Knee Surgery, Sports Traumatology, Arthroscopy

https://doi.org/10.1007/s00167-018-4879-5

KNEE

The WOMAC score can be reliably used to classify patient satisfaction

after total knee arthroplasty
Lucy C. Walker1 · Nick D. Clement1 · Michelle Bardgett1 · David Weir1 · Jim Holland1 · Craig Gerrand1 ·
David J. Deehan1

Received: 23 October 2017 / Accepted: 12 February 2018

© European Society of Sports Traumatology, Knee Surgery, Arthroscopy (ESSKA) 2018

Abstract
Purpose  The primary aim of this study was to define a classification in the WOMAC score after total knee arthroplasty (TKA)
according to patient satisfaction. The secondary aims were to describe patient demographics for each level of satisfaction.
Methods  A retrospective cohort consisting of 2589 patients undergoing a primary TKA were identified from an established
arthroplasty database. Patient demographics, Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC),
and short form (SF) 12 scores were collected pre-operatively and 1 year post-operatively. In addition, patient satisfaction
was assessed at 1 year with four responses: very satisfied, satisfied, dissatisfied or very dissatisfied. Receiver operating
characteristic (ROC) curves were used to identify values in the components and total WOMAC scores that were predictive
of each level of satisfaction, which were used to define the categories of excellent, good, fair and poor.
Results  At 1 year, there were 1740 (67.5%) very satisfied, 572 (22.2%) satisfied, 190 (7.4%) dissatisfied and 76 (2.9%)
very dissatisfied patients. ROC curve analysis identified excellent, good, fair and poor categories for the pain (> 78, 59–78,
44–58, < 44), function (> 72, 54–72, 41–53, < 41), stiffness (> 69, 56–69, 43–55, < 43) and total (> 75, 56–75, 43–55, < 43)
WOMAC scores, respectively. Patients with lung disease, diabetes, gastric ulcer, kidney disease, liver disease, depression,
back pain, with worse pre-operative functional scores (WOMAC and SF-12) and those with less of an improvement in the
scores, had a significantly lower level of satisfaction.
Conclusion  This study has defined a post-operative classification of excellent, good, fair and poor for the components and
total WOMAC scores after TKA. The predictors of level of satisfaction should be recognised in clinical practice and patients
at risk of a lower level of satisfaction should be made aware in the pre-operative consent process.
Level of evidence III.

Keywords  WOMAC · Outcome · Total knee arthroplasty · Classification · Satisfaction

* Lucy C. Walker Craig Gerrand

lcwalker86@gmail.com craig.gerrand@nuth.nhs.uk
Nick D. Clement David J. Deehan
nickclement@doctors.org deehan1@hotmail.co.uk
Michelle Bardgett 1
Department of Orthopaedics, Newcastle‑upon‑Tyne
michelle.bardgett@nuth.nhs.uk
University Hospitals Foundation Trust,
David Weir Freeman Hospital, Freeman Road, High Heaton,
david.weir@nuth.nhs.uk Newcastle upon Tyne NE7 7DD, UK
Jim Holland
jim.holland@nuth.nhs.uk

