Journal of Gerontology: MEDICAL SCIENCES
1999, Vol. 54A, No. 12, M613-M620
Copyright 1999 by The Gerontological Society of America
Disability Fingerprints: Patterns of Disability in Spinal
Cord Injury and Multiple Sclerosis Differ
Helen Hoenig, 1,4 Lauren Mcintyre." JenniferHoff} Gregory Samsa," and LaurenceG. Branch>'
'Physical Medicineand Rehabilitation Service,2Health ServicesResearch and Development,
and 3Veterans AdministrationNationalCenter for Health Promotion,DurhamVeterans Administration Medical Center;
fDivision of Geriatrics, Departmentof Medicine;5Division of Biometry, Departmentof Communityand Family Medicine;
and 6Centerfor ClinicalHealth Policy Research,Departmentof Medicine and Departmentof Communityand Family Medicine,
Duke University Medical Center,Durham, North Carolina.
Background. Models for causation of functional disability differ as to whether different diseases lead to common expressions of disability versus producing unique "disability fingerprints." Multiple sclerosis (MS) and Spinal Cord Injury (SCI) both
affect the spinal cord; however, their pathophysiologies differ (progressive vs nonprogressive; multifocal vs unifocal).
Methods. Patterns of disability were compared among veterans who reported in a national survey that they had MS
(n
= 1789) or SCI (n =6361) as the sole cause of their spinal cord dysfunction. The study used self-reported information on
disease duration, physical impairments, and self-care skills to compare the two samples for differences in disability overall and
after stratification according to (a) disease duration, and (b) specific physical impairments.
Results. Patterns of disability differed significantly among persons with MS compared to SCI (p =.001). Differences in
level of disability between the two samples remained statistically significant after stratification on disease duration. There were
substantial, statistically significant differences between the two samples in the amount and kinds of physical impairment.
However, differences in level of disability between the two conditions remained highly significant after stratifying on number
of affected limbs (p = .(03), amount of useful movement (p = .(01), overall motor impairment (p = .(03), amount of sensation
(p = .000, impairment in memory and thinking (p = .(01), and visual impairment (p = .(01).
Conclusions. This study shows differing diseases indeed have unique disability fingerprints, which remain unique after
controlling for disease duration and for population-specific differences in physical impairment. These findings point out the
need to explain the disablement process more fully.
N 1980 the World Health Organization (WHO) (la) adopted a
theoretical model for the causation of disability which proposes that disease causes organ system impairment which in
tum causes disability. One extension of this model is the final
common pathway thesis, which proposes that different diseases
lead to common expressions of disability or "frailty," because
different diseases may adversely affect the same organ system
with similar resulting impairments, (1b). This can be termed the
disability fingerprint thesis (Ic). The similarities in organ system impairment across diseases would then result in similar
functional deficits across the diseases. And, indeed, work by
Tinetti and colleagues (2) supports this thesis. From another
perspective, diseases were classified as unique entities initially
because they had unique presentations; diseases continue to be
classified as unique diseases beeause they have unique pathophysiology; thus, different diseases might be expected to have
unique patterns of disability. And, indeed, work by Guccione
and colleagues (3) supports this alternate hypothesis. If the
WHO theoretical model is valid, it should be possible to predict
the amount and kinds of disability in a population based on the
impairments in the population.
Multiple sclerosis (MS) is a disease that is usually progressive over time and that causes damage at multiple sites in the
central nervous system. In contrast, the pathology in traumatic
spinal cord injury (SCI) occurs at a single point in time and acts
at a discrete location in the central nervous system (the spinal
I
cord). The effects of SCI are not incremental and are focal,
whereas the effects of MS are both incremental and diverse.
The progressive versus nonprogressive pathophysiology would
be expected to cause the disability fingerprints for the two conditions to vary with duration of illness (e.g., loss of function in
all four limbs, and consequent disability, would be increasingly
common among persons with MS for a prolonged period of
time, compared to persons whose disease duration was short;
whereas the rate of quadriparesis and related disability would
vary less according to duration of injury among persons with
SCI). The difference in the unifocal versus multifocal nature of
the two conditions would be expected to produce differences in
the kinds of impairments seen in the two populations (e.g., cognitive impairment would be more common in MS, because MS
frequently affects the brain as well as the spinal cord, compared
to SCI, which affects only the spinal cord. In SCI, impaired
cognition would be due to some other comorbid illness such as
head injury at the time of spinal cord injury.). Despite these differences in pathophysiology, the WHO model would predict
similar patterns of disability across the two populations among
persons with a similar degree of organ system impairment (e.g.,
irrespective of their underlying diagnosis of SCI or MS, people
with loss of cognition would have a similar level of functional
disability, and people with quadriparesis would have a similar
level of functional disability).
