTIPLE INJURIES AND EVALUATING
EMERGENCY CARE

e f SUSAN P. BAKER, M.P.H., BRIAN O'NEILL, B.Se.,

i WILLIAM HADDON, In., M.D., AND WILLIAM B. LONG, M.D.
li'mm the Division of Forensic Pathology, The Johns Hopkins University School of Hygiene

i
\ and Public Health, Baltimore, Maryland; Insurance Institute for Highway Safety, Wash-
I ington, D.C.; and Maryland Institute for Emergency Medicine, University of Maryland,
Baltimore
i

I
Injuries are a serious problem common to all The AIS and CRIS pertain to individual
societies. Yet even within a single community, injuries. Even though most deaths following
1 groups of injured persons differ as to the nature automotive crashes involve injury to more than
1
f
!~ and severity of their injuries. The difficulty of
adjusting for such variation has hampered sci-
entific study of injured persons. Nevertheless, it
one part of the body, a scale for describing
multiply injured patients has been lacking.
The present study was undertaken to deter-
is essential to take differences in severity of in- mine the extent to which AIS ratings correlate
( jury into account when comparing the morbidity with mortality. The analysis led to development
l'
'I
and mortlllity.of various groups for purposes of
evaluating their emergency and subsequent care
of a simple method-based on the AIS-of ad-
justing for multiple injuries. TIns method, the

~ (1) .
Two basic research approaches can be used in
"Injury Severity Score," makes possible a valid
numerical description of the overall severity of

l dealing with this problem. The first is to com-

pare only persons with similar injuries. When
feasible this is often the best approach, but the
injury in persons who have sustained injury to
more than one area of the body.

