Available online at www.sciencedirect.com
, Ricardo Pietrobon
, Chad Cook
Brigham and Women’s Hospital, Department of Orthopaedic Surgery, Boston, MA, USA
Behavioral Medicine Program, Department of Psychiatry and Behavioral Sciences,
Division of Orthopedic Surgery, Center for Excellence in Surgical Outcomes,
Duke University Medical Center, Durham, NC, USA
Division of Physical Therapy, Duke University Medical Center, Box 3907, Durham, NC 27710, USA
Received 14 September 2006; accepted 31 January 2007
Available online 6 August 2007
Background: Patient-oriented questionnaires are important measures of clinical outcomes in medical practice but require systematic testing of
reliability and validity. The International Knee Documentation Committee (IKDC) Subjective Knee Form is a patient-oriented questionnaire that
assesses symptoms and function in daily living activities. The purpose of this study was to validate the IKDC Subjective Knee Form in a large
patient population with various knee disorders.
Methods: One thousand five hundred and thirty-four knee patients seen at a sports medicine clinic at a large medical center completed the IKDC
Subjective Knee Form. Factor structure was determined by exploratory factor analysis with promax rotation, and internal consistencies of the
identified subscales were calculated with Cronbach’s alpha. Concurrent validity was assessed by correlating the IKDC Subjective Knee Form
dimensions to the summary scales of the SF-12. Finally, item characteristics were analysed using graded response item response theory (G-IRT).
Results: Exploratory factor analyses yielded a two-factor solution, and the dimensions were termed as follows: symptom and knee articulation
(SKA), and activity level (AL). Confirmatory factor analysis confirmed the selection of 15 items within the study. Both SKA and AL demon-
strated good internal consistency (0.87 for SKA; 0.88 for AL). Both SKA and AL demonstrated statistically significant correlations to the SF-12
total score, more substantively to the physical component summary scale than to the mental component summary scale of the SF-12.
Discussion: G-IRT analyses revealed that nearly all questionnaire items demonstrated clear response patterns, with higher levels of the latent trait
corresponding to more adaptive clinical endpoints concerning pain, symptoms, function, and sport activity.
Conclusions: The IKDC is a reliable and valid instrument worthy of consideration for use in a broad patient population.
� 2007 Elsevier Masson SAS. All rights reserved.
Keywords: IKDC Subjective Knee Form; IRT; SF-12; Validation
Conventional approaches to the measurement of outcomes
in knee function have typically involved objectively defined
parameters, such as radiological findings, strength, range of
retain their value, patient-oriented questionnaires increasingly
have become recognized as important complementary mea-
sures of clinical outcomes. To ensure their utility, two objec-
tives are essential. First, it must be feasible to compare
outcomes concerning symptoms, function, and activity levels
Reliability and validity of the Internati
Laurence D. Higgins
, Marcus K. Tay
Joint Bone Spine 74 (2007)
motion, and ligamentous laxity . While these strategies
* Corresponding author. Tel.: þ1 (919) 684 9575; fax: þ1 (919) 668 3024.
E-mail address: email@example.com (C. Cook).
1297-319X/$ - see front matter � 2007 Elsevier Masson SAS. All rights reserv
onal Knee Documentation Committee
tive Knee Form
, Daniel Park
, Neil Ghodadra
across diverse knee conditions and patient populations. Sec-
ond, to ensure utility, patient-oriented questionnaires must be
validated both systematically and extensively, including the
testing of reliability, validity, and responsiveness.
Many different scoring systems for patient-focused instru-
ments have been developed to assess knee disability and func-
tion before and after treatment [2e8]. One such instrument is
the International Knee Documentation Committee (IKDC) Sub-
jective Knee Form, an 18-item, region specific instrument de-
signed to measure symptoms, function, and sports activity
[9,10]. The IKDC Subjective Knee Form was created from
a committee of international knee experts from the American
Orthopedic Society for Sports Medicine (AOSSM) and the
European Society for Sports Traumatology, Knee Surgery and
Arthroscopy (ESSKA). These groups formed the International
Knee Documentation Committee (IKDC) with the goal of stan-
dardizing an international documentation system for assessing
patient outcomes after knee surgery or treatment .
