Atorn meniscus is one of the most common reasons why patients undergo
arthroscopic surgery. Various instruments have been used to measure the
outcomes of treatment of meniscal injuries of the knee. These have included
the Tapper and Hoover system, Knee Injury and Osteoarthritis Outcome Score
(KOOS), Lysholm knee score, International Knee Documentation Committee (IKDC)
subjective knee form, Cincinnati knee rating scale, and Tegner activity
scale1-6.
The Lysholm knee score is a condition-specific outcome measure that
contains eight domains: limp, locking, pain, stair-climbing, use of supports,
instability, swelling, and
squatting2. An
overall score of 0 to 100 points is calculated, with 95 to 100 points
indicating an excellent outcome; 84 to 94 points, a good outcome; 65 to 83
points, a fair outcome; and <65 points, a poor outcome. Originally designed
for assessment of ligament injuries of the knee, the Lysholm knee score has
been used for a variety of knee
conditions2,7-9.
It was recently validated for the evaluation of patients with chondral
disorders of the
knee8 and those with
an acute patellar
dislocation9.
The Lysholm score was designed to document the patient's evaluation of
function. Following publication of the Lysholm score, Tegner and Lysholm
concluded that, in addition to a functional score, a score for activity level
needed to be measured on a separate scale. The Tegner activity scale is a
numerical scale ranging from 0 to 106. Each value indicates the
ability to perform specific activities. An activity level of 10 corresponds to
participation in competitive sports, including soccer, football, and rugby at
the elite level; an activity level of 6 points corresponds to participation in
recreational sports; and an activity level of 0 is assigned if a person is on
sick leave or receiving a disability pension because of knee problems. An
activity level of 5 to 10 is recorded only if the patient participates in
recreational or competitive sports. Several authors have examined the
psychometric properties of the Tegner activity
scale9,10.
Since both that scale and the Lysholm knee score were developed for the
assessment of patients with ligament injuries, it is unclear whether they are
appropriate for evaluation of outcomes in patients with other knee conditions.
Studies in which more than one score was used to assess the outcomes of
meniscal surgery have shown different results depending on which score was
utilized11,12.
To decide how to interpret these scores, it is important to study the
psychometric properties of the individual scores when they are used for the
assessment of various knee conditions.
It is essential to use outcome instruments whose psychometric properties
have been vigorously established. The important psychometric properties of an
outcome instrument include reliability, validity, and
responsiveness1,13.
Reliability refers to the reproducibility of an outcome measure, either
between subjects (test-retest reliability) or between observers (interobserver
reliability). Validity is a measure of whether an outcome instrument actually
measures what it intends to measure. Components of validity include content
validity ("face" validity and floor and ceiling effects),
criterion validity (how an instrument compares with an accepted "gold
standard" instrument), and construct validity (whether the instrument
produces results that conform with expected, noncontroversial hypotheses).
Responsiveness is the ability of the instrument to reflect change over time or
after treatment.
The purpose of this institution-review-board-approved study was to
determine the psychometric properties of the Lysholm knee score and the Tegner
activity scale when those instruments were used to evaluate patients with a
meniscal lesion of the knee.
Test-Retest Reliability
Test-retest reliability was studied in a group of 122 patients who had a
variety of arthroscopically documented meniscal disorders in the knee, with
some having an isolated lesion and others having additional lesions of the
knee. The mean age of these patients was forty-eight years (range, fourteen to
seventy-six years); there were seventy-seven male and forty-five female
patients. A medial meniscal lesion was seen in fifty-nine (48%) of the knees;
a lateral meniscal lesion, in forty (33%); and a medial and lateral meniscal
lesion, in twenty-three (19%). Twenty-eight patients had an isolated meniscal
lesion with no other ligament or chondral surface disorder.
The Lysholm score and Tegner activity level (see Appendix) were measured at
a minimum of two years postoperatively and then again within the following
four weeks. There was no change in the patient's postoperative condition in
the four-week interval, as determined by a review of a patient questionnaire,
history form, and reinjury status. The intraclass correlation coefficient was
determined for the overall Lysholm score, the domains of the Lysholm score,
and the Tegner activity scale. An intraclass correlation coefficient of
>0.70 was considered
acceptable13.
The standard error of the measurement was calculated as described
previously14. This
value can be used to determine the 95% confidence interval for individual
scores, which provides an estimate of where the actual score may lie. To
further define this interval, the minimum detectable change can be calculated
to determine the smallest change that can be considered a true difference
after measurement error and noise have been taken into
account14.