i
I
numbers of injuries of each specific type and
severity are often too small to support statisti-
cally sound conclusions.
The second approach is to compare persons
J\ whose injuries, although not necessarily the
same anatomically, are of the same severity.
Combining patients into groups on the basis of
severity of injury requires the use of scales such
as the Abbreviated Injury Scale (AIS) (4, 12)
and the Comprehensive Research Injury Scale
(CRIS) (5,12), which were developed to provide
a method for rating and comparing injuries
incurred in automotive crashes. The widely used
AIS is the simpler of the two scales; the CRIS
is a detailed extension of the AIS. These two
scales were based primaril:y on the professional
experience and judgment 6f the physicians who
constructed them. The degree to whicn they re-
late to morbidity and mortality has remained
conjectural, despite the medical reasonableness
of the scales themselves.
Address for reprints: William Haddon, Jr., M.D.,
Insurance Institute for Highway Safety, Watergate
600, Washington, D.C. 20037.
187
METHODS
The study group included 2,128 motor vehicle
occupants, pedestrians, and other road users
whose injuries resulted in hospitalization or
caused death. All such patients at eight Balti-
more hospitals during the 2-year period 1968-
1969 were included. The eight hospitals were
selected on the basis of having record systems i;
that made it feasible to identify patients ad-
mitted because of vehicle-related injuries.
Seven of the hospitals were participants in the
Professional Activity Study (PAS), and at these
hospitals the coded PAS data were used. At the
eighth hospital data were obtained from patients'
charts. Patients transferred to other hospitals or
extended care facilities were followed to deter-
mine their status as of March 1971.
Records from the Office of the Chief Medical
Exanllner of Maryland provided a second source
of information for persons who died following
hospital admission, and also made it possible to
include persons who were dead on arrival (DOA)
or who died in emergency rooms prior to admis-
 188
TABLE I
The Abbreviated Injury Scale
Examples of Codes faT
Chest Injuries
AIS Code
1
Injury Description
Muscle ache or chest wall stiffness
2 Simple rib or sternal fractures
3 Multiple rib fractures without res-
4
5
piratory embarrassment
Flail chest
Aortic laceration
sion-deaths that might not appear in hospital
admission or discharge records.
Injuries were categorized according to the
AIS (Table I) modified for the present study in
two respects. First and most important, we did
not use the AIS codes 6 through 9, which are
normally assigned to any fatality occurring
within 24 hr, irrespective of injury severity. All
such fatalities were coded as if the outcome
wete not known. Thus all injuries were rated by
severity, irrespective of outcome, and the most
severe injury code used was 5. Use of the fatal
codes 6 to 9 would have made it impossible to
compute meaningful death rates for the various
severity codes, and in addition some details of
the injuries themselves would have been lost.
Second, facial injuries were separated from
cranial and neck injuries-in part because facial
injuries, being common in automobile crashes,
might otherwise have overshadowed other head
injuries, and in part because the disfigurement
often associated with facial injuries may have in-
fluenced their AIS rating.
With the AIS, each injury was categorized by
body area (head or neck, face, chest, abdominal
or pelvic contents, extremities or pelvic girdle,
and general) and severity (1, minor; 2, moder-
ate; 3, severe, not life-threatening; 4, severe,
life-threatening, survival probable; 5, critical,
survival uncertain). _For injuries coded by the
International Classification of Diseases, Adapted
(ICDA) at the PAS hospitals, each ICDA code
was translated into an AIS grade. For example,
the ICDA code 835, denoting dislocation of the
hip, was converted to "grade 3 injury, ex-
tremity."
After grading of all injuries for a given pa-
tient, each body area was categorized by the
most severe injury in that area. For example, if a
THE JOURNAL OF TRAUMA, March 1974
person had two chest injuries graded 1 and 3,
his grade for chest injury was 3.
Autopsies had been performed on 74% of all
persons who died. When information was avail-
able from both hospital and autopsy data,
severity grading was based on the autopsy. I,
(
RESULTS
The overall ratio of hospital admissions to
deaths (including DOA's) was 8: 1. The ratios
of admissions -to deaths ranged from approxi-
I
mately 5: 1 to 60: 1 in individual hospitals-re-
flecting, in general, differences in the proportion
of severely injured patients each received.
There were large differences between hospitals
in the proportion of children and elderly pa-
tients. There were also differences between hos-
pitals in the proportion of patients who had
I
sustained injury to specific body areas, such as
the face; but at all eight hospitals the ex-
!
tremities and pelvic girdle were the most fre-
quently injured parts of the body. Forty-nine'
percent of all patients sustained injuries to ex- I
I
tremities or the pelvic girdle, and in 35% of all
patients this was the most severely injured area.
Table II shows the distribution of injury
severity. (It is important to remember that this \
distribution does not represent the entire spec-
trum of highway injuries, since many injuries,
especially minor ones, do not result in admission f
to hospitals.) In apprm6mately half (49%) of
all patients in the study, the most severe injury
(
'1
was grade 3. I
The percentage of patients who died increased
with the AIS grade of the most severe injury
(Fig. 1), as did the proportion of deaths that
[
were DOA.
I
Patients in each of these five AIS severity --.-~
groups had a wide spectrum of additional in-
juries. For instance, some patients whose most
severe injury was grade 4 had no injury else-
where, while others had minor to severe in-
juries in other parts of the body. Figure 2A -
shows that for persons whose most severe in-
jury was grade 4, the death rate increased with
injury severity in the second area, ranging from
6% in persons with no injury or only a grade 1
injury in a second area to 60% of those with a
second grade 4 injury. Similarly, Figure 2B
shows that for persons whose most. severe injury
 h 1974

and 3, J I
Vol. 14, No.3 / INJURY SEVERITY SCORE

TABLE II
189

;, of aU
3 avail-
f
1I AlS Grade Most
Outcome by AlB Grade of Most Severe Injury
"Admitted" Total
Severe Injury Dead on Arrival
, data, Died Later' Survived Unknown No. Percent
I
I 1 0 0 80 1 81 (4)
I 2 0 2 437 1 440 (20)
3 6 23 997 20 1,046 (49)
4 13 30 229 3 275 (13)
Lons to 5 93 80 97 3 273 (13)
~ ratios Unknown 1 0 12 0 13 (1)
pproxi-
Total 113 135 1,852 28 2,128 (100)
tls-re- (1) (100)
Percent (5) (6 ) (88)
portion
* Includes 34 patients who were alive on arrival but died before admissions procedures were com-
)spitals pleted.
:ly pa-
;n hos-
10 had
:uch as
he ex-
st fre-
&y-nine
to ex-
75%
) of all
area.
, 64
injury
at this r\
~ spec- % DIED 50%
ljuries,
mSSlOn
%) of
i
injury (
25%
Teased
injury
.s that