13
Vol.:(0123456789)
 Knee Surgery, Sports Traumatology, Arthroscopy

Introduction Materials and methods

Patient-reported outcome measures (PROMs) are rou- Patients for this study were identified retrospectively from
tinely collected for audit and research purposes after total a prospectively compiled arthroplasty database held at the
knee arthroplasty (TKA) [10]. The Western Ontario and study centre. During a 10-year period, 3641 patients under-
McMaster Universities Osteoarthritis Index (WOMAC) going primary TKA at the study centre were asked to com-
[4] is thought to be the primary measure of efficacy for plete a pre-operative patient questionnaire. Only patients
osteoarthritis trials [11], and is a self-administered health with primary osteoarthritis were included. Patients that
status measure that assesses the dimensions of pain, stiff- underwent simultaneous bilateral TKA during the study
ness and function either separately or as an overall index period were excluded, and for those patients that underwent
[29]. Despite collecting this data, it is not clear how this a second TKA, after the index procedure, only the outcome
should be interpreted or what a post-operative score of of the first knee was used for analysis. Patients who had a
X points means to a patient. The concept of the Patient deep infection, did not complete the outcome assessments, or
Acceptable Symptom State (PASS) has gained attention were revised at before 1-year follow-up were also excluded
over the last decade, which is determined using patient from analysis. There were 2578 TKA performed during the
satisfaction with current symptoms or performance then study period with complete pre- and post-operative data that
determining the score at which a particular percentage of met the inclusion criteria (see Fig. 1).
subjects reports meeting that benchmark [27]. Numerous Basic patient demographics, body mass index (BMI) and
authors have defined the PASS for the WOMAC score, but comorbidities were collected pre-operatively. Comorbidities
this merely reflects a single score beyond which a patient were recorded as a categorical yes and no for: heart disease,
is deemed to have had an acceptable outcome. Rather than hypertension, lung disease, diabetes, stomach ulcer, kidney
a specific PASS score to assess outcome, there is possibly disease, liver disease, anaemia, cancer, depression, neuro-
a sliding scale from excellent to poor outcomes that equate logical disease, and back pain. The WOMAC [4] score and
to a specific range of post-operative WOMAC scores. Short Form (SF-) 12 score [28] were assessed pre-opera-
The post-operative Oxford score [8, 19, 21] and Knee tively and 1 year post-operatively.
Society score (KSS) [2, 12] have been used to define the
patient outcome as excellent, good, fair and poor. This
classification is useful when interpreting the patient’s post-
operative outcome on a graded scale, and has been widely Gave consent
used. To the authors knowledge, no such classification has Underwent pre-operative
been described for the WOMAC score, which would be assessment
of clinical and research interest to describe the outcome n=3641
of cohort studies. Patient satisfaction is recognised as an
important measure when assessing the outcome of TKA
Excluded
[17]. The thresholds at which each level of satisfaction is
attained could help to interpret the WOMAC score and Lost to 1 year
aid communication with patients, e.g. X post-operative assessment or revised
score is associated with X level of satisfaction. This n=222 (6.1%)
could then be used in pre-operative counselling to man-
age the patients’ expectations of post-operative outcome.
Did not meet
In addition, the identified thresholds in the post-operative
WOMAC score for patient satisfaction could be used to inclusion criteria
grade outcome. The thresholds to achieve each level of n=898 (26.7%)
satisfaction (very satisfied, satisfied, dissatisfied or very
dissatisfied) could be used to subdivide the post-operative Did not complete
WOMAC score according to the patient’s perception of satisfaction question
their outcome. n=11 (0.3%)
The primary aim of this study was to define a classi-
fication in the post-operative WOMAC score after TKA
Analysed
according to patient satisfaction. The secondary aims
were to describe patient demographics for each level of n=2521
satisfaction.
Fig. 1  Flow chart demonstrating cohort patient exclusions