Internal Consistency and Content Validity
Internal consistency and content validity were determined in two groups.
One group had an isolated meniscal lesion—i.e., no other ligament or
chondral lesion was diagnosed at arthroscopy—and the other group had a
meniscal lesion as well as additional intra-articular knee disorders. Complete
Lysholm scores were calculated for 191 patients with an isolated lesion. Their
mean age was forty years (range, thirteen to eighty-one years), and there were
129 male patients (68%) and sixty-two female patients (32%). A medial meniscal
lesion was seen in 117 (61%) of the 191 knees, a lateral meniscal lesion was
seen in sixty (31%), and a medial and lateral meniscal lesion was seen in
fourteen (7%). Because the Tegner activity level was added to the
questionnaire after collection of the Lysholm scores had already begun, the
Tegner score was determined for only eighty patients with an isolated meniscal
lesion. The mean age of those eighty patients was forty-two years (range,
sixteen to eighty-one years), and there were thirty female patients (38%) and
fifty male patients (63%). A medial meniscal lesion was seen in forty-two
knees (53%), a lateral meniscal lesion was seen in twenty-nine (36%), and a
medial and lateral meniscal lesion was seen in nine (11%).
The group with a meniscal lesion and associated intra-articular disease
consisted of 477 patients with a mean age of thirty-nine years (range,
eighteen to sixty-two years); 367 (77%) of the patients were male, and 110
(23%) were female. An anterior cruciate ligament procedure was performed in
120 patients in this group, and 261 patients had had at least one previous
operation (range, one to eight operations) on the knee.
The Lysholm score, Tegner activity level, demographic data, and patients'
subjective assessments were recorded preoperatively and at least one year
postoperatively, and these measurements were maintained in a computerized
database. Preoperative Lysholm knee scores were used to establish internal
consistency. Overall internal consistency for all eight domains was
determined. A Cronbach alpha of >0.70 was considered
acceptable13.
Preoperative Lysholm knee scores and Tegner activity levels were also used to
establish content validity. Floor effects (percentage of lowest possible
score) and ceiling effects (percentage of highest possible score) were
determined for the overall Lysholm score, for the eight domains of the Lysholm
score, and for the Tegner activity scale. Floor and ceiling effects of <30%
were considered
acceptable13.
Criterion Validity
The criterion validity was determined in the group of 477 patients with a
meniscal lesion and other associated intra-articular lesions, as described
above. Preoperatively, these patients completed the Short Form-12 (SF-12)
Health Survey in addition to the Lysholm score and Tegner activity scale
questionnaires. Correlation of the overall Lysholm score and the Tegner
activity scale with the physical score of the SF-12 Health Survey was
performed to establish criterion validity. The Pearson correlation coefficient
was used for the continuous variables.
Construct Validity
Construct validity was studied in the group with an isolated lesion, in
which 191 patients had complete Lysholm knee scores and eighty had Tegner
scores as well. The preoperative Lysholm scores and Tegner activity levels
were used to establish construct validity. Eight hypotheses (constructs) were
developed by consensus and were tested in this group of patients
(Table I). The constructs were
developed on the basis of a literature review and were included if they were
agreed on and considered to be noncontroversial by the three participating
clinicians.