~verity
o 0.5
wI in- 1 2
: most
T else- AIS GRADE OF MOST SEVERE
re In-
re 2A
INJURY
FIG. 1. Mortality by AIS grade of most severe injury. DOA's included in calculations.
,re in-
:! with
; from was grade 5, death rates ranged from 22% to than for others whose most severe injury was
100%, depending upon the degree of injury in grade 5. Such relationships pointed to the need
rade 1
the second most severely injured area. for a method to derive, in each case, a summary
iVith a
Comparison of Figures 2A and 2B shows that of injury severity that would adjust for varia-
:e 2B
death rates were higher for some groups of pa- tions in mortality associated with the number of
injury body areas involved and the severity of trauma
tients whose most severe injury was grade 4
 190 THE JOURNAL OF TRAUMA, March 1974 VoL. 1.

II,"

75%r----------------------H--------------

% DIED
50%~------------

25% \ - - - - - - - -

1
I
AIS GRADE OF 2ND MOST 0-' 2 3 4 0-1 2 3 5
SEVERE INJURY
I
) AIS
AISGRADE OF MOST 5
SEVERE INJURY
4
.
I
~

NO. OF PERSONS 272 267 r AI!

FIG. 2. Mortality by AlS grade of second most severe injury: (A) when most severe
injury was grade 4 and (B) when most severe injury was grade 5. DOA's included in calcu-
lations. AI

in each. The authors of the AIS caution against
adding or averaging the AIS ratings, stating that
"the quantitative relationship of the AIS codes
is not known and is almost certainly nonlinear"
(12). Evidence of nonlinearity is illustrated by
Figure 1, which shows that mortality increases
disproportionately with AIS rating of the most
severe injury. Additional nonlinearity is demon-
strated by comparison of Figures 2A and 2B:
the death rate for persons with two injuries of
third most severely injured area also influenced
mortality.
When the AIS grades for each of the three
\
most severely injured areas were squared and
the results added together, comparable t.otals
again proved to be associated with similar
mortality rat.es, and the correlation between
total injury severity and mortality was further
I one-f
grad,
.. Fi!
improved. Including the grade of the fourth twee:
most severely injured area had no appreciable ~.Deat

grades 4 and 3 was not comparable to that of effect. :69y'

persons with two injuries of grades 5 and 2 In view of the foregoing, an "Injury Severity incn
I =
(sum 7 in both cases). Score" was defined as the su.rn of the squares of a,ge-:
'peci:
The simplest nonlinear relationship is quad- the highest AIS grade in each of the thj'ee most
ratic. TIns led to investigation of the possibility severely injured areas, In illustration, a person Eigl:
that squaring the AIS grades for the most with a laceration of the aorta (AIS = 5), '·for
severe injury in each body area before adding multiple closed long bone fractures (AIS = 4), ',ago
them together would provide a valid adjustment and retroperitoneal hemorrhage (AIS = 3), 'Sev!
for multiple injuries. When the AIS grades for would have an Injury Severity Score of 50 (25 + Il!jt
each of the two most severely injured areas were 16 + 9), The highest possible score for a person :·was
squared and the two results added together, it with trauma to a single area is 25, 'for
Use of the Injury Severity Score dramatically 'and
was found that death rates were usually similar
for comparable totals. For instance, for persons increased the correlation between severity of in- peri
~]i
whose two most severely injured areas were jury and mortality, as compared to th3 AIS
·WeI
graded 5 and 0 and for those graded 4 and 3 grade for the most severe injury. [Statistically,
,:tha
=
(sum of squares 25 in both cases), death rates approximately half (490/0) of the variance. in
were almost identical (220/0 and 240/0, respec- mortality was explained by using the Injury Se-
tively). Figure 3 shows that the grade of the verity Score to measure severity, compare.d to
 'ch 1974
1 Vol. 14, No.3 / INJURY SEVERITY SCORE 191

I!
1

%DIED

I )