13
Knee Surgery, Sports Traumatology, Arthroscopy

The WOMAC score [4] used in this study was the Lik- Statistical analysis
ert version 3.1 standardised with English for a British
population, consisting of 24 self-administrated questions Statistical analysis was performed using Statistical Package
that were answered for each item on a five-point Likert for Social Sciences version 17.0 (SPSS Inc., Chicago, IL,
scale (none, mild, moderate, severe and extreme). It was USA). The data assessed demonstrated a normal distribution
reported as three separate subscales: pain, physical func- and parametric tests were used to assess continuous vari-
tion, and stiffness. The WOMAC pain subscale had five ables for significant differences between groups. A Student’s
questions scored 0–4 and was considered invalid if more t test, unpaired and paired, was used to compare linear vari-
than one item was missing; hence, it had a range of 0 ables between groups. Dichotomous variables were assessed
(no pain) to 20 (maximal pain). In the event of a missing using a Chi-square test. Receiver operating characteristic
item, the remaining four items were averaged and then (ROC) curve analysis was used to identify thresholds (cut
multiplied by five [7]. The WOMAC function subscale points) in the components and total WOMAC scores that
has 17 questions scored 0–4 and was considered invalid were predictive of each of the satisfaction groups. The area
if more than three items were missing. It had a range of 0 under the ROC curve ranges from 0.5, indicating a test with
(maximal function) to 68 (minimal function). In the event no accuracy, to 1.0 where the test is perfectly accurate by
of missing items, the remaining items were averaged and identifying all satisfied patients. The threshold is equivalent
then multiplied by 17. The WOMAC stiffness subscale to the point (WOMAC score) at which the sensitivity and
had two items scored 0–4 and was considered invalid specificity are maximal in predicting patient satisfaction
if either was missing; hence it had a range from 0 (no [15]. A p value of < 0.05 was defined as significant.
stiffness) to 8 (maximal stiffness). The final scores were
determined by adding the corresponding items for each
dimension, and standardising to a range of values from 0 Results
to 100. According to recent recommendations, we have
used the reverse option, from 0 (worst) to 100 (best) [26]. There were 1182 males and 1396 females with a mean age of
The Short Form (SF-) 12 is a generic assessment tool 68.9 [standard deviation (SD) 9.6]. At 1 year, there were 1740
used to measure a patient’s wellbeing, which is assessed (67.5%) very satisfied, 572 (22.2%) satisfied, 190 (7.4%) dis-
using a physical component summary (PCS) and a mental satisfied and 76 (2.9%) very dissatisfied patients. Patients
component summary (MCS) [28]. Both the SF-12 PCS with lung disease (p = 0.04), diabetes (p = 0.01), gastric ulcer
and MCS range from 0 (worst level of functioning) to 100 (p = 0.001), kidney disease (p = 0.04), liver disease (p = 0.01),
(best level of functioning). depression (p < 0.0001), back pain (p < 0.0001), with worse
Patient satisfaction was assessed by asking the ques- pre-operative functional scores (WOMAC and SF-12) and
tion “How satisfied are you with the results of your knee those with less of an improvement in the scores (p < 0.0001)
replacement surgery?” at 1 year following surgery. The had a lower level of satisfaction (Table 1).
response was recorded using a four-point Likert scale: Patients with a higher level of satisfaction demonstrated
very satisfied, somewhat satisfied, somewhat dissatisfied, a greater post-operative component and total WOMAC
and very dissatisfied. scores compared to those with a lower level of satisfaction
The level of patient satisfaction was used to classify (Table 2). The same pattern was observed for the change
the grades in the WOMAC scores of excellent, good, in the scores when compared to the patients pre-operative
fair and poor. Three threshold values were identified to score (Table 2). Interestingly, all the groups had a signifi-
separate the groups using very satisfied versus satisfied, cant improvement in the pain and functional components and
satisfied versus dissatisfied and dissatisfied versus very total WOMAC score, except for the stiffness component in
dissatisfied. The threshold value represents a score that the dissatisfied group (Table 2).
most likely predicts one group over another. The threshold values predictive of very satisfied patients
There was no additional patient contact, and as such, from satisfied patients varied from 69 for the stiffness compo-
this project was performed as a service evaluation with- nent to 78 for the pain component with sensitivities and spe-
out the need for formal ethical approval. The project was cificities between 70 and 80% (Table 3). These thresholds were
registered with the institutions audit department (New- reliable with an area under curve (AUC) of approximately 80%
castle Hospitals NHS Foundation Trust, Project Record for all except the stiffness component with an AUC of 76%
Number 2840) and was conducted in accordance with the (Fig. 2). The threshold values predictive of satisfied patients
Declaration of Helsinki and the guidelines for good clini- from dissatisfied patients varied from 54 for the functional
cal practice. component to 58 for the pain component with sensitivities and
specificities between 60 and 75% (Table 4). The threshold val-
ues predictive of dissatisfied patients from very dissatisfied

13
 Knee Surgery, Sports Traumatology, Arthroscopy

Table 1  Patient demographics and pre-operative functional scores according level of satisfaction after TKR
Demographic Descriptive All patients, Level of satisfaction p value*
n = 2578
Very Satisfied, n = 572 Dissatisfied, Very dis-
satisfied, n = 190 satisfied,
n = 1740 n = 76