Although recently introduced, the IKDC Subjective Knee
Form is gaining recognition in the literature [11e17]. The
instrument contains 18 selected items designed to measure
symptoms assess pain, stiffness, swelling, joint locking, and
joint instability, while other items designed to measure knee
function assess the ability to perform activities of daily living.
Items purported to measure activity levels assess the respon-
dent’s ability to run, jump and land, stop and start quickly, as-
cend and descend stairs, stand, kneel on the front of the knee,
squat, sit with the knee bent, and rise from a chair. Examples
of questions include, ‘‘what is the highest level of activity that
you can perform without significant knee pain?’’, ‘‘what is the
highest level of activity that you can participate in on a regular
basis?’’, and ‘‘how does your knee affect your ability to go up
stairs?’’ Response types include 5-point Likert scales, 11-point
Likert scales, and dichotomous ‘‘yeseno’’ responses. A copy of
the IKDC Subjective Knee Form is available elsewhere .
The IKDC Subjective Knee Form has been validated in
Dutch  and Italian  as well as English . Previous
research  has suggested a one-factor solution using poly-
choric correlations, with an internal consistency coefficient
(Cronbach’s alpha) of 0.92 and testeretest reliability of
0.95. Concurrent validity for the IKDC Subjective Knee
Form was demonstrated, as it related significantly to the
SF-36  physical function subscales (r¼ 0.44e0.66), but
not to emotional function subscales (r¼ 0.16e0.26). Partial
credit item response theory (PC-IRT) analysis suggested that
patients with lower levels of function appropriately scored se-
lections of the lowest ordinal response categories . Finally,
analyses of differential item functioning, advocates that ques-
tions functioned similarly for men versus women, young ver-
sus old, and for those with different diagnoses .
The purpose of our study was to explore further the psycho-
metrics of the IKDC Subjective Knee Form using graded
response model IRT (G-IRT). G-IRT differs from the PC-
IRT used initially by Irrgang and colleagues in that the model
is specifically used when items have different numbers of
response categories (as with the IKDC Subjective Knee
Form) and when discrete items are examined independent of
the sum of the exponentials of the instrument [19,20]. Findings
may broaden support of the use of the IKDC Subjective Knee
Form and allow further examination of each individual item
L. D. Higgins et al. / Joint Bo
within the form for use in an item bank.
The study was approved by the Duke University Institu-
tional Review and Ethics Committee. Subjects were eligible
for the study if they were currently treated for any form of
knee dysfunction in an orthopedic sports medicine clinic at
a medical center in the southeast United States. Before the ini-
tial orthopedic examination, patients were provided with
a questionnaire packet including background/demographic in-
formation, the IKDC Subjective Knee Form, and the medical
outcomes study short form-12 (SF-12) , which allowed
patients time to answer all questions prior to the surgeon’s
examination. In addition, background information regarding
comorbidities was collected to determine other potential sour-
ces of contribution to dysfunction. No surgeons were involved
in any of the data collection or transfer procedures. Data were
collected by a nurse clinician who was also responsible in
transferring the data forms to a clinic database technician,
located in the clinic. The SF-12 is a brief version of the widely
used SF-36 [22,23,18]. It has been argued that the SF-36 is too
long for inclusion in certain research and clinical circum-
stances . The SF-12 includes 12 items drawn from the
SF-36, and identifies two summary scales, termed the physical
component summary scale (PCS) and the mental component
summary scale (MCS). The SF-12 subscales have demon-
strated multiple R-squares of 0.91 and 0.92 in predictions of
the SF-36 PCS and MCS, respectively . This instrument
has also consistently differentiated between groups known to
differ relative to presence and seriousness of physical and
mental conditions, acute symptoms, age and aging, and
depression. Finally, adequate testeretest reliabilities for the
SF-12 PCS and MCS scales have been demonstrated .
2.2. Statistical analyses
All statistical analyses were performed using Stata version
8.0 for Linux (Stata Corporation, College Station, TX, USA).
Initially, descriptive analyses using means and percentages
with 95% confidence intervals were used to establish the
demographic and clinical characteristics of the sample. Factor
structure was analysed using exploratory factor analysis with
promax rotation which was verified using confirmatory factor
analysis. Cronbach’s alpha reliability coefficients were com-
puted for each subscale. Instrument validity was determined
by correlating the IKDC Subjective Knee Form dimensions
to the SF-12 physical and mental component summary scores.