The eight constructs for the Lysholm knee score were:
Patients with lower activity levels would have lower Lysholm scores. The
activity level was measured on an ordinal scale ranging from 1 to 5, according
to the criteria of the
IKDC1, with 1
indicating an inability to perform activities and 5 indicating the performance
of very strenuous activities. Analysis of variance was used to determine
significance, with eta2 used as the measurement of association.Because of the pain, swelling, possible use of crutches, and overall
disability associated with an acute knee injury, patients with such an injury
would have a lower mean Lysholm score than would patients with a chronic knee
injury. For this study, a chronic knee injury was defined as more than four
weeks between the injury and the surgery. The independent-samples t test was
used to determine significance.Patients with a Workers' Compensation claim would have a lower mean Lysholm
score than would patients without a Workers' Compensation claim. The
independent-samples t test was used to determine significance.Patients with more difficulty with activities of daily living would have
lower Lysholm scores than would patients with less difficulty with activities
of daily living. Difficulty with activities of daily living was measured on an
ordinal scale ranging from 1 to 10, with 1 indicating an inability to perform
activities of daily living and 10 indicating no difficulty with activities of
daily living. The Spearman rho test was used to determine significance.Patients with more difficulty working because of problems with the knee
would have lower Lysholm scores than would patients with less difficulty
working because of prob-lems with the knee. Difficulty working because of
problems with the knee was measured on an ordinal scale ranging from 1 to 10,
with 1 indicating an inability to work because of problems with the knee and
10 indicating no difficulty working because of problems with the knee. The
Spearman rho test was used to determine significance.Patients with more difficulty participating in sports activity because of
problems with the knee would have lower Lysholm scores than would patients
with less difficulty participating in sports activity because of problems with
the knee. Difficulty participating in sports activity because of problems with
the knee was measured on an ordinal scale ranging from 1 to 10, with 1
indicating an inability to participate in sports activity because of problems
with the knee and 10 indicating no difficulty participating in sports activity
because of problems with the knee. The Spearman rho test was used to determine
significance.Patients who assessed their overall knee function as abnormal or severely
abnormal would have a lower mean Lysholm score than would patients who
assessed it as normal or nearly normal. The independent-samples t test was
used to determine significance.Patients with a degenerative and/or complex meniscal tear would have a
lower mean Lysholm score than would patients with a simple tear of the
meniscus. The independent-samples t test was used to determine
significance.
Patients with lower activity levels would have lower Lysholm scores. The
activity level was measured on an ordinal scale ranging from 1 to 5, according
to the criteria of the
IKDC1, with 1
indicating an inability to perform activities and 5 indicating the performance
of very strenuous activities. Analysis of variance was used to determine
significance, with eta2 used as the measurement of association.
Because of the pain, swelling, possible use of crutches, and overall
disability associated with an acute knee injury, patients with such an injury
would have a lower mean Lysholm score than would patients with a chronic knee
injury. For this study, a chronic knee injury was defined as more than four
weeks between the injury and the surgery. The independent-samples t test was
used to determine significance.
Patients with a Workers' Compensation claim would have a lower mean Lysholm
score than would patients without a Workers' Compensation claim. The
independent-samples t test was used to determine significance.
Patients with more difficulty with activities of daily living would have
lower Lysholm scores than would patients with less difficulty with activities
of daily living. Difficulty with activities of daily living was measured on an
ordinal scale ranging from 1 to 10, with 1 indicating an inability to perform
activities of daily living and 10 indicating no difficulty with activities of
daily living. The Spearman rho test was used to determine significance.
Patients with more difficulty working because of problems with the knee
would have lower Lysholm scores than would patients with less difficulty
working because of prob-lems with the knee. Difficulty working because of
problems with the knee was measured on an ordinal scale ranging from 1 to 10,
with 1 indicating an inability to work because of problems with the knee and
10 indicating no difficulty working because of problems with the knee. The
Spearman rho test was used to determine significance.
Patients with more difficulty participating in sports activity because of
problems with the knee would have lower Lysholm scores than would patients
with less difficulty participating in sports activity because of problems with
the knee. Difficulty participating in sports activity because of problems with
the knee was measured on an ordinal scale ranging from 1 to 10, with 1
indicating an inability to participate in sports activity because of problems
with the knee and 10 indicating no difficulty participating in sports activity
because of problems with the knee. The Spearman rho test was used to determine
significance.
Patients who assessed their overall knee function as abnormal or severely
abnormal would have a lower mean Lysholm score than would patients who
assessed it as normal or nearly normal. The independent-samples t test was
used to determine significance.
Patients with a degenerative and/or complex meniscal tear would have a
lower mean Lysholm score than would patients with a simple tear of the
meniscus. The independent-samples t test was used to determine
significance.
The eight constructs for the Tegner activity scale were:
Patients with moderate-to-severe pain with activity would have a lower mean
Tegner activity level than would patients with no or mild pain with activity.
The independent-samples t test was used to determine significance.Patients with moderate-to-severe difficulty running would have a lower mean
Tegner activity level than would patients with no or mild difficulty running.