Gender (n, % of Male 1182 (45.8) 795 (45.7) 266 (46.5) 95 (50.0) 26 (34.2) n.s.**
group) Female 1396 (54.2) 945 (54.3) 306 (53.5) 95 (50.0) 50 (65.8)
Mean age (years: mean, SD) 68.9 (9.6) 69.2 (9.3) 68.6 (10.0) 67.6 (10.7) 67.0 (10.2) n.s
BMI (kg/m2: mean, SD) 29.8 (6.8) 29.4 (4.9) 30.6 (11.0) 30.0 (5.1) 30.5 (6.1) n.s.**
Comorbidity (n, % Heart disease 433 (16.8) 279 (16.0) 105 (18.4) 35 (18.4) 14 (18.4) n.s.**
of group) Hypertension 1406 (54.5) 932 (53.6) 333 (58.2) 98 (51.6) 43 (56.6) n.s.**
Lung disease 398 (15.4) 252 (14.5) 89 (15.6) 40 (21.1) 17 (22.4) 0.04**
Cancer 125 (4.8) 82 (4.7) 29 (5.1) 13 (6.8) 1 (1.3) n.s.**
Neurological 153 (5.9) 97 (5.6) 38 (6.6) 11 (5.8) 7 (9.2) n.s.**
disease
Diabetes mellitus 364 (14.1) 226 (13.0) 83 (14.5) 34 (17.9) 21 (27.6) 0.01**
Gastric ulceration 325 (12.6) 188 (10.8) 91 (15.9) 33 (17.4) 13 (17.1) 0.001**
Kidney disease 79 (3.1) 47 (2.7) 19 (3.3) 12 (6.3) 1 (1.3) 0.04**
Liver disease 44 (1.7) 25 (1.4) 9 (1.6) 9 (4.7) 1 (1.3) 0.01**
Anaemia 245 (9.5) 161 (9.3) 54 (9.4) 21 (11.1) 9 (11.8) n.s.**
Depression 374 (14.5) 193 (11.1) 102 (17.8) 48 (25.3) 31 (40.8) < 0.0001**
Back pain 1299 (50.4) 755 (43.4) 359 (62.8) 132 (69.5) 53 (69.7) < 0.0001**
Functional measures (mean, SD)
 WOMAC Total 36.5 (16.4) 37.9 (16.6) 34.2 (15.7) 33.5 (14.9) 29.8 (15.2) < 0.0001
Pain 35.4 (17.7) 36.8 (17.8) 33.1 (17.0) 32.9 (16.5) 27.9 (18.0) < 0.0001
Function 30.0 (17.0) 38.1 (17.4) 34.4 (16.3) 33.3 (15.3) 30.0 (15.8) < 0.0001
Stiffness 37.2 (20.50) 38.0 (20.8) 35.5 (19.8) 36.5 (20.0) 33.2 (19.1) 0.02
 SF-12 PCS 27.6 (7.4) 27.9 (7.50) 27.3 (7.1) 26.6 (7.3) 24.7 (6.7) 0.001
MCS 47.0 (13.6) 48.5 (13.5) 44.9 (13.3) 43.5 (12.7) 39.8 (14.2) < 0.0001

*ANOVA unless stated otherwise

**Chi square

patients varied from 41 for the functional component to 44 the predictive values for each level of patient satisfac-
for the pain component with sensitivities and specificities tion. This can then be used to contribute to pre-operative
between 62 and 76% (Table 5). The thresholds predictive of patient counselling and help manage patient expectations
satisfied from dissatisfied and dissatisfied from very dissatis- of post-operative outcome. Although the components of
fied patients were not as reliable with an AUC of approxi- the WOMAC score assess different aspects of the patient’s
mately 70% (Figs. 3, 4). outcome, the defined boundary scores were similar for
Using the threshold values identified for the components each subgroup: excellent > 70, good 55–70, fair 40–54
and total WOMAC scores, boundaries were set to define excel- and poor < 40 approximately. Specific comorbidities and
lent, good, fair and poor outcome categories (Table 6). Inter- worse pre-operative functional scores were also demon-
estingly the categories for each of the components and total strated to be associated with a lower level of post-operative
score have similar values to define each of the groups: excel- satisfaction.
lent > 70, good 55–70, fair 40–54 and poor < 40 approximately. The major limitation of this study was the retrospec-
tive design, with lost to follow-up data, may have skewed
the defined threshold factors as patients lost to follow-up
Discussion may represent a certain cohort whose satisfaction levels
and their relationship to their WOMAC scores were not
The most important outcome of this study was the clas- accounted for. However, there was no significant (p > 0.1
sification of the post-operative WOMAC scores into four n.s.) differences in the pre-operative demographics and
outcome subgroups of excellent, good, fair and poor using functional measures between those lost to follow-up and

13
Knee Surgery, Sports Traumatology, Arthroscopy

Table 2  Post-operative outcome measures and the difference relative to pre-operative scores for the all patients and according to level of patient
satisfaction
Functional measure All patients Level of satisfaction p value*
Very satisfied Satisfied Dissatisfied Very dissatisfied