The purpose of this study was to examine the disablement
M6l3
M614
HOENIG ET AL.
process among persons with two diseases of differing pathophysiology, MS and SCI. Research questions examined in this
study were:
1. Does the level of disability differ among persons with MS
compared to persons with SCI?
2. Do persons with MS versus SCI differ in level of disability
after stratifying by duration of disease or by specific physical
impairments?
METHODS
Patient Sample
The Paralyzed Veterans of America provided initial funding
to establish a National Registry of Spinal Cord Dysfunction
(SCD). Using Veterans Administration databases, 49,468 veterans with a high probability of SCD were identified and mailed
a questionnaire. Individuals who had not responded within 60
days were mailed a second questionnaire. The VA SCD
National Survey contained measures of disease, impairment,
and disability. The VA SCD National Survey and its development have been fully described previously, and its measures
have been shown to be valid and reliable (4-6). Disease measures on the VA SCD National Survey were self-reports of (i)
the cause or causes of their spinal cord condition using a check
list of diseases causing SCD, and (ii) the date of onset ofthe
condition. The duration of disease or injury was calculated from
the reported date of diagnosis or injury to 1995, and was categorized as 0-5 years, 6-10 years, 11-20 years, and >20 years
of duration.
Outcome Measures
Impairment in the VA SCD National Survey was measured
according to self-reported measures of sensory, cognitive, visual, and motor impairment, measures which we have shown to
be both reliable and valid (4,5). Sensory impairment was defined as full, some, or no feeling reported in affected areas.
Cognitive impairment was defined as any versus no reported
problems with thinking or memory. Visual impairment was defined as any versus no reported difficulty with vision.
Motor impairment was defined according to (i) the reported
number of affected limbs (no affected limbs reported, 1,2,3, or
4 affected limbs), (ii) the amount of useful movement reported
in affected limbs (full, some, no useful movement in affected
areas), and (iii) a composite index of the amount of overall
motor impairment (little, some, moderate, severe motor impairment, see Table 1 for definitions). These measures of motor impairment were suggested by a multidisciplinary panel convened
to help develop the VA SCD National Survey (4,6). Current
measures of impairment used in the SCI population (American
Spinal Injury Association [ASIA] Standards for Neurological
and Functional Classification) (7) and the MS population
(Kurtzke Functional Neurological Status Evaluation) (8) were
deemed inappropriate to this application for two reasons. The
ASIA measure is a pathophysiologically based measure developed for the SCI population and it does not necessarily apply to
nontraumatic pathology of the spinal cord that may affect multiple areas of the central nervous system (e.g., MS) which may
preclude a clear-cut "level of injury." Similarly some aspects of
the Kurtzke measure, which was developed with the patho-
physiology of MS in mind, do not apply to persons with SCI
(e.g., measures of central nervous system function like ataxia
and nystagmus are not pertinent to persons with SCI, even those
with concomitant head trauma). Secondly, both the ASIA and
the Kurtzke scales require assessment by a trained physician.
Measures of motor impairment designed for the geriatric population (e.g., 9,10) do not capture important differences in motor
function in a population that is largely wheelchair dependent
(e.g., ability to stand up or walk across a room will not distinguish among persons with paraplegia vs quadriplegia). The
panel therefore recommended a simple count of affected limbs,
amount of sensation, and amount of useful movement (this latter concept being similar to the ASIA scoring of motor function). The composite index of motor impairment was derived
from these as follows: persons with full, useful movement in
affected limbs have little motor impairment by definition (i.e.,
loss of sensation in affected limbs is not counted as motor impairment). Clinical judgment indicated that persons who reported that all four limbs were affected and that they had no
useful movement were most comparable to a person with highlevel, complete quadriplegia and represented the most severe
level of motor impairment. The remaining persons all had some
or no useful movement in the affected limbs. They were distinguished using clinical judgment into two groups based on the
number of affected limbs, in that persons who reported two or
fewer affected limbs likely were comparable to persons with incomplete or complete paraplegia and persons with three or
more affected limbs likely were comparable to persons with incomplete or complete, but low level (e.g., C 6-7) quadriplegia.
In fact, 71% of persons with SCI and 53% of persons with MS
who reported two limbs were affected reported that both legs
were affected, and 88% of persons with SCI and 81% of persons with MS who reported three or more limbs were affected
reported that all four limbs were affected.