The independent-samples t test was used to determine significance.Patients with more difficulty with activities of daily living would have
lower Tegner activity levels than would patients with less difficulty with
activities of daily living. The Spearman rho test was used to determine
significance.Patients with more difficulty working because of problems with the knee
would have lower Tegner activity levels than would patients with less
difficulty working because of problems with the knee. The Spearman rho test
was used to determine significance.Patients with more difficulty participating in sports activity because of
problems with the knee would have lower Tegner activity levels than would
patients with less difficulty participating in sports activity because of
problems with the knee. The Spearman rho test was used to determine
significance.Patients who assessed their overall knee function as abnormal or severely
abnormal would have a lower mean Tegner activity level than would patients who
assessed it as normal or nearly normal. The independent-samples t test was
used to determine significance.Because of the pain, swelling, possible use of crutches, and overall
disability associated with an acute knee injury, patients with such an injury
would have a lower mean Tegner activity level than would patients with a
chronic knee injury. For this study, a chronic knee injury was defined as more
than four weeks between the injury and the surgery. The independent-samples t
test was used to determine significance.Patients with a degenerative and/or complex meniscal tear would have a
lower mean Tegner activity level than would patients with a simple meniscal
tear. The independent-samples t test was used to determine significance.
Patients with moderate-to-severe pain with activity would have a lower mean
Tegner activity level than would patients with no or mild pain with activity.
The independent-samples t test was used to determine significance.
Patients with moderate-to-severe difficulty running would have a lower mean
Tegner activity level than would patients with no or mild difficulty running.
The independent-samples t test was used to determine significance.
Patients with more difficulty with activities of daily living would have
lower Tegner activity levels than would patients with less difficulty with
activities of daily living. The Spearman rho test was used to determine
significance.
Patients with more difficulty working because of problems with the knee
would have lower Tegner activity levels than would patients with less
difficulty working because of problems with the knee. The Spearman rho test
was used to determine significance.
Patients with more difficulty participating in sports activity because of
problems with the knee would have lower Tegner activity levels than would
patients with less difficulty participating in sports activity because of
problems with the knee. The Spearman rho test was used to determine
significance.
Patients who assessed their overall knee function as abnormal or severely
abnormal would have a lower mean Tegner activity level than would patients who
assessed it as normal or nearly normal. The independent-samples t test was
used to determine significance.
Because of the pain, swelling, possible use of crutches, and overall
disability associated with an acute knee injury, patients with such an injury
would have a lower mean Tegner activity level than would patients with a
chronic knee injury. For this study, a chronic knee injury was defined as more
than four weeks between the injury and the surgery. The independent-samples t
test was used to determine significance.
Patients with a degenerative and/or complex meniscal tear would have a
lower mean Tegner activity level than would patients with a simple meniscal
tear. The independent-samples t test was used to determine significance.
Responsiveness
Responsiveness to change was assessed in the group with an isolated
meniscal lesion (191 patients for whom the Lysholm score was determined and
eighty patients for whom the Tegner activity level was determined) and the
group with a meniscal lesion combined with other intra-articular lesions (477
patients). The preoperative Lysholm and Tegner activity scale scores were
compared with the scores at twelve months after treatment of the meniscal
lesion. The effect size was calculated by subtracting the mean preoperative
score from the mean postoperative score and dividing the result by the
standard deviation of the preoperative score. The standardized response mean
was calculated by subtracting the mean preoperative score from the mean
postoperative score and dividing the result by the standard deviation of the
change in the score. Values between 0.20 and 0.50 were considered to be small
effects; those between 0.51 and 0.80, moderate effects; and those of >0.80,
large
effects13.
Test-Retest Reliability
Lysholm score: The overall Lysholm score had acceptable
test-retest reliability (intraclass correlation coefficient, >0.70)
(Table II). The standard error
of measurement was 3.6. The minimum detectable change for the knees with an
isolated meniscal lesion was 10.1. The domains of the Lysholm score for
instability, pain, stair-climbing, limping, support, swelling, and squatting
also had acceptable test-retest reliability (intraclass correlation
coefficient, >0.70). The domain for locking had less than acceptable
test-retest reliability (intraclass correlation coefficient, 0.67).
Tegner activity scale: The Tegner activity scale had acceptable
test-retest reliability (intraclass correlation coefficient, 0.817; 95%
confidence interval, 0.75 to 0.87). The standard error of measurement was 0.4.
The minimum detectable change was 1 for knees with an isolated meniscal
lesion.
Internal Consistency
The Lysholm score had acceptable internal consistency (Cronbach alpha,
0.729).