AIS GRADE OF 3RD MOST

,I SEVERE INJURY

I AIS GRADE OF 2ND MOST

3 3 4

I
SEVERE INJURY
~re

:ti-
AIS GRADE OF MOST
I SEVERE INJURY 4 5 5
\
fluenced
NO. OF PERSONS 102 78 38
1e three r
red and FIG. 3. Mortality by AIS grade of third most severe injury. DOA's included in calculations.
e totals
similar
between i
(
one-fourth (25%) using only the highest AIS
grade.]
Figure 4 shows the observed relationship be-
Among patients who died, the higher the score
the shorter survival tended to be (Fig. 6). In
further illustration, among those with an Injury Severity
, fourth /
tween Injury Severity Score and mortality. Score below 20, half the deaths occurred over a
lreciable I Death rates were higher for persons in the 50- week after injury and 18% over a month after
69 year age group than for younger persons, and injury. (No patients in the study were known
Severity increased markedly for those age 70 and over. The to have died over 3 months after injury.) At the
uares of age-associated increase in mortality was es- other extreme, when the score was 50 or greater
'ee most pecially pronounced for less severe injuries. all deaths in this series occurred within a week
. person Figure 5 shows that the ratio of the death rate of injury and three-fourths within the first hour .
= 5) , for age 70 and over to the death rate below Statistical correlation analysis indicated that,
~= 4), ago 50 increased exponentially as the Injury when the Injury Severity Score and age of pa-
= 3), Severity Score decreased from 55 to 15. For tient were taken into consideration, survival
J (25 + Injury Severity Scores of 50 and higher there was not significantly influenced by race or sex,
, person was almost no age difference in mortality, but or by whether a person was a vehicle occupant
for scores of 10-19 the death rate for ages 70 or a pedestrian.
I'
latically and over was more than eight times the rate for
persons less than 50 years old. DISCUSSION
y of in-
h3 AIS For Injury Severity Scores less than 10, there The results demonstrate that death rates in-
stieally, were no deaths at any time among patients less crease in the presence of injuries in a second or
lnce in than 50 years old and no DOA's at any age. The third body area (Figs. 2, 3), even when the
ury Se- highest score in a surviving patient was 50, in additional injuries would not normally, in them-
~
ared to ( this particular series. selves, be life-threatening. That patients with
I
 192 THE JOURNAL OF TRAUMA, March 19"14

100
/ I
+ / I
90 70 (N=l09), / I
""'" I
80
,
,
,'" 50-69 (N=316) /
~/
"I
70 , I
J1
, " ;""
."."" 0-49 (N=l540)
60 I .".
%DIED I
50 I
.:
i
'
I
40

30

20

10 \

0
10 20 30 40 50 60
I
\
INJURY SEVERITY SCORE (
FIG. 4. Mortality by Injury Severity Score for three age groups. DOA's excluded from
calculations. Dotted lines connect points based upon less than 10 persons. I
(
,1
substantial injuries to multiple body areas have firm the belief of the authors of the AIS that
worse prognoses has long been known, but not adding or averaging the ratings for various in-
I A
previously quantified to a similar extent. This juries will not adequately adjust for multiple
effect of additional injuries on mortality under-
scored the need for a method of summarizing the
overall severity of injury.
injuries.
The score produced by adding together the
squares of the AIS ratings for each patient's
I k
Y
Although the results show that the AIS, modi-
r
three most severely injured body areas proved II
fied by non-use of ratings over 5, does correlate to correlate even more closely with death rates II
substantially with death rates,* they also con- than did the AIS itself. (The quadratic relation- \'
ship between AIS grades and the threat to life '1
* The more recent Comprehensive Research In-
jury Scale (CRIS) was not available when this in- of injuries in multiple body areas may reflect a
vestigation was initiated. Comparison of CRIS fundamental aspects of response to injury that d
threat-to-life grades and AIS grades shows that for should 'be the subject of research on changes II
87% of injuries the two are the same or only one over time in basic biochemical and physiological
grade apart. This correlation suggests, first, that the s·
Injury Severity Score can be based on this threat- variables.) Because of its much higher correla-
to-life scale as well as on the AIS and, secondly, tion with mortality and its solution to the prob- s
that results of the present investigation validate lem of summarizing multiple injuries, this ''In-
both scales. It is expected that further research
would demonstrate that death rates correlate even Injury Severity Scores derived from them) than a
more closely with CRIS threat-to-life grades (and with AIS grades. ~ e:
I