Item characteristics were analysed using G-IRT, with probabil-
ity of item responses being regressed on the latent trait (knee
function). Item patterns were obtained using Stata 8.0, while
all G-IRT models were computed using MULTILOG .
For descriptive analyses and factor analyses, observations
with missing values for any two or more items were excluded.
For IRT analyses, observations with any missing values were
595ne Spine 74 (2007) 594e599
AL¼ activity level, PCS¼ physical component summary scale, and
3.1. Baseline characteristics
This study included 1517 patients seen for conditions such
as osteoarthritis, ligamentous tears, general knee pain, and pre-
and post-surgery status. The participants were mostly male
(n¼ 893, 58.2%) and Caucasian (n¼ 1119, 73.0%). Average
age was 37.5 years (SD¼ 14.9). The average total IKDC Sub-
jective Knee Form score was 50.2 (SD¼ 12.6) (Table 1). The
most prevalent comorbid pathological condition among this
patient population was osteoarthritis (n¼ 135, 8.8%), fol-
lowed by hypertension (n¼ 133, 8.7%), depression (n¼ 97,
6.32%), and diabetes (n¼ 58, 3.8%). Other comorbidities
were present in small percentages, including heart disease,
anemia and blood disease, rheumatoid arthritis, and back
pain (Table 2).
3.2. Factor analysis and dimension reliabilities
Exploratory factor analysis with promax rotation of the 18
items was performed with one-, two-, and three-factor solu-
tions, and the two-factor solution was most interpretable. A
follow-up confirmatory factor analysis (traditional method)
verified the factors loaded on the pre-established latent
Participants’ demographics and IKDC Subjective Knee Form score
Variable N Mean/frequency SD/percentage Min Max
Age 1517 37.48 14.89 6 100
Male 893 893 58.21%
Female 624 624 40.68%
African 255 255 16.62%
Asian 32 32 2.09%
Caucasian 1119 1119 72.95%
Hispanic 32 32 2.09%
Native 16 16 1.04%
Other 15 15 0.98%
IKDC score 985 50.16 12.58 3.45 82.76
Comorbid medical conditions
Variable Frequency Percentage
Heart disease 38 2.48
Hypertension 133 8.67
Lung disease 7 0.46
Diabetes 58 3.78
Kidney disease 12 0.78
Liver disease 9 0.59
Anemia/blood disease 46 3.00
Cancer 22 1.43
Depression 97 6.32
Osteoarthritis 133 8.80
Back pain 24 1.56
Rheumatoid arthritis 50 3.26
596 L. D. Higgins et al. / Joint B
Other 191 12.45
variables. The findings suggest that the IKDC Subjective
Knee Form can be identified as having two distinct dimen-
sions. The two dimensions were termed as follows: symptom
and knee articulation (SKA; 11 items) and activity level
(AL; 4 items) (Table 3). Three items were factorially complex
and failed to load in either factor, therefore were not consid-
ered in descriptive analyses or reliability analyses associated
with either subscale. Cronbach’s alpha reliabilities of the iden-
tified subscales were 0.87 (SKA) and 0.88 (AL). After the
two-factor solution was verified, further validation procedures
were conducted, involving a comparison of the identified sub-
scales to the summary scales of the SF-12 , and analysis of
item characteristics using item G-IRT.
3.3. Construct and convergent comparison of IKDC
Subjective Knee Form to SF-12
IKDC Subjective Knee Form SKA dimension scores pre-
sented significant correlations, in the expected directions,
with the total SF-12 (r¼ 0.45, p< 0.0001), as well as the
physical component summary scale (r¼ 0.51, p< 0.0001)
and the mental component summary scale (r¼ 0.12,
p< 0.0001) (Table 4). The AL dimension scores demonstrated
significant links to the total SF-12 score (r¼ 0.40, p< 0.0001)
and the physical component summary scale (r¼ 0.48,
p< 0.0001). The relationship between the AL subscale and
Factor loadings for two-factor solution of the IKDC Subjective Knee Form
Item Description Factor 1 Factor 2
1 Highest activity level without pain 0.7
2 Pain frequency 0.5
3 Pain intensity 0.6
4 Stiff or swelling 0.5
5 Highest activity without swelling 0.8
6 Locking or catching 0.3 0.04
7 Highest activity without instability 0.8
8 Highest level of activity 0.8
9 Go up stairs 0.8
10 Go down stairs 0.7
11 Kneel on front of the knee 0.7
12 Squat 0.7
13 Sit with your knee bent 0.6
14 Rise from a chair 0.7
15 Run straight ahead 0.4 0.4
16 Jump and land on the involved leg 0.4 0.4
17 Stop and start quickly 0.5
18 Current function of the knee 0.5
Three items (#6, #15, and #16) did not clearly load in this two-factor solution.