Total
 1 year 74.8 (20.1) 83.0 (15.0) 63.3 (16.6) 49.9 (15.7) 35.5 (17.5) < 0.001
 Change (95% CI) 38.2 (37.4–39.0) 45.1 (44.3–46.0) 29.1 (27.7–30.1) 16.4 (14.1–18.6) 5.6 (2.5–8.8) < 0.001
 p value** < 0.001 < 0.001 < 0.001 < 0.001 0.001
Pain
 1 year 80.2 (20.8) 88.3 (14.3) 69.4 (17.7) 54.8 (18.7) 38.2 (18.0) < 0.001
 Change (95% CI) 44.7 (43.8–45.6) 51.5 (50.6–52.5) 36.3 (42.7–38.0) 21.9 (18.8–25.0) 10.2 (6.5–14.0) < 0.001
 p value** < 0.001 < 0.001 < 0.001 < 0.001 < 0.001
Function
 1 year 73.7 (22.5) 81.8 (16.5) 62.0 (18.0) 48.5 (16.3) 34.5 (18.9) < 0.001
 Change (95% CI) 36.8 (35.9–37.6) 43.7 (42.8–44.6) 27.6 (26.1–29.1) 15.2 (12.9–17.5) < 0.001
 p value** < 0.001 < 0.001 < 0.001 < 0.001 < 0.001
Stiffness
 1 year 71.6 (22.5) 79.5 (18.4) 60.0 (20.5) 49.1 (19.3) 37.7 (20.1) < 0.001
 Change (95% CI) 34.4 (33.4–35.4) 41.5 (40.3–42.6) 24.5 (22.7–26.4) 12.7 (9.5–15.7) 4.4 (− 0.7 to 9.6) < 0.001
 p value** < 0.001 < 0.001 < 0.001 < 0.001 0.09

*t test
**Paired t test

Table 3  ROC curve analysis WOMAC Threshold value Sensitivity Specificity AUC​ 95% CI p value
identifying threshold values
for the components and total Pain > 78 80 76 80.6 78.0–82.2 < 0.001
WOMAC scores that predict
Function > 72 76 77 80.1 78.1–82.0 < 0.001
very satisfied from satisfied
patients Stiffness > 69 77 71 75.9 73.7–78.2 < 0.001
Total > 75 75 75 81.7 83.5–79.8 < 0.001

the study cohort. Another limitation of this study is the subjective entity, being closely linked with patient expec-
assessment of patient satisfaction at 1 year after surgery. tations [20]. This is likely reflected by the relatively low
Potentially some patients’ perception of pain and func- level of significance for the sensitivities, specificities and
tion may continue to improve after this time point and AUC. However, the p values still offered a moderate level of
hence their level of satisfaction may change [6]. However, reliability (> 60%). We also included a 10-year time period
a study of over 27,000 TKA performed in Sweden found of retrospective data collection, during which it could be
the level of patient satisfaction to be “remarkably con- argued that patient expectations of knee arthroplasty have
stant” 1 year after surgery for unrevised cases, with no evolved. However, it would be difficult to evaluate histori-
significant change with time [23]. Furthermore, we did not cal changes in patient expectations and there is no evidence
analyse the effect of factors which have previously been currently available demonstrating such an effect.
shown to influence patient satisfaction, such as gender, This study focused purely on patient-reported outcome
diagnosis, comorbidity, and mental health [3, 25], upon measures (PROMs) using WOMAC scores which are
the identified threshold values. However, these variables thought to be the primary measure of efficacy for osteo-
also influence the post-operative WOMAC score and so arthritis trials [11]. However, there are multiple validated
do not influence the threshold levels, i.e. a patient with PROM scores available, a systematic review by Ram-
diabetes is less likely to be satisfied and they also have a kumar et al. [22] included 38 studies using 47 PROMs.
lower post-operative WOMAC score, which predicts their With so many assessment tools available, there is a lack
lower level of satisfaction. of consistency throughout the literature regarding which
A further limitation of this study is the inherent diffi- most accurately reflects patient outcome. Alviar et al. [1]
culty in quantifying satisfaction which is a multifactorial and assessed eight instruments and found that the Knee Injury

13
 Knee Surgery, Sports Traumatology, Arthroscopy

Fig. 2  ROC curve for predict-

ing very satisfied patients from
satisfied patients using the post-
operative components and total
WOMAC scores

Table 4  ROC curve analysis WOMAC Threshold value Sensitivity Specificity AUC​ 95% CI p value
identifying threshold values
for the components and total Pain > 58 75 60 71.8 67.5–76.0 < 0.001
WOMAC scores that predict
Function > 54 68 65 71.1 67.1–75.1 < 0.001
satisfied from dissatisfied
patients Stiffness > 56 54 72 65.5 61.2–69.9 < 0.001
Total > 56 68 67 72.3 68.3–76.2 < 0.001