The VA SCD National Survey measured disability with the
self-reported functional measure (SRFM), a reliable and valid
self-reported measure developed from the 13 motor questions
of the Functional Independence Measure (FIM) (4,5,11). The
SRFM uses four response levels instead of the traditional seven
response levels on the FIM (4 = No Help, 3 = Extra Time or
Tool, 2 = Some Help from Another Person, 1 = Never Do/Total
Help). Summing the 13 individual items provides the total
SRFM score, with total scores ranging from 13-52 (higher
score = better function). The SRFM score is not normally distributed. In fact, the summary score appears multimodal.
Therefore, in order to capture the overall shape of the distribution and preserve the ordinality inherent in the score, the SRFM
score was categorized into the following four ordinal groups:
13-22,23-32,33-42, and 43-52. These groupings represent
equal intervals over the range of possible values, and the group-
Table 1. Definitionsfor the Measure of Overall Motor Impairment
oLimbs
I Limbs
2 Limbs
3 Limbs
4 Limbs
Full Movement
Some Movement
No Movement
Little impairment
Little impairment
Little impairment
Little impairment
Little impairment
Some impairment
Some impairment
Some impairment
Moderate impairment
Moderate impairment
Some impairment
Some impairment
Some impairment
Moderate impairment
Severe impairment
DISABILITY FINGERPRINTS
ings reflect the multimodality of the data as well as preserving
its ordinality.
Data Analysis
Research Question 1 (whether or not there was a difference in
the pattern of disability among persons with SCI and compared
to persons with MS) was analyzed by comparing the distribution
of the SRFM score among the SCI versus the MS population
using the Mantel-Haenszel chi-square test for trend. This test
statistic is based upon ordinal rather than nominal categories and
recognizes the ordinal nature of the SRFM score (12). We examined the possibility of ordinal regression; however, the assumption of proportionality of odds was quite clearly violated,
making this approach untenable. The second question was addressed by determining if there were condition-specific differences in (a) the relationship of duration of illness to SRFM
score (tested with the Breslow-Day statistic) (12), and (b) proportion of the population with motor, sensory, visual, or cognitive impairment (Mantel-Haenszel chi-square test for trend).
The MS and SCI patient samples then were stratified on duration of the SCD and on each of the impairment variables, and
the distribution of disability in the two samples compared using
the Mantel-Haenszel chi-square test for trend.
RESULTS
A total of 18,038 unique responses were received from persons who reported they had SCD. This analysis uses the 8150
veterans out of this total population who reported either MS
(n = 1789) or SCI (n = 6361) as the sole cause of their SCD,
and who completed all survey items on impairment and disability. Specifically, of the 18,038 survey respondents, 12,447 indicated they had SCI, 8346 of these had SCI alone, 1985 of these
were excluded because they had not answered one or more of
our questions on function or on impairment, leaving 6361 persons in the SCI sample. The SCI sample had a mean age of 52.5
years, mean disease duration was 20 years (Pearson correlation
coefficient between age and disease duration = .52),2% were
female, and 81% were white. Three thousand eight hundred
and nine survey respondents indicated they had MS, 2590 of
these had MS alone, 801 of these were excluded because they
had not answered one or more of our questions on function or
on impairment, leaving 1789 in the MS sample. The MS sam-
M615
ple had a mean age of 54.5 years, mean disease duration was
20 years (Pearson correlation coefficient between age and disease duration was .53), 9% were female and 90% were white.
Patterns of disability differed significantly among persons
with MS compared to persons with SCI (p = .001) (Table 2).
Overall, a greater proportion of the MS sample scored in the
lowest SRFM category (the category with the greatest functional deficits) compared to the SCI sample (34% compared to
22%). The proportionately greater level of dependency reported
by the MS sample holds for almost all types of basic activities
of daily living (Figure 1). There was little difference in mean
age or mean disease duration between the two conditions; yet,
as predicted by the progressive nature of MS, there was a statistically significant interaction between disease and duration of
illness and SRFM score (p = .001). That is, soon after onset,
people with SCI were more disabled than people with MS. The
longer the time since onset, the more likely that people with
MS would have greater disability than people with SCI.
However, the differences in pattern of disability across the two
conditions remained statistically significant after stratification
on years since onset of the SCI. Irrespective of years since diagnosis, SCI always showed a multimodal pattern with the largest
mode in the highest SRFM functional quartile. On the other
hand, among persons with MS for more than 10 years, the
largest mode was at the lower end of function.