Comparison of IKDC Subjective Knee Form to SF-12
SF-12 p-Value PCS p-Value MCS p-Value
SKA 0.65 <0.0001 0.71 <0.0001 0.20 <0.0001
AL 0.40 <0.0001 0.48 <0.0001 0.08 0.013
Log value of activity level was used; SKA¼ symptom and knee articulation,
one Spine 74 (2007) 594e599
MCS¼mental component summary scale.
sion, items 4 (how does your knee affect your ability to go up
stairs?) and 5 (how does your knee affect your ability to go
down stairs?) demonstrated the highest level of discrimination,
suggesting that these items have the greatest ability to distin-
guish between trait levels of knee function above and below
their respective thresholds (as demonstrated by their steep
slopes at threshold values). Items 6 (how does your knee affect
your ability to kneel?), 7 (how does your knee affect your abil-
ity to squat?), 9 (how does your knee affect your ability to rise
from a chair?) and 10 (how does your knee affect your ability
to start and stop quickly?), possessed moderate levels of dis-
crimination, the remaining exhibiting adequate levels. For
the AL dimension, items 1 (what is the highest level of activity
that you can perform without significant knee pain?), 3 (what
is the highest level of activity that you can perform without
significant giving way in your knee?) and 4 (what is the high-
est level of activity that you can participate in on a regular ba-
sis?) demonstrated particularly high levels of discrimination,
while item 2 (what is the highest level of activity that you
can perform without significant swelling in your knee?) scored
very poorly and was not discriminative.
Furthermore, as evidenced by the threshold values associ-
onstrated across each response alternative, along with the rel-
atively high levels of discrimination imposed by most items,
designates that this instrument provides substantial informa-
tion across a broad range of trait levels of knee function.
The purpose of this study was to validate the IKDC Subjec-
tive Knee Form using G-IRT. G-IRT differs from other forms
of item response analyses, as it is appropriate for Likert type
scales or ordinal responses that are dissimilar in nature and
in number . Furthermore, G-IRT examines each item spe-
cific to its singular association with the latent construct of the
instrument or form . G-IRT analyses revealed that with the
exception of one questionnaire item (what is the highest level
of activity that you can perform without significant swelling in
your knee?), all demonstrated clear response patterns, with
higher levels of the latent trait (knee function) corresponding
to more adaptive clinical endpoints concerning pain, symp-
toms, function, and sport activity.
Our factor analyses findings may be the most prevalent var-
iation compared to the initial validation study of the IKDC
the mental component summary scale, however, did not reach
statistical significance (r¼ 0.07, p¼ 0.013). As expected,
then, the two IKDC Subjective Knee Form dimensions demon-
strated a more convincing relationship to the physical, as com-
pared to the mental elements of the SF-12.
3.4. Item characteristics of the IKDC Subjective Knee
Item characteristics analysed using G-IRT, that assessed the
probability of each item response regressed on the latent trait
(knee function) are presented in Table 5. For the SKA dimen-
IRT analysis for symptom and knee articulation subscale items
Symptom and knee articulation subscale
1. During the past 4 weeks, or since your injury, how often have you had pain?
2. If you have pain, how severe is it?
3. During the past 4 weeks, or since your injury, how stiff or swollen was your
4. How does your knee affect your ability to go up stairs?
5. How does your knee affect your ability to go down stairs?
6. How does your knee affect your ability to kneel on the front of your knee?
7. How does your knee affect your ability to squat?
8. How does your knee affect your ability to sit with your knee bent?
9. How does your knee affect your ability to rise from a chair?
10. How does your knee affect your ability to stop and start quickly?
11. Current function of the knee
Activity level subscale
1. What is the highest level of activity that you can perform without significant
2. What is the highest level of activity that you can perform without significant
3. What is the highest level of activity that you can perform without significant
4. What is the highest level of activity that you can participate in on a regular b
Higher slope values suggest better item discrimination per latent variable. Thres
values which indicate selections that are considered ‘‘easy’’ versus positive value
they demonstrate a wide variety of selections (negative to positive).