Table 5  ROC curve analysis WOMAC Threshold value Sensitivity Specificity AUC​ 95% CI p value
identifying threshold values
for the components and Pain > 44 76 64 74.4 68.0–80.9 < 0.001
total WOMAC scores that
Function > 41 68 68 71.5 64.4–78.6 < 0.001
predict dissatisfied from very
dissatisfied patients Stiffness > 43 64 62 65.1 57.7–72.5 < 0.001
Total > 43 68 70 73.7 66.8–80.5 < 0.001

and Osteoarthritis Outcome Score (KOOS) had the widest are similar to the threshold values of very satisfied and satis-
coverage for body function. However, the Arthritis Impact fied patients in the current study, hence using these values
Measurement Scales (AIMS) had the broadest bandwidth as a simple descriptive of an acceptable outcome would not
for activity and participation. Furthermore, the group who include those patients defining their outcome as satisfactory.
developed the tool, Roos et al. [24], intended for it to be used A PASS score of 83.5 points in the total WOMAC score was
on populations of young to middle-aged patients with post- defined more recently by Giesinger et al. [16], who declared
traumatic osteoarthritis (OA), or injuries predisposing that this higher score may be due to the anchor question used in
to such, not all patients with degenerative OA. their study. Using the 83.5 score would result in only half
The PASS score for the WOMAC scores have been of very satisfied patients and no satisfied patients achieving
defined for the pain, functional and total scores. Escobar a PASS in the WOMAC total score. The defined classifica-
et al. [14] defined the PASS score for pain to be 75 and 66 tion of the post-operative WOMAC scores allows a graded
for the functional WOMAC components. These two values assessment of the score and aids interpretation of different

13
Knee Surgery, Sports Traumatology, Arthroscopy

Fig. 3  ROC curve for predict-

ing satisfied patients from
dissatisfied patients using the
post-operative components and
total WOMAC scores

Fig. 4  ROC curve for predicting

dissatisfied patients from very
dissatisfied patients using the
post-operative components and
total WOMAC scores

13
 Knee Surgery, Sports Traumatology, Arthroscopy

Table 6  Outcome categories in the post-operative components and for this tool to be utilised in post-operative knee arthro-
total WOMAC scores as defined by the threshold values predictive of plasty patients for more focused assessment of the physical
level of patient satisfaction
as well as psychosocial outcomes of TKR.
WOMAC Excellent Good Fair Poor In addition, patients with lung disease, diabetes, gastric
ulcer, kidney disease, liver disease, with worse pre-oper-
Pain > 78 59–78 44–58 < 44
ative functional scores (WOMAC and SF-12) and those
Function > 72 54–72 41–53 < 41
with less of an improvement in the scores had a lower
Stiffness > 69 56–69 43–55 < 43
level of satisfaction and were shown to influence level of
Total > 75 56–74 43–55 < 43
patient satisfaction. The majority of studies assessing pre-
dictors of satisfaction assess this with a simple satisfied
and dissatisfied dichotomous variable [17]. The current
levels of outcome. However, the satisfaction classification is study assessed the effect of patient factors over four dif-
to be used in conjunction with the post-operative WOMAC fering levels of satisfaction, which may explain, in part,
scores, not as a replacement. the reason why more influencing factors were identified.
Classification of the post-operative score is a useful tool In addition, this study did not adjust for confounding fac-
and has been used to help interpret the Oxford score [8, 19, tors when assessing predictors of satisfaction, and if this
21] and KSS [2, 12] after TKA. The classification defined were done some factors may not reach significance. How-
by Kalairajah et al. [19] was based upon the correlation ever, it was not the primary aim of the study to identify
between the Harris hip score (which was arbitrarily clas- predictors of level of satisfaction, but this could be the
sified) [18] and Oxford hip score, which was then used to focus of future studies that could employ the classifica-
define the classification of the Oxford score [21]. Clement tion defined.
et al. [8] used patient satisfaction to defined the subgroups
in the post-operative Oxford knee score, which was found to
be different from that defined by Kalairajah et al. [19] using
the Oxford hip score. The classification of the KSS by Asif Conclusion
et al. [2], then subsequently quoted by Dowsey et al. [12],
was not based on any firm statistical method, simply being This study has defined a post-operative classification of
defined by the authors. The classification defined in the cur- excellent, good, fair and poor for the components and
rent study for the WOMAC score after TKA is based on the total WOMAC scores after TKA. The predictors of level
level of satisfaction and represents a meaningful categorisa- of satisfaction should be recognised in clinical practice
tion according to the patient’s perception of their outcome. and patients at risk of a lower level of satisfaction could be
This is a useful addition to how post-operative outcome is made aware in the pre-operative consent process. This may
assessed. The subjective patient perception of the result of help manage their expectations and facilitate the process
their surgery may have a greater personal impact upon them of informed consent.
than the more objective measures such as stiffness and func-
tion. Using this new categorisation may highlight patients Funding  The authors received no financial support for the research,
authorship, and/or publication of this article.
at higher risk of dissatisfaction leading to early interven-
tion to address problem areas and improve satisfaction with
outcome. Compliance with ethical standards 
Previous studies have also demonstrated that depres-
Conflict of interest  The authors declare no conflict of interest with the
sion [25], associated back pain [9], and a lower (worse) content of this study.
pre-operative WOMAC pain score [5] are independent
predictors of patient dissatisfaction 1  year after TKA, Ethical approval  The data collected formed part of the study centre’s
local joint registry which is registered as an ongoing service evalua-
which supports the findings of the current study. Driban
tion with Caldicott approval (reference number 2840). There was no
et al. [13] evaluated the use of patient-reported outcome additional patient contact, and as such, this project was performed as
measurement information system (PROMIS) instruments a service evaluation without the need for formal ethical approval. The
in patients with symptomatic knee osteoarthritis. These project was registered with the institutions audit department (registra-
tion number 8161) and was conducted in accordance with the Declara-
use several static short-form patient-reported outcome
tion of Helsinki and the guidelines for good clinical practice.
measures. They demonstrated good performance of the
tool in these patients, in particular, the PROMIS anxi- Informed consent  The patients gave informed written consent to have
ety and depression targets general mental health and pain their anonymised data collected onto the study centre’s registry and for
analysis for service evaluation purposes.
interference and physical function static short forms target
whole body outcomes. Therefore, there may be potential