As expected, there were substantial, statistically significant
differences between the two conditions in the amount and kinds
of physical impairment (Figure 2). The distribution of number
of affected limbs is different in MS versus SCI (p = .001 ). A
greater proportion of persons with MS, compared to SCI, reported that all four limbs were affected by their condition (52%
vs 33%) and fewer persons with MS reported that just two
limbs were affected (29% vs 51%). However, compared to MS,
SCI more frequently caused complete loss of motor and sensory abilities in the affected limbs. Forty-two percent of the SCI
population reported having no useful movement (42%) and no
sensation (36%) compared to persons with MS, of whom 26%
reported no useful movement and 9% reported no sensation
(p = .001 for differences in motor and for differences in sensory impairment). However, when amount of useful movement
in the limbs was coupled with the number of affected limbs, as
a rating of overall motor impairment (see Table 1 for defini-
Table 2: Distribution of Disability Among Persons With MS and SCI, Overall and Stratified by Disease Duration,
as Measured by the Percent of the Sample With Differing Levels of Self-Care Function (SRFM Score)
MS (n = 1789)t
SCI (n =6361)t
SRFM Score*
13-22
23-32
33-42
43-52
13-22
23-32
33-42
43-52
p valuer
Overall
21.8
11.8
22.8
43.7
34.1
13.3
21.8
30.8
.001
Duration
0-5 years
6-10 years
11-20 years
>20 years
21.0
24.9
25.3
19.9
8.4
10.5
10.7
14.0
21.3
19.9
22.7
24.9
49.3
44.7
41.3
41.2
9.2
14.4
33.2
45.8
11.9
11.0
15.9
12.5
21.1
27.7
23.7
19.7
57.8
47.0
27.2
22.0
.001
*Higher SRFM score =better function.
t458 persons with SCI (7.2%) and 110 persons with MS (6.1 %) were missing data for date of injury/disease onset.
:j:Mantel-Haensze1 chi-square test for trend.
HOENIG ET AL.
M6l6
100
, --
90
80
~~
li>' :::
70
"J, :::
:::
60
.....
=
..
<J
",
,"
",
.::.
,"
",
50
-::.
,"
":';
,"
",
,"
",
Q.,
30
~
1-
20
10
0
!
io
.....:
:::
"M¥
:-:.
"
"
"-::.
",
,"
",
,"
..::.
",
""
,"",
......
,"
,"
",
,"
",
,"
",
,"
",
,"
",
,"
..":
r:
",
; ~~:~':
,"
..-:.
. :!
",
~~;.:.
-;:.
",
,"
::
:
",
,"
:::
",
",
.::"
::~
",
,"
",
,"
",
,"
-:...
~'
",
,"
.
r:::'
-::: '::,
to,:," "::'
i'
"
......
,"
r ,"
",
",
,"
",
,"
.:.:
,"
,"
"
.::"
-::.
,"
40
r:::'
,"
",
,"
",
,"
",
-;:.
to'!
"",
,"
",
5~
"
,"
",
.::,.
.::"
e- c:
,"
j
-
,•
",
,"
",
,"
",
,"
"
:~~
:::
..
",
,"
",
-::.
,"
, .::.
~~
~... ",
,"
,"
' " ,
,"
'«; ,"
",
I~:
":';
,"
",
,"
",
,"
",
,"
",
,"
~~
-, ,"
;(
":';
r:::'
,"
",
,"
",
..":.
,"
",
,"
",
,"
~ ",
Y:
j::
:::
",
"~::
,"
»
e",
»
",
r:::'
,"
",
,"
",
,"
",
,"
",
,"
",
,"
",
::
:
",
....
-::.
,"
.~%
::.
....:.
,"
",
,"
,"
,"
",
,"
",
,"
",
",
,"
",
,"
",
,"
",""",
,"
.::",.
, ":::
",
,"
:.:.
",
","" ,
;.:.
::,:
::
,:
",
",
,"
",
,"
",
,"
-;:.
",
......
,"
":';
,"
",
",
,"
",
,"
",
,"
.::,.
,"
",
",
.;:.
,"
",
,"
:::
",
.......
,"
",
~~
",
-::.
,"
",
,"
s-,
,"
",
,"
",
,"
",
,"
- .::"
",
,"
",
,"
",
.;:.
":.:
,"
0:::.
,"
",
,"
",
,"
",
,"
"0":
",
......
,"
",
,"
t ,> ,"
",
~.
.,
",
":';
":';
c;
"",
:::
,"
",
,"
",
,"
",
,"
",
,"
",
,"
",
:.::.
",
,"
",
o:
,"
",
,"
",
,"
",
.::.
,"",
,"
",
.0
:
r:::'
,"
",
......
,"
",
"
",
,"
";.-:
",
:::
",
",
,"
",
,"
",
,"
",
,"
",
,"
",
,"
:.:
:
",
,"
",
";.":
",
","",
,"
,"
",
,"
",
,"
",
";.::",
,"
. -;:.