L. D. Higgins et al. / Joint Bo
ated with each response, all questionnaire items in both
subscales demonstrated clear response patterns, with higher
levels of the latent trait (knee function) corresponding to
more adaptive clinical endpoints concerning pain, symptoms,
function, and sport activity. Items 8 (how does your knee
affect your ability to sit with your knee bent?) and 9 (how
does your knee affect your ability to rise from a chair?)
exhibited lower threshold values in the Likert selections asso-
ciated with unable to do, extremely difficult, moderately diffi-
cult, and minimally difficult, suggesting that these item
response were considered less descriptive of their condition
at hand. Nonetheless, the broad continuum of thresholds dem-
1.37 0.10 0.78 1.60 2.62
1.05 �1.68 �0.34 0.89 2.61
knee? 1.25 �2.42 �0.88 0.60 2.17
4.05 �1.90 �0.79 0.21 1.04
3.39 �2.04 �0.80 0.20 1.02
2.0 �0.87 �0.03 0.79 1.68
2.21 �0.87 0.01 0.92 1.92
1.44 �2.14 �1.39 �0.35 0.64
2.61 �2.50 �1.35 �0.31 0.64
2.09 �0.97 �0.21 0.65 1.47
1.86 �1.07 �0.17 0.72 1.69
knee pain? 3.42 �1.49 0 1.72 1.98
swelling in your knee? 0 �1.33 0 1.38 1.66
giving way in your knee? 4.41 �1.49 0 1.29 1.46
asis? 2.86 0 0 1.55 1.75
hold (ß) reflects variations between ordinal selections within the item. Negative
s which are indicative of more difficult selections. Items are more descriptive if
597ne Spine 74 (2007) 594e599
Subjective Knee Form . Although the initial study 
suggested a single dominant component, exploratory and con-
firmatory factor analytic procedures in the present study sug-
gested a two-factor solution. Identification of factor structure
can be a complicated issue, and at least two possible
explanations exist for the discrepant results between these
two studies. First, Irrgang et al.  may have implemented
different criteria in the identification of factor structure. In
their study, the largest factor had an eigenvalue of 9.03, which
accounted for 50.2% of the total variance, and the second and
third factors possessed eigenvalues of 1.76 and 1.19, respec-
tively. Although extracted factors with eigenvalues greater
than 1 are typically retained, the authors may have determined
that the initial factor was quite substantial and accounted for
a significant portion of total variance, such that the second
and third factors should not be considered because these
contained ambiguous factor loadings. In the present study,
the two-factor solution was chosen because the eigenvalues
for the identified factors were 7.76 and 1.11, respectively,
and the two-factor solution resulted in the cleanest (least am-
biguous) separation between factors. Proper selection of facto-
rial structure is essential for correct performance of item
response analysis . Three items were factorially complex
and did not load clearly on either factor, a finding that was dis-
similar to that of Irrgang and associates . We elected not to
include these items within the G-IRT and suggest the exclusion
of these items in the IKDC Subjective Knee Form. Commonly,
items that fail to distill into a dimension also do not demon-
strate strong association with the latent construct of interest.
Our findings support that the items (1) locking or catching,
(2) run straight ahead, and (3) jump and land on the involved
leg do not stand alone with the IKDC Subjective Knee Form
and should not be used in an item bank.
The internal consistency coefficients for the IKDC Subjec-
tive Knee Form subscales in the present study were notable
(0.87 and 0.88 for SKA and AL indices, respectively). Typi-
cally, internal consistency coefficients above 0.70 are consid-
ered acceptable. Since this is the first study to suggest
a two-factor structure, no previous research has calculated
internal consistency coefficients for these subscales. More
research is needed to confirm the factor structure identified
in the present study, and to investigate further the internal
consistencies of these scales.
As hypothesized, both IKDC Subjective Knee Form
subscales demonstrated statistically significant correlations to
the SF-12 total score, yet both related more substantively to
the physical component summary than to the mental compo-
nent summary. Both dimensions of the IKDC Subjective
Knee Form correlated, in the expected directions, with the
total SF-12 as well as the physical component summary scale.