Ii
 mock 1974 Vol. 14, No.3 193
/ INJURY SEVERITY SCORE

10

I
I
I
9

8
I

40)
I DEATH RATE AGE 70+
DEATH RATE AGE < 50
7

4
)

! 3
1 "" ...... ,
2
" .........
I ....
60
I( 10 20 30
INJURY SEVERITY SCORE
40

FIG. 5. Age differential in mortality by Injury Severity Score. DOA's excluded from
50 60

.) calculations. Dotted lines co=ect points for which there were less than 10 persons age 70+ .
from
\
jury Severity Score" was used for analysis of ad- treatment. [An investigation of 33 deaths due to
(
./ ditional factors in relation to mortality. abdominal injuries revealed that none of the pa-
AIS that tients died within an hour of injury, and that
rarious in- I AGE approximately half the deaths might have been
r multiple

gether the
1 patient's
I Results of the present study reflect the well-
known fact that elderly patients, compared with
younger persons similarly injured, have worse
prevented with prompt and proper diagnosis
and treatment (6)]. The association between
severity of injury and rapidity of death (Fig. 6)
suggests that patients with Injury Severity
prognoses. Use of the Injury Severity Score
:as proved Scores below 50 may have the greatest potential
. made it .possible to determine that this increased
.eath rates for improved survival rates, since the majority
mortality in the elderly is most pronounced
c relation- who died with scores below 50 were still alive
when the injuries are least severe (Figs. 4, 5).
eat to life an hour after injury.
The magnitude of the effect of age qn mortality,
lay reflect
and the difference between hospitals in the age
ljury that POTENTIAL USES OF THE INJURY SEVERITY
distribution of their patients, underscore the
n changes p SCORE
( necessity of adjusting for age whenever two
ysiological
. series of patients are to be compared. Since patients with scores below 10 rarely die,
3r correla-
the prob- survival rates for groups with scores between
SURVIVAL TIME 10 and 50 may prove to be important evaluative
, this "In-
Improvements in survival rates are especially indices. This group includes the patients who,
hem) than achievable among persons who survive long as Waller (13) suggests, provide the real key
enough to make possible the initiation of medical to evaluating the emergency care system: those
 vo~.
194 THE JOURNAL OF TRAUMA, March 1974
types
WITHIN 3 MONTHS _
-100 death
hand,
realis
arriv:
WITHIN 1 MONTH ~
-80 the f:
DOA
dead
heen
-60 of n
TIME FROM INJURY
WITHIN 1 WEEK
- %OF
DEATHS
use i
the 1

MAJ'

F
the
WITHIN 6 HOURS ______
and
-20
dev
pat
WITHIN 1 HOUR
r sur'

I
cal
-0
0-19 20-29 30-49 50+ cot
alt,
INJURY SEVERITY SCORE l
baf
FIG. 6. Length of time from injury until death by Injury Severity Score. Includes all deaths.
'I sm
me
who are sick enough to be adversely affected by injured may also find the Injury Severity Score :1
poor care, but not so sick that they will not sur- useful. For example, mortality rates for a trauma m'
'0 stl
vive even with optimum care. It is important to unit such as the Maryland Institute for Emer- i