" ,
;.:
",
,"
",
.0:
Iffiru SCI
I
MS
,
,"
",
,"
",
,"
Figure 1. Distribution of specific depend encies in activities of daily living among perso ns with MS and SCI, measured as the percent of the sample needing some
help from another person , total help from another person, or never performin g the activity (compared to persons reportin g co mplete independence or usc of ext ra
time or a tool to perform the activity).
tions), a greater proportion of the MS sample fell into the moderate and severe categories of motor impairment compared to
the SCI sample, Sixteen percent of persons with MS had severe
motor impairment compared to 9% severe impairment in the
SCI popu lation , 44 % of MS patients had moderate impairment
compared to 27 % moderate impairment among persons with
SCI (p =,001). Additionally, a substantially greater proportion
of MS sample, compared to the SCI sample, report visual impairment and cognitive impairment Sixty-seven percent of people with MS reported problems with their vision compared to
only 6% of the SCI population ( p ,00 1), and 52% reported
problems with their memory and thinking compared to only
13% of the SCI population (p = .00 1).
Although there were statistically significant differences in
amount and kinds of impairment across the two conditions,
these difference s did not fully account for the differences in disability across the two conditions. Table 3 shows the distribution
of disability across the two patient samples after stratifying according to specific impairments, There still were highly significant differences in disability between the MS and SCI sam ples
after stratifying amount of sensation (p = .001), memory and
=
thinking (p =,(01), vision (p =.00 1), number of affected limbs
=.(03), amount of useful movement ( p =,(0 1), and overall
motor impairment ( p = .00 3) . For example, 93.2% of the
severely motor impaired MS persons were in the lowest functional category compared to 76.4% of severely motor impaired
SCI persons; 71% of the MS population who reported they lack
sensation in affected limbs were in the lowest SRFM category
compared to 23,5% of the SCI population who reported a lack
of sensation; 34,7% of MS persons with cognitive impairment
and 36.8 % of those with visual impairment were in the lowest
functional category compared to 22.1 % and 21.8% of the SCI
population with visual and cognitive impairment. Thu s, across
all categories of impainnent and after stratifying for disease duration, persons with MS report greater functional dependency
compared to persons with SCI.
(p
DISCUSSION
This study shows there are statistically significant differences
in pattern of disability across two populations with differing
conditions affecting the central nervous system. The two conditions indeed have unique disability fingerprints, However, the
M617
DISABILITY FINGERPRINTS
100
90
60
70
60
....'"
.
C
IS SCI
50
@ MS
c..
i
40
30
20
10
0
~.,
~.,
v.. .~
.
-,
..:.,.s:'
I::l
~.,
..:.,.s:'
..:.,....<S'
"J
""
~.,
..:.,.s:'
~
100
90
60
,
70
....'"
.
60
C
-;:
-:-:-:-:.;.
-:.
50
~
40
.
'
~.
c.~.
20
10
~.
~.
I
, -c:
0
it c.
.
c.
~.
c.
~.
c.-::
.::
-::
.::
'Co
~
c.~
c.
~:
-:0
-:0
~.
~.
~ ".
~:
~.
..
-:"
'
....
~
~
~.
-;.
~.
~.
~.
.'
~
,-;.
-;:
-;.
..":
~
-;.
~
-;.
-::
c
c.~.
-::
.
!O
-:0
-;.
-:.
0;.
0;.
-;:
-;:
.'
~
~,.
.::
~.
~.
~
.'
[I !..'
1m
SCI I
@MS
-::
~
~
-;.
-;:
0;:
«
.::
c.
-;:
.'-;"~
-;:
c.
f-: .
-;:
•-:0c.
0;:
.::
~.
-::
30
0;.
-:.
-: .
-:"
-:-
r
e,
-::
c.
!.
~
~.
~
~.
~.
-;:
-::
~.
-::
."
Figure 2. Distribution of specific imp airm ents among persons with MS und SCI, as measured by the perce nt of the sample reponing the impairment , There was a
statistically significant difference between the MS and SCI popula tion at .001 level using the Mantel-Haenszel chi-square test for trend for all of the categories of impairment depicted.
HOENIG ET AL.