Both subscales, however, related less convincingly to the men-
tal component summary scale of the SF-12. In fact, the corre-
lation between the activity level dimension (subscale) and the
mental component summary scale did not reach statistical
significance. These findings are in agreement with previous
research that has examined the relationships between these
598 L. D. Higgins et al. / Joint Bo
two scales . A validated Italian version of the IKDC Sub-
jective Knee Form , for instance, correlated significantly
with the PCS and the MCS of the SF-36, although more con-
vincingly for the PCS than for the MCS.
Analyses of item characteristics demonstrated that, for both
the SKA and AL dimensions, most items demonstrated high
levels of discrimination, meaning that these items effectively
distinguish between trait levels of knee function above and
below their respective thresholds. Furthermore, as evidenced
by the threshold values associated with each response, all
questionnaire items in both subscales demonstrated clear re-
sponse patterns, with higher levels of the latent trait (knee
function) corresponding to more adaptive clinical endpoints
concerning pain, symptoms, function, and sport activity. Addi-
tionally, a broad continuum of thresholds was demonstrated
across each response alternative for most items. Altogether,
these data are the first to provide support for the validity of
the IKDC Subjective Knee Form at the item level, and further
suggest that the instrument may be suitable for use in patient
populations with a broad range of levels of knee function or
We employed G-IRT analyses to examine the validity and
performance of the IKDC Subjective Knee Form at the item
level, which possesses several advantages over classical test
theory (CTT) approaches and is distinct from the PC-IRT
used in an earlier validation . In contrast to CTT, scale
properties obtained from G-IRT are population-independent,
whereas properties derived from CTT are limited to specific
populations. PC-IRT creates a common item slope that is
used to rescale the item parameters, while G-IRT analyses
each item individually [19,20]. In addition, G-IRT allows
transformation of scores from different scales measuring the
same trait (i.e., knee function) into a common and comparable
metric  and is able to validate single item for use in a sep-
arate item bank.
Limitations of our study should be recognized. In this
investigation, we neither examined testeretest reliability of
the instrument, nor analysed the responsiveness of the instru-
ment to changes in knee function across time. Although the in-
clusion criteria outlined the acceptance of a wide variety of
patients with knee-related disorders the data collectors did
not obtain individual diagnoses from each patient respondent.
Subsequently, the actual percentage of each diagnosis within
the study is unknown. Nonetheless, IRT is purported to be
diagnosis independent and should exhibit strong psychometric
properties within a variety of disorders of the knee. Addition-
ally, the majority of our patient population was Caucasian
(n¼ 1119, 73.0%), and the average age was relatively young
(37.5 years, SD¼ 14.9).
In summary, we have demonstrated that the IKDC Subjective
Knee Form possesses adequate scale reliability and validity,
with factor analysis resulting in a two-factor solution and inclu-
sion of 15 of the original 18 items. As hypothesized, both IKDC
Subjective Knee Form subscales demonstrated statistically sig-
ne Spine 74 (2007) 594e599
nificant correlations to the SF-12 total score, yet both related
more substantively to the physical component summary scale
than to the mental component summary scale. G-IRT analyses
revealed that with the exception of one items, all questionnaire
items demonstrated clear response patterns, with higher levels
of the latent trait (knee function) corresponding to more
adaptive clinical endpoints concerning pain, symptoms, func-
tion, and sport activity. Future validation studies should exam-
ine the testeretest reliability of the instrument, as well as its
responsiveness to changes in knee function across time.
 Marx RG. Knee rating scales. Arthroscopy 2003;19:1103e8.
 Barber-Westin SD, Noyes FR, McCloskey JW. Rigorous statistical reli-
ability, validity, and responsiveness testing of the Cincinnati knee rating
system in 350 subjects with uninjured, injured, or anterior cruciate
 Anderson AF. Transepiphyseal replacement of the anterior cruciate liga-
ment using quadruple hamstring grafts in skeletally immature patients. J
Bone Joint Surg Am 2004;86:201e9.
 Delcogliano A, Caporaso A, Menghi A, Rinonapoli G, Chiossi S. Results
of autologous osteochondral grafts in chondral lesions of the knee. Mi-
nerva Chir 2002;57:273e81.
 Haverkamp D, Sierevelt IN, Breugem SJM, Lohuis K, Blankevoort L,
van Dijk CN, et al. Translation and validation of the Dutch version of
the international knee documentation committee subjective knee form.