note, however, that certain patients whose In-
jury Severity Scores would be well in excess of
50 may survive if promptly treated by per-
tinently trained specialists who have at their
disposal all necessary resources. Therefore,
evaluation of emergency response systems should
include patients throughout the range of the
Injury Severity Score.
The Injury Severity Score can easily be added
to data coded for research purposes, and to the
hospital chart itself. This description of injury
severity would enhance the value of patient rec-
ords, from the simplest records to those in the
Trauma Registry (3). After grouping patients
on the basis of overall injury severity, any given
emergency room, hospital, region, or country
could describe the proportion of its trauma popu-
lation that is injured to a specified extent.
Ability to compare groups of patients classi-
fied by overall injury severity makes it possible
to evaluate methods of treatment, identify prob-
lem areas, and document progress. Professional
service review organizations wishing to compare
various institutions that provide care to the
gency Medicine, where the typical patient has ! at
sustained multiple severe injuries, cannot be
compared meaningfully with mortality rates for
I sc
e,
all admitted injured patients at another hos- ( SE
ir
pital; however, a comparison based on compara-
ble Injury Severity Scores would be useful. r
)
I tE
i CI
METHODOLOGICAL CONSIDERATIONS
A
The present study does not include all vehicle- \
I
E
related injuries and deaths from a defined geo- I t
graphic area. This would be important if over-
all injury death rates were under consideration.
However, a major advantage of the Injury
Severity Score is that even when biases are
present or suspected, valid comparisons can
usually be made between the death rates of
sub-groups with similar Injury Severity Scores
and age distributions, provided follow-up in-
formation is available for both groups for corre-
sponding time periods.
The appropriateness of including DOA's in
calculation of mortality rates will depend upon
the objectives of a.n investigation. DOA's should
be included, for e~ample, in studies fci~using on
 Vol. 14, No.3 / INJURY SEVERITY SCORE
ch 1974
types of vehicles or in calculation of the total
death rate for a geographic area. On the other
hand, when evaluating hospital care it is not
realistic to include patients who were dead on
arrival. The latter situation is complicated by
the fact that a patient who would be pronounced
DOA at one hospital might not bc pronounced
dead at another hospital until some time had
195
produce them-for example, mechanical, ther-
mal, electrical, ionizing, chemical (8-1O)-we be-
lieve that subdivision of this type is necessary
to rationalize such codes.
Most patients in this study had been injured
by excessive transfers of mechanical energy. The
findings reported cannot, therefore, be general-
ized to injuries due to transfer of other kinds of

I
)
been spent in resuscitative attempts. Resolution
of this dilemma might require that each facility
use the same objective criteria for vital status at
energy. Similarly, since the AIS and hence the
Injury Severity Score based oil it are oriented to
the types of mechanical energy damage that
_ %OF
DEATHS ! the moment of arrival. commonly result from road crashes, ell.'trapola-
tion of the findings beyond this major group may
MAJOR CURRENT NEEDS
be inappropriate except where similar injuries

I )
Further improvement in ability to evaluate
the effectiveness of emergency response systems
and medical care of the injured depends upon
developing the ability both to classify the injured
are involved. For example, injuries sustained in
falls might be suitably so scored, but gunshot
wounds might not. Rating systems analogous to
the AIS and Injury Severity Score need to be

I patient, before and after admission, and to mea-

sure his outcome. Other indices, including clini-
developed and applied to the full range of hu-
man injury to help bring the science of this rela-

I
cal signs and biochemical measurements (11),
could be used either in conjunction with or as an
alternative to injury severity rating systems
I based on clinical diagnosis. Nor should the mea-
.8.
surement of outcome be limited to analyses of
fiy Score mortality rates. For example, other outcome
tively neglected area of human wastage to the
breadth and quality its importance demands (2).
SUMMARY AND CONCLUSION
A method for comparing death rates of groups
of injured persons was developed, using hospital

I
. trauma measurements suggested by Gibson (7) for use in
r Emer-
studying emergency medical systems include dis-
ient has ability days, residual impairment, and symptom
mot be scores 6 and 12 months after injury. For some
'ates for evaluative purposes such measurements are more
ler hos- ( sensitive than death rates, and need to be studied
Jmpara-
! in relation to severity of injury.
I Another major need is for revision of the In-
ternational Classification of Disease trauma
and medical examiner data for more than two
thousand persons. The first step was determina-
tion of the extent to which injury severity as
rated by the Abbreviated Injury Scale correlates
with patient survival. Substantial correlation was
demonstrated. -Controlling for severity of the pri-
mary injury made it possible to measure the
effect on mortality of additional injuries. Injuries

I
that in themselves would not normally be life-
codes. At present, conversion of ICDA codes to threatening were shown to have a marked effect
vehicle- AlS grades (and therefore Injury Severity on mortality when they occurred in combination
Scores) is' difficult because the ICDA index for with other injuries. An Injury Severity Score was
led geo- I trauma is extremely limited, especially in the
if over- ~ developed that correlates well with survival and
.eration. I classification of blunt injury. Only 15 whole num-
provides a numerical description of the overall
Injury bers, of the 159 assigned to trauma by the ICDA,