Content Validity
Lysholm score: The mean preoperative overall Lysholm score (and
standard deviation) was 58.4 ± 19.4 points (range, 2 to 100 points) in
the group with an isolated meniscal lesion and 61.7 ± 20.4 points
(range, 6 to 100 points) in the group with combined lesions (see Appendix).
The overall Lysholm score had acceptable floor and ceiling effects (<30%)
(Table III). In the group with
an isolated meniscal lesion, the domains of pain, swelling, limp, instability,
support, stair-climbing, and locking had an acceptable floor effect (<30%),
and the domains of pain, swelling, squatting, and stair-climbing had an
acceptable ceiling effect (<30%). The domain of squatting had a high floor
effect (>30%), and the domains of limp, instability, support, and locking
had a high ceiling effect (>30%).
Tegner activity scale: The mean Tegner activity level (and
standard deviation) in the group with an isolated meniscal lesion was 4
± 2.3 (range, 0 to 10) (see Appendix). The mean Tegner activity level
in the group with combined lesions was 3 ± 1.7 (range, 0 to 10) (see
Appendix). The Tegner activity scale had acceptable floor and ceiling effects
(<30%) (Table III).
Criterion Validity
There were significant (p < 0.05) correlations between the overall
Lysholm score and the physical component score of the SF-12 scale (r = 0.551).
There were also significant (p < 0.05) correlations between the Tegner
activity scale and the physical component score of the SF-12 scale (r =
0.46).
Construct Validity
Lysholm score: All eight hypotheses (constructs) were significant
(p < 0.05). Patients with lower activity levels had lower Lysholm scores
(eta2 = 0.15, p < 0.001). Patients with an acute injury had a
lower mean Lysholm score (51.6 ± 20.1 points) than did patients with a
chronic knee injury (60.3 ± 19.2 points) (p = 0.012). Patients with a
Workers' Compensation claim had a lower mean Lysholm score (47.1 ± 18.9
points) than did patients without a Workers' Compensation claim (59.4 ±
22.5 points) (p = 0.015). Patients with more difficulty with activities of
daily living had lower Lysholm scores than did patients with less difficulty
with activities of daily living (r = 0.536, p < 0.001). Patients with more
difficulty working because of problems with the knee had lower Lysholm scores
than did patients with less difficulty working because of problems with the
knee (r = 0.555, p < 0.001). Patients with more difficulty with sports
because of problems with the knee had lower Lysholm scores than did patients
with less difficulty with sports because of problems with the knee (r = 0.407,
p < 0.001). Patients who assessed their overall knee function as abnormal
or severely abnormal had a lower mean Lysholm score (49.9 ± 17.7
points) than did patients who assessed it as normal or nearly normal (72.0
± 14.5 points) (p < 0.001). Patients with a degenerative and/or
complex meniscal tear had a lower mean Lysholm score (55.4 ± 18.8
points) than did patients with a simple meniscal tear (63.9 ± 19.4
points) (p = 0.004).
Tegner activity scale: All eight constructs were significant.
Patients with moderate-to-severe pain with activity had a lower mean Tegner
activity level (3.5 ± 2.1) than did patients with no or mild pain with
activity (5.0 ± 2.4) (p = 0.004). Patients with moderate-to-severe
difficulty running had a lower mean Tegner activity level (3.6 ± 2.1)
than did patients with no or mild difficulty running (5.5 ± 2.4) (p =
0.004). Patients with more difficulty with activities of daily living had
lower Tegner activity levels than did patients with less difficulty with
activities of daily living (r = 0.583, p < 0.001). Patients with more
difficulty working because of problems with the knee had lower Tegner activity
levels than did patients with less difficulty working because of problems with
the knee (r = 0.632, p < 0.001). Patients with more difficulty with sports
because of problems with the knee had lower Tegner activity levels than did
patients with less difficulty with sports because of problems with the knee (r
= 0.593, p < 0.001). Patients who assessed their overall knee function as
abnormal or severely abnormal had a lower mean Tegner activity level (3.2
± 2.1) than did patients who assessed it as normal or nearly normal
(5.2 ± 2.1) (p < 0.001). Patients with an acute injury had a lower
mean Tegner activity level (3.1 ± 2.0) than did patients with a chronic
knee injury (4.3 ± 2.3) (p = 0.017). Patients with a degenerative
and/or complex meniscal tear had a lower mean Tegner activity level (3.09
± 2.2) than did patients with a simple meniscal tear (4.2 ± 2.2)
(p = 0.046).