M618
Table 3. Distribution of Disability Among Persons With SCI and MS, as Measured by the Percent of the Sample
With Differing Levels of Self-Care Function (SRFM Score), Stratified According to Specific Impairments
SCI (n = 6361)
MS (n = 1789)
13-22
23-32
33-42
43-52
13-22
23-32
33-42
43-52
P value]
Overall
21.8
11.8
22.8
43.7
34.1
13.3
21.8
30.8
.001
Limbs Affected
0
1
2
3
4
10.7
2.5
6.3
17.6
51.8
6.3
4.7
10.1
15.1
16.3
9.3
12.6
29.3
28.8
16.1
73.7
80.1
54.4
38.5
15.8
26.1
2.3
13.2
30.5
48.8
6.1
13.6
13.8
14.3
13.2
4.6
20.5
26.8
26.0
19.3
63.1
63.6
46.3
29.1
18.7
.003
Movement
Full
Some
None
1.7
24.7
23.3
1.2
13.7
12.1
6.8
21.7
27.8
90.2
39.8
36.9
3.5
21.5
77.6
1.1
16.4
10.0
8.1
29.3
8.3
87.4
32.9
4.1
.001
Motor Impairment
Little
Some
Moderate
Severe
1.7
7.3
40.7
76.4
1.2
10.5
18.8
10.7
6.8
28.9
21.1
8.1
90.2
53.4
19.5
4.7
3.5
15.8
31.2
93.2
1.1
15.2
18.1
3.7
8.1
27.8
28.1
2.4
87.4
41.3
22.6
0.7
.003
Feeling
Full
Some
None
10.1
22.7
23.5
4.3
12.5
12.6
9.8
21.5
28.0
75.8
43.3
35.9
26.8
31.6
71.2
8.7
14.5
11.1
14.4
24.6
11.8
50.2
29.3
5.9
.001
Memorytrhinking
Affected
Not Affected
22.1
19.9
11.7
12.6
22.7
22.8
43.5
44.7
34.7
33.6
12.0
14.4
20.1
23.4
33.2
28.6
.001
Eyes
Affected
Not Affected
21.8
21.5
11.7
12.5
22.7
23.3
43.7
42.7
36.8
32.9
10.6
14.5
20.2
22.6
32.4
30.0
.001
SRFMScore*
*Higher SRFM score = better function.
tMantel-Haenszel chi-square test.
etiology for the differences in the pattern of disability among
the two populations does not appear to be solely influenced by
the known differences in pathophysiologies of the two conditions. Consistent with the progressive nature of MS versus the
nonprogressive nature of SCI, the level of disability appears to
be comparatively more unstable over time for persons with MS
than for persons with SCI. However, stratification on years
since onset does not eliminate the population differences in
level of disability. The differences seen in physical impairments
also are consistent with the known pathophysiologies of the
two conditions. The instances of reported visual and cognitive
impairment among persons with SCI likely are due to concomitant head injury or comorbid illness and, thus, are relatively less
common in the SCI sample compared to the MS sample
wherein the disease itself affects visual and cognitive function.
The multifocal nature of the MS disease process also likely accounts for the greater prevalence of impairment in all four limbs
among persons with MS compared to SCI. However, the differences in level of disability between the two conditions persist
after stratification according to visual impairment, cognitive
impairment, as well as motor and sensory impairment. These
findings provide additional evidence for the disability fingerprints hypothesis. Efforts by the World Health Organization to
modify their model for disability should account for the extent
to which the disablement process appears to be unique for different diseases and conditions, and the factors that determine
that uniqueness.
The findings in this study supplement previous studies that
focused solely on showing the existence of a relationship between diseases, impairments, and disability, and did not examine the extent to which impairment alone accounted for the
amount of disability in the populations studied, or whether that
relationship differed for differing diseases. For example, in a
cohort of community dwelling older persons, poor performance
in lower extremity motor tasks was associated with a fourfold
greater increased risk of disability 4 years later (10). However,
we do not know what other factors may have influenced which
persons with lower extremity limitations did or did not develop
disability. Correlations as high as 80% have been reported between measures of motor impairment and functional disability
in persons recovering from acute SCI (13). However, correlations are somewhat lower among persons with chronic SCI.
Curt and colleagues showed that ambulatory capacity among
persons with chronic SCI has a correlation of .61 with overall
motor impairment (14). We do not know what factors may
account for differences in functional capacity among chronic
versus recently injured SCI patients. Because much of our treatment for disability is based on the medical model and is di-
DISABILITY FINGERPRINTS
rected at specific diseases and impairments (e.g., nonsteroidal
anti-inflammatory agents for inflammatory arthritis, exercises
to increase strength), findings which highlight the relationship
between disease, impairment, and disability are reassuring to
practitioners using the medical model. However, the question
remains, how good is this model and might there be other nonmedical factors that influence the likelihood of disability? For
example, some studies have shown that psychosocial factors
may influence the likelihood of being disabled (15-17). Our
results indicate that although physical impairments may vary
for conditions with differing pathophysiologies, populationspecific differences in physical impairment do not fully account
for population-specific differences in disability. As such, our
findings emphasize the need to more fully explain the disablement process.