I
severity of injury for patients with multiple
.ses are describe internal injuries to the head, chest, and
trauma .
illS can
abdomen. Thus, less than 10% of the. codes de-
Results of this investigation indicate that the
'ates of scribe those injuries that account for the ma-
jority of trauma deaths. If the potential useful- Injury Severity Score represents an important
Scores step in solving the problem of summarizing in-
-up in- ness of the ICDA injury codes is ever to be re-
alized, it is imperative that they be revised to jury severity, especially in patients with multiple
r corre- trauma. The score is easily derived, and is based
pravide greater detail in describing the wide
spectrum of damage resulting from the physical on a widely used injury classification system, the
)A's in
hazards of man's environment. Furthermore, Abbreviated Injury Scale. Use of the Injury Se-
ld upon
since the most basic categorization of injuries is verity Score facilitates comparison of the mor-
I should
in terms of the kinds of energy transfers that tality experience of varied groups of trauma pa-
.sing on
 196
tients, thereby improving ability to evaluate care
of the injured.
Addendum
It has come to our attention that Ryan and
Garrett* also converted a linear injury scale to
one in which the injuries with the higher ratings
were heavily weighted, and proposed that the
weights for each injury could be added together
to give a total injury score for an individual.
However, the method involved an eight-class in-
jury scale that has not been widely used, and the
validity of the proposed weights was not tested.
Acknowledgments
This investigation was supported by the In-
surance Institute for Highway Safety and the
,-
,.( Maryland Medical-Legal -Foundation. The co-
operation of the eight hospitals that made infor-
~ation available is gratefully acknowledged. Ap-
preciation is also due to Mr. Manuel Machiran,
a student at the University of Maryland Medical
School, who abstracted information from hospi-
tal records and coded injury data.
REFERENCES
1. BAKER SP: Evaluation of medical care of the
injured. J Trauma 11 :892-894,1971
2. BAKER SP: Injury control. In SARTWELL PE
(Editor): Preventive Medicine and Public
Health. Appleton-Century-Crofts, 1973, p. 987-
1006
3. BOYD DR, LOWE RJ, BAKER RJ, et al: Trauma
* Ryan GA and Garrett JW: A quantitative
scale of impact injury. CAL No. VJ-1823-R34 Buf-
falo, Cornell Aeronautical Laboratory, Inc., 1968.
THE JOURNAL OF TRAUMA, Ma1·ch 1974
registry. New computer method for multifac_
torial evaluation of major health problem
JAMA 223:422-428,1973 .
4. COMMITTEE ON MEDICAL ASPECTS OF A UTOMOTIVEJ
SAFETY: Rating the severity of tissue damage
1. The Abbreviated Scale. JAMA 215 :277-280·
1971 '
5. COMMITEE ON MEDICAL ASPECTS OF AUTOMOTlVEJ
SAFETY: Rating the severity of tissue damage.
II. The Comprehensive Scale. JAMA 220 :717-
72Q, 1972
6. GERTNER HR JR, BAKER SP, RUTHERFORD RB
et al: Evaluation of the management of ve:
hicular fatalities secondary to abdominal in-
jury. J Trauma 12:425--431, 1972 81
7. GIBSON G: Research and evaluation of emer_ sl
gency medical services. Health Services Re-
ports, in press n
8. HADDON W JR: A note concerning accident the- p
ory and research with special reference to o
motor vehicle accidents. Ann NY Acad Sci 107:
635--646, 1963
9. HADDON W JR: The prevention of accidents. In
CLARK DW, MACMAHON B (Editors): Preven-
tive Medicine. Little, Brown and Co., 1967,
p.591-621
10. HADDON W JR: Energy damage and the ten
countermeasure strategies. J Trauma 13 :321-
331,1973
11. SACCO ViTJ, COWLEY RA, GOLDFARB MA, et al:
A prognostic index in critical care medicine.
Aberdeen Proving Ground, Md., Edgewood
Arsenal Technical Report EATR 4777,1973
I,
12. STATES JD, FENNER HA JR, FLAM.BOE EE, et al:
Field application and research development of
the Abbreviated Injury Scale. SAE Print
710873. New York, Society of Automotive En-
gineers, 1971
13. WALLER JA: How do we evaluate services and
set priorities? Proceedings of Workshop on
i
Emergency Health Services in Vermont. Bur-
lington, University of Vermont, 1970
[