Responsiveness
Lysholm score: The Lysholm score had a large overall effect size
in the group with an isolated lesion (1.2) and in the group with combined
lesions (1.2). It also had a large overall standardized response mean in the
two groups (0.97 and 1.13, respectively). The individual domains of the
Lysholm score had large, moderate, and small effect sizes in the group with an
isolated meniscal lesion (Table
IV).
Tegner activity scale: The Tegner activity scale had a moderate
effect size in the group with an isolated lesion (0.61) and a large effect
size in the group with combined lesions (0.836). It had a moderate
standardized response mean in the two groups (0.60 and 0.704,
respectively).
In this study, the Lysholm knee score and Tegner activity scale
demonstrated, in general, acceptable psychometric parameters (test-retest
reliability, internal consistency, floor and ceiling effects, criterion
validity, construct validity, and responsiveness) to justify their use as
outcome measures for patients with a meniscal lesion of the knee.
Content validity represents a score's ability to cover all dimensions of
the variable. Often, expert input from the research community determines if
items leave out important aspects of each domain. Other investigators have
used ceiling and floor effects as measures of content validity to show that an
instrument uses the full range of
scores9,15,16.
In the Lysholm score, the domain of squatting had a high floor effect and the
domains of limp, instability, support, and locking had a high ceiling effect.
Thus, these domains may lack the discriminative ability to differentiate the
functional status of patients with a meniscal injury of the knee.
Although the Lysholm score is generally not interpreted on the basis of its
individual items, it is important to determine the properties of each of its
domains. It is important to know that the domain of locking, which can be a
common symptom of acute meniscal injuries, had less than acceptable
test-retest reliability and thus may not be sufficiently reliable for
scientific precision. The individual domain of locking therefore should not be
used alone, and the question in this domain may require further refinement to
improve its reliability for documenting locking associated with meniscal
injuries.
Overall, the Lysholm score had adequate test-retest reliability. To
estimate the repeatability of the test on the basis of this reliability, the
standard error of the measurement was calculated. This measurement together
with the appropriate z score for the desired level of confidence can be used
to determine an interval within which the actual score may
occur14. This has
been expanded to the concept of minimum detectable change, which provides a
number that determines when an improvement is large enough to be considered
beyond measurement error. The minimum detectable change calculated in our
study indicated that a documented change in the Lysholm score of <10 points
following treatment of a meniscal lesion may be attributable to measurement
error.
The analysis of responsiveness showed less than a small effect size for the
domain of instability in knees with an isolated meniscal lesion. This was due
to the fact that this symptom was not experienced by patients with an isolated
meniscal lesion. As a result of the high preoperative score for instability,
changes in instability were not expected to affect the response to treatment
of meniscal injuries. This finding may have been noted early in the
development of the Lysholm knee score. In 1982, Lysholm and
Gillquist2 presented
a modified version of the score for assessment of meniscal lesions. The only
difference between that score and the score commonly used today is that, in
the 1982 version, a maximum of 20 points could be assigned for instability
compared with a maximum of 25 points in the current version. A total score of
95 points was possible with the 1982 version compared with a total of 100
points in the current version.
The general psychometric performance of the Lysholm knee score has been
previously examined in studies of patients with various diagnoses. Irrgang et
al.17 reported
Cronbach alpha values of 0.60 to 0.73, which are similar to the value of 0.729
observed in our study. Their analysis of criterion validity demonstrated
correlations between the Lysholm score and a global rating of function of 0.54
to 0.57, which again were similar to the findings in the present study, in
which the values were 0.42 to 0.55. Their analysis of the responsiveness of
the Lysholm score showed an effect size of 0.82 to
1.1317, whereas we
found a value of 1.2 regardless of whether the patient had an isolated or
combined meniscal injury.
Marx et al.18
found an intraclass correlation coefficient of 0.95 for the Lysholm score,
which was similar to the intraclass correlation coefficient of 0.93 that we
observed. Their assessment of the criterion validity of the Lysholm score
revealed significant correlations with the physical function, role physical,
and bodily pain subscales of the SF-36 (r = 0.66, 0.49 and 0.57, respectively;
p < 0.01) findings that were similar to the significant correlations that
we observed with these subscales of the SF-12. Marx et al. found no ceiling or
floor effects of the Lysholm score in their patient population. Their analysis
of the responsiveness of the Lysholm score demonstrated a standardized
response mean of 0.9, which was similar to the value of 0.97 observed in our
study.