This study has several limitations. Factors unmeasured in
this study that affect disability may be present differentially
among persons with MS versus SCI. For example, impoverishment may affect ability to cope with physical impairment, for
example by affecting availability of technological and personal
assistance. If, for example, veterans with SCI were more likely
to be eligible for health care in the VA on the basis of war injury, but veterans with MS were more likely to be eligible for
health care in the VA due to impoverishment (rather than from
a condition connected to their military service), this would
greatly influence eligibility for a variety of veterans' benefits,
including equipment and home modifications (18). As another
example, the incremental nature of MS compared to the static
nature of SCI may affect ability to adapt and compensate for
specific impairments. Our sample was derived from the veteran
population, which differs in important ways from the general
population. For example, among SCI patients in the Model
Systems, 85% were male compared to 98% of our sample, 71%
were white compared to 81 % of our sample. Although we know
of no data that indicate either sex or race are associated with
functional outcomes in the SCI or MS populations, our findings
cannot necessarily be generalized to the nonveteran population.
Given the fair correlation between age and disease duration in
the MS and SCI populations in this study, we cannot rule out the
possibility that some unknown age-disease-specific interaction
is accounting for the differences in disability seen in the two
conditions. Stratification alone on either age or duration will not
definitively answer that question; rather, a longitudinal study design would be required. Although some investigators (19) have
noted age to be an independent predictor of functional decline
in the SCI population (most likely due to age-related acquisition of comorbid conditions like osteoarthritis, heart disease),
we know of no data to indicate that such conditions would be
acquired differentially in SCI versus MS. Finally, there may be
important interactions between impairments, which together
differentially affect the likelihood of disability in the two populations, which we were unable to examine in this study due to
sample size limitations. These are important directions for
future research.
Some of our findings are surprising. For example, similar
rates of disability are seen among persons with and without
cognitive impairment and visual impairment. Speculating on
the reason for this brings up several possibilities: there may be a
dose-response relationship between cognitive impairment (or
visual impairment) and function that was covered up by use of
M619
dichotomous measures for these variables. Basic activities of
daily living are the main functional tasks measured by the
SRFM and these may be less affected by visual and cognitive
impairment than the more complex instrumental activities of
daily living. Finally, patients often self-report less disability
than observed by providers completing a performance-based
measure, in which case only the most potent relationships
would be detected (9).
In the SCI and MS populations, performance measures of
impairment and disability are common [e.g., ASIA Standards
for Neurological and Functional Classification, Kurtzke
Functional Neurological Status Evaluation, the Functional
Independence Measure (7,8,20)]. Direct comparison between
our self-reported measures and these performance-based measures is not appropriate, in that self-reported measures are
widely believed to assess somewhat different, although related,
constructs to performance-based measures (9). This is because
self-reported measures allow the respondent to incorporate
other parameters (e.g., depression, environmental barriers) and
therefore may lack specificity. Performance-based measures of
impairment and disability are less affected by subjective factors
and therefore might show a tighter correlation with one another
than self-reported measures do with one another. However, selfreported measures have the benefit of being grounded in the patient's reality rather than an artificial laboratory setting. Thus,
although this study provides important epidemiological information about impairment and disability in the MS and SCI population, it cannot be directly compared to similar information
derived from performance-based data (e.g., 19,21).
Functional decline may be the final outcome for diverse diseases, but the paths for functional decline appear to be complex
and unique for individual diseases. It is said that all rivers lead to
the ocean, but experience shows that they do so by markedly differing routes. Analogously, functional decline may be represented
more accurately by rapids with differing declivities and obstacles
rather than as a single whirlpool. Now we need to map more accurately the pathways to disability, including the medical and
nonmedical, extrinsic and intrinsic modifiers of disability.
ACKNOWLEDGMENTS
This study was supported in part by Spinal Cord Research Foundation Grant
No. 1632 from the Paralyzed Veterans of America. Dr. Hoenig is now a Paul
Beeson Faculty Scholar in Aging Research. We would like to express our gratitude to Mr. Michael Zolkewitz for his assistance with data management.
Address correspondence to Helen Hoenig, MD, Physical Medicine and
Rehabilitation Service (117), Durham VA Medical Center, 508 Fulton Street,
Durham, NC 27705. E-mail: [email protected]
REFERENCES
I a. World Health Organization. International Classification of Impairments,
Disabilities, and Handicaps. Geneva: World Health Organization; 1990.
lb. Ferrucci L, Guralnik 1M, Simonsick E, Salive ME, Corti C, Langlois 1.
Progressive versus catastrophic disability: a longitudinal view of the disablement process. 1 Gerontal Med Sci. 1996;51AM I 23-M 130.