13
Knee Surgery, Sports Traumatology, Arthroscopy

References 15. Farrar JT, Young JP Jr, LaMoreaux L, Werth JL, Poole RM (2001)
Clinical importance of changes in chronic pain intensity measured
on an 11-point numerical pain rating scale. Pain 94:149–158
1. Alviar MJ, Olver J, Brand C, Hale T, Khan F (2011) Do patient-
16. Giesinger JM, Hamilton DF, Jost B, Behrend H, Giesinger K
reported outcome measures used in assessing outcomes in reha-
(2015) WOMAC, EQ-5D and knee society score thresholds for
bilitation after hip and knee arthroplasty capture issues relevant to
treatment success after total knee arthroplasty. J Arthroplasty
patients? Results of a systematic review and ICF linking process.
30:2154–2158
J Rehabil Med 43(5):374–381
17. Gunaratne R, Pratt DN, Banda J, Fick DP, Khan RJK, Rob-
2. Asif S, Choon DS (2005) Midterm results of cemented press fit
ertson BW (2017) Patient dissatisfaction following total knee
condylar sigma total knee arthroplasty system. J Orthop Surg
arthroplasty: a systematic review of the literature. J Arthroplasty
(Hong Kong) 13:280–284
32(12):3854–3860
3. Baker PN, van der Meulen JH, Lewsey J, Gregg PJ (2007) The
18. Harris WH (1969) Traumatic arthritis of the hip after dislocation
role of pain and function in determining patient satisfaction after
and acetabular fractures: treatment by mold arthroplasty. An end-
total knee replacement. Data from the National Joint Registry for
result study using a new method of result evaluation. J Bone Joint
England and Wales. J Bone Joint Surg Br 89:893–900
Surg Am 51:737–755
4. Bellamy N, Buchanan WW, Goldsmith CH, Campbell J, Stitt LW
19. Kalairajah Y, Azurza K, Hulme C, Molloy S, Drabu KJ (2005)
(1988) Validation study of WOMAC: a health status instrument
Health outcome measures in the evaluation of total hip arthroplas-
for measuring clinically important patient relevant outcomes to
ties—a comparison between the Harris hip score and the Oxford
antirheumatic drug therapy in patients with osteoarthritis of the
hip score. J Arthroplasty 20:1037–1041
hip or knee. J Rheumatol 15:1833–1840
20. Mahomed NN, Liang MH, Cook EF, Daltroy LH, Fortin PR, Fos-
5. Bourne RB, Chesworth BM, Davis AM, Mahomed NN, Charron
sel AH, Katz JN (2002) The importance of patient expectations
KD (2010) Patient satisfaction after total knee arthroplasty: who
in predicting functional outcomes after total joint arthroplasty. J
is satisfied and who is not? Clin Orthop Relat Res 468:57–63
Rheumatol 29(6):1273–1279
6. Brander V, Gondek S, Martin E, Stulberg SD (2007) Pain and
21. Murray DW, Fitzpatrick R, Rogers K, Pandit H, Beard DJ, Carr
depression influence outcome 5 years after knee replacement sur-
AJ, Dawson J (2007) The use of the Oxford hip and knee scores.
gery. Clin Orthop Relat Res 464:21–26
J Bone Joint Surg Br 89:1010–1014
7. Burch FX, Tarro JN, Greenberg JJ, Carroll WJ (2008) Evaluating
22. Ramkumar PN, Harris JD, Noble PC (2015) Patient-reported
the benefits of patterned stimulation in the treatment of osteoar-
outcome measures after total knee arthroplasty. Bone Joint Res