Marx et al.18
and Irrgang et
al.17 evaluated
patient populations with a variety of knee disorders, and the performance of
the scoring system was similar to that in our study. We previously examined
the properties of the Lysholm score in a study of patients with chondral
disorders, and again the scoring system performed similarly, with acceptable
test-retest reliability, floor and ceiling effects, construct and criterion
validity, and responsiveness to
change8. Although
this may indicate that the Lysholm score does not need to be tested for
patients with various knee disorders, we found that individual domains
performed differently in our study of patients with chondral defects than they
did in our current study of patients with meniscal lesions. This difference
may lead to differences in the overall performance of the score between
pathological entities. In the present study, we also found differences in the
score's performance between knees with an isolated meniscal lesion and those
with concurrent anterior cruciate ligament injury. These differences may have
been important when we measured the standard error of the measurement to
determine the minimal detectable change. If the minimal detectable change
differs between pathological entities, then analysis of patient change in
score should take this into account.
Paxton et al.9
evaluated the Lysholm score in a study of patients with an acute patellar
dislocation. They found no floor or ceiling effect for the overall score, and
they demonstrated correlations with the SF-36 that were similar to those that
we observed with the SF-12. Paxton et al. also found test-retest reliability
and internal consistency that were similar to those in our study.
Although our results were similar to those in the above studies, we believe
that we are one of the first to establish the validity of the Lysholm score
for patients with meniscal injuries.
We found that the one-item Tegner activity scale performed well. However,
with regard to responsiveness, the scale had only a moderate effect size and
standardized response mean when used for patients with an isolated meniscal
injury. This finding indicates that the Tegner activity scale has the ability
to measure moderate changes in activity level, which, given how the scale is
set up (i.e., large changes are not expected), may be adequate.
The Tegner activity scale, which was originally described in
19856, has been the
most widely used activity scoring system for patients with various knee
disorders9,19,20.
The scale is easy to use, but not all sports are represented in the
categories. Also, the original score was based on common sports in Europe, so
perceptions of sports activity levels may differ from those in the United
States. A few studies have documented the psychometric properties of the
Tegner activity
scale9,10.
The scale was shown to correlate with the activity rating scale described by
Marx et al.20 and
with an activity score developed for patients with ankle
instability13. In a
study of patients with an acute patellar dislocation, Paxton et
al.9 reported that
the Tegner scale had acceptable floor and ceiling effects and acceptable
test-retest reliability, findings similar to those in the present study.
However, in their analysis of criterion validity, Paxton et al. reported a
correlation with the SF-36 (r = 0.22 to 0.37) that was lower than the
correlation that we observed with the SF-12 (r = 0.46).
With meniscal injuries being the most common reason for arthroscopic
surgery, there needs to be a consensus regarding a standardized approach for
measuring outcomes of treatment. The validation of knee-condition-specific
measures for patients with meniscal injuries is essential to allow rigorous
comparison of treatment methods. In this study, we validated a score that
measured function and a scale that measured activity level. Marx et
al.20 noted that
function must be evaluated with the patient's activity level, and they
considered activity level to be the most important prognostic factor in the
athletic population. The interaction between function and activity level is an
important area requiring further research. How decreased function affects
activity level or vice-versa is an important consideration when documenting
outcomes, and it may be affected by patient age, gender, and expectations.
Criterion-related validity can be assessed by correlating a measure with a
gold standard or an already well-established measure of the characteristic,
and a limitation of this study was the lack of such a gold standard. The SF-12
has undergone vigorous psychometric testing and has been previously used to
measure outcomes following arthroscopic knee
surgery11,21,22.
However, it is not a gold standard of measurement of outcome following
arthroscopic knee surgery, and that limits the results regarding
criterion-related validity in this study.
Use of the Lysholm score, the Tegner activity scale, a generic measure of
health-related quality of life, and a measure of patient satisfaction would
provide a comprehensive outcome assessment for patients with meniscal
injuries. In general, the Lysholm knee score and the Tegner activity scale
demonstrated acceptable psychometric parameters. Psychometric testing of other
condition-specific knee instruments for patients with meniscal injuries of the
knee would be useful to allow comparison of psychometric properties and to
help determine the most appropriate instruments to use for evaluating patients
with this condition.