Am J Sports Med 2006;34:1680e4.
 Ma CB, Francis K, Towers J, Irrgang J, Fu FH, Harner CH, et al. Ham-
string anterior cruciate ligament reconstruction: a comparison of bioab-
sorbable interference screw and endobuttonepost fixation. Arthroscopy
 Padua R, Bondi R, Ceccarelli E, Bondi L, Romanini E, Zanoli G, et al.
Italian version of the international knee documentation committee sub-
jective knee form: cross-cultural adaptation and validation. Arthroscopy
599L. D. Higgins et al. / Joint Bone Spine 74 (2007) 594e599
ligament-reconstructed knees. Am J Sports Med 1999;27:402e16.
 Irrgang JJ, Snyder-Mackler L, Wainner RS, Fu FH, Harner CD. Develop-
ment of a patient-reported measure of function of the knee. J Bone Joint
Surg Am 1998;80:1132e45.
 Johnson DS, Smith RB. Outcome measurement in the ACL deficient
knee e what’s the score? Knee 2001;8:51e7.
 Marx RG, Jones EC, Allen AA, Altchek DW, O’Brien SJ, Rodeo SA,
et al. Reliability, validity, and responsiveness of four knee outcome scales
for athletic patients. J Bone Joint Surg Am 2001;83:1459e69.
 Paxton EW, Fithian DC, Stone ML, Silva P. The reliability and validity of
knee-specific and general health instruments in assessing acute patellar
dislocation outcomes. Am J Sports Med 2003;31:487e92.
 Roos EM, Roos HP, Lohmander LS, Ekdahl C, Beynnon BD. Knee injury
and osteoarthritis outcome score (KOOS) e development of a self-admin-
istered outcome measure. J Orthop Sports Phys Ther 1998;28:88e96.
 Ryd L, Karrholm J, Ahlvin P. Knee scoring systems in gonarthrosis.
Evaluation of interobserver variability and the envelope of bias. Acta Or-
thop Scand 1997;68:41e5. Score Assessment Group.
 Anderson AF, Irrgang JJ, Kocher MS, Mann BJ, Harrast JJ, International
Knee Documentation Committee. The international knee documentation
committee subjective knee evaluation form: normative data. Am J Sports
 Irrgang JJ, Anderson AF, Boland AL, Harner CD, Kurosaka M, Neyret P,
et al. Development and validation of the international knee documenta-
tion committee subjective knee form. Am J Sports Med 2001;29:600e13.
 Anderson AF. Transepiphyseal replacement of the anterior cruciate liga-
ment in skeletally immature patients. A preliminary report. J Bone Joint
Surg Am 2003;85:1255e63.
 Vitkus L, Fiodorovas M, Pocius G. Knee joint evaluation after anterior
cruciate ligament plasty. Medicina (Kaunas) 2002;38:510e4.
 Ware Jr JE, Sherbourne CD. The MOS 36-item short-form health survey
(SF-36). I. Conceptual framework and item selection. Med Care
 Embretson S, Reise SP. Item response theory for psychologists. Hillsdale,
NJ: Erlbaum; 2000.
 Masters GN. A Rasch model for partial credit scoring. Psychometrika
 Ware Jr JE, Kosinski M, Keller SD. A 12-item short-form health survey:
construction of scales and preliminary tests of reliability and validity.
Med Care 1996;34:220e33.
 McHorney CA, Ware Jr JE, Lu JF, Sherbourne CD. The MOS 36-item
short-form health survey (SF-36): III. Tests of data quality, scaling as-
sumptions, and reliability across diverse patient groups. Med Care
 McHorney CA, Ware Jr JE, Raczek AE. The MOS 36-item short-form
health survey (SF-36): II. Psychometric and clinical tests of validity in
measuring physical and mental health constructs. Med Care
 Thissen D. MULTILOG user’s guide: multiple categorical analysis and
test scoring using item response theory. Chicago: Scientific Software In-
 van der Linden W, Hambleton R. Handbook of modern item response
theory. New York: Springer; 1997.
 Bot SD, Terwee CB, van der Windt DA, Bouter LM, Dekker J, de
Vet HC. Clinimetric evaluation of shoulder disability questionnaires:
a systematic review of the literature. Ann Rheum Dis 2004;63:335e41.