thritis of the knee: a multi-center, randomized, single-blind, con-
4(7):120–127
trolled study with an independent masked evaluator. Osteoarthr
23. Robertsson O, Dunbar M, Pehrsson T, Knutson K, Lidgren L
Cartil 16:865–872
(2000) Patient satisfaction after knee arthroplasty: a report on
8. Clement ND, MacDonald D, Burnett R (2013) Predicting patient
27,372 knees operated on between 1981 and 1995 in Sweden.
satisfaction using the Oxford knee score: where do we draw the
Acta Orthop Scand 71:262–267
line? Arch Orthop Trauma Surg 133:689–694
24. Roos EM, Roos HP, Lohmander LS, Ekdahl C, Beynnon BD
9. Clement ND, MacDonald D, Simpson AH, Burnett R (2013) Total
(1998) Knee injury and osteoarthritis outcome score (KOOS)—
knee replacement in patients with concomitant back pain results in
development of a self-administered outcome measure. J Orthop
a worse functional outcome and a lower rate of satisfaction. Bone
Sports Phys Ther 28(2):88–96
Joint J 95-B:1632–1639
25. Scott CEH, MacDonald D, Howie CR, Biant LC (2010) Predicting
10. Dawson J, Doll H, Fitzpatrick R, Jenkinson C, Carr AJ (2010) The
patient dissatisfaction following total knee replacement: a prospec-
routine use of patient reported outcome measures in healthcare
tive study of 1217 cases. J Bone Joint Surg Br 92-B:1253–1258
settings. BMJ 340:c186. https​://doi.org/10.1136/bmj.c186
26. Singh J, Sloan JA, Johanson NA (2010) Challenges with health-
11. Dieppe PA (1995) Recommended methodology for assessing the
related quality of life assessment in arthroplasty patients: prob-
progression of osteoarthritis of the hip and knee joints. Osteoarthr
lems and solutions. J Am Acad Orthop Surg 18:72–82
Cartil 3:73–77
27. Tubach F, Ravaud P, Baron G, Falissard B, Logeart I, Bellamy N,
12. Dowsey MM, Choong PF (2013) The utility of outcome measures
Bombardier C, Felson D, Hochberg M, van der Heijde D, Dou-
in total knee replacement surgery. Int J Rheumatol 2013:506–518
gados M (2005) Evaluation of clinically relevant states in patient
13. Driban JB, Morgan N, Price LL, Cook KF, Wang C (2015)
reported outcomes in knee and hip osteoarthritis: the patient
Patient-reported outcomes measurement information system
acceptable symptom state. Ann Rheum Dis 64:34–37
(PROMIS) instruments among individuals with symptomatic knee
28. Ware J Jr, Kosinski M, Keller SD (1996) A 12-item short-form
osteoarthritis: a cross-sectional study of floor/ceiling effects and
health survey: construction of scales and preliminary tests of reli-
construct validity. BMC Musculoskelet Disord 16:253. https:​ //doi.
ability and validity. Med Care 34:220–233
org/10.1186/s1289​1-015-0715-y
29. Woolacott NF, Corbett MS, Rice SJ (2012) The use and reporting
14. Escobar A, Riddle DL (2014) Concordance between important
of WOMAC in the assessment of the benefit of physical therapies
change and acceptable symptom state following knee arthroplasty:
for the pain of osteoarthritis of the knee: findings from a system-
the role of baseline scores. Osteoarthr Cartil 22:1107–1110
atic review of clinical trials. Rheumatology 51:1440–1446

13