Ie. Fried LP, Guralnik JM. Disability in the older adult: evidence regarding
significance, etiology, and risk. lAm Geriatr Soc. 1997;45:92-100.
2. Tinetti ME, Inouye SK, Gill TM, Doucette JT. Shared risk factors for
falls, incontinence, and functional dependence. lAMA. 1995;273:
1348-1353.
3. Guccione AA, Felson DT, Anderson Jl, et al. The effects of specific medical conditions on the functional limitations of elders in the Framingham
Study. Am 1 Public Health. 1994;84:351-358.
M620
HOENIG ET AL.
4. Hoenig H, McIntyre L, Sloane R, Branch LG, Truncali A, Homer R. The
reliability of a self-reported measure of disease, impairment, and function
in persons with spinal cord dysfunction. Arch Phys Med Rehabil. 1998;79:
378-387.
5. Hoenig H, Branch L, McIntyre L, Hoff J, Homer R. The validity in persons with spinal cord injury of a self-reported functional measure (the
SRFM) derived from the HM. Spine. 1999;24:539-544.
6. Samsa G, Hoenig H, Homer R, Sloane R. Developing a registry for veterans with spinal cord dysfunction. Spinal Cord. 1998;36:57-62.
7. Ditunno JF. ASIA Standards for Neurological and Functional Classification of Spinal Cord Injury. Chicago, IL: American Spinal Injury
Association; 1992.
8. Kurtzke JE On the evaluation of disability in multiple sclerosis.
Neurology. 1998;50:317-326. Reprint from Neurology 1961;18.
9. Reuben DB, Valle LA, Hays RD, Siu AL. Measuring physical function in
community-dwelling older persons: a comparison of self-administered,
interviewer-administered, and performance-based measures. JAm Geriatr
Soc. 1995;43:17-23.
10. Guralnik lM, Ferrucci L, Simonsick EM, Salive ME, Wallace RB. Lower
extremity function in persons over the age of 70 years as a predictor of
subsequent disability. N Engl J Med. 1995;332:556--561.
11. Heinemann AW, Linacre JM, Wright BD, Hamilton BB, Granger C.
Relationships between impairment and physical disability as measured by
the functional independence measure. Arch Phys Med Rehabil. 1993;
74:566--573.
12. Agresti A. Categorical Data Anlaysis. New York: John Wiley and Sons;
1991.
13. Saboe LA, Darrah lM, Pain KS, Guthrie J. Early predictors of functional
independence 2 years after spinal cord injury. Arch Phys Med Rehabil.
1997;78:644-650.
14. Curt A, Keck ME, Dietz V. Functional outcome following spinal cord injury: significance of motor-evoked potentials and ASIA scores. Arch Phys
Med Rehabil. 1998;79:81-86.
15. Cunningham LS, Kelsey JL. Epidemiology of musculoskeletal impairments and associated disability. Am J Public Health. 1984;74:574-579.
16. Hughes SL, Edelman PL, Singer RH, Chang RW. Joint impairment and
self-reported disability in elderly persons. J Gerontal Soc Sci.
1993;48:S84-S92.
17. Badley EM, Ibanez D. Socioeconomic risk factors and musculoskeletal
disability. J Rheumatol. 1994;21:515-522.
18. Department of Veterans Affairs. Federal Benefits for Veterans and
Dependents. Washington, DC: US Government Printing Office; 1988.
19. Menter RR, Whiteneck GG, Charlifue SW, et aI. Impairment, disability
handicap and medical expenses of persons aging with spinal cord injury.
Paraplegia. 1991;29:613--619.
20. Granger CV, Cotter AC, Hamilton BB, Fiedler RC, Hens MM. Functional
assessment scales: a study of persons with multiple sclerosis. Arch Phys
Med Rehabil. 1990;71:87G--875.
21. Devivo MJ, Rutt RD, Black KJ, Go BK, Stover SL. Trends in spinal cord
injury demographics and treatment outcomes between 1973 and 1986.
Arch Phys Med Rehabil. 1992;73:424-430.
Received July 3 I, 1998
Accepted April 3, 1999
3 BC/BE Geriatricians or IM/FP's
Utah-Salt Lake City and Ogden
#1 Best Place to Live in the Country - geographically and culturally. Locations are ideally
suited to raising families or attaining one's quest
for unsurpassed outdoor recreation!
3 BC/BE geriatricians or IM/FP's with strong interest/experience in geriatrics sought for hospital
affiliated senior centers. Employment with excellent compensation/benefits/relocation package.
Call NOW! PPS 800·824·9275. Sorry no
visa support available.