255considered more of a measure of pain during ADL, as it does not consider how OA
pain impacts on other areas of life (e.g., sleep and mood) [ 24 , 25 ]. There may also
be issues of content validity for more active patients with OA, especially younger
adults with post-traumatic OA, as there are no items in either the pain or function
subscales that relate to more vigorous activities (e.g., running).
All three WOMAC subscales demonstrate adequate reliability for use in groups
of patients with hip OA (intraclass correlation coeffi cient [ICC] ≥ 0.8). The function
subscale is suffi ciently reliable for groups of patients with knee OA, but there is
confl icting evidence for the pain and stiffness subscales. Clinicians wishing to use
WOMAC in individual patients also need to consider the evidence regarding test–
retest reliability. While the function subscale consistently demonstrates adequate
reliability for use in individual patients with hip OA (ICC ≥ 0.9), there is confl icting
evidence for knee OA. For the pain subscale in both hip and knee OA, some studies
show adequate reliability, and others inadequate reliability. The stiffness subscale
consistently demonstrates inadequate reliability for individual patients with knee or
hip OA. This may be because this subscale only contains two items. Therefore,
clinicians should be aware of the likelihood that changes in WOMAC scores
observed in individual patients may represent error in the instrument, rather than
true change in the patient’s condition.
Clinicians can use values for measurement error (minimal detectable change
[MDC]) to determine the minimum score that represents a true change in an indi-
vidual patient’s condition. In 95 % of cases (MDC 95 ), individual patients with hip
OA will have experienced a real change if their function subscale score has changed
by at least 9.1 points (when scored using the Likert version, score range 0–68) [ 26 ].
Values for the pain and stiffness subscales are unknown for patients with hip OA. For
individual patients with knee OA, the change in pain subscale score should exceed
18.8, stiffness 27.1, and function 13.3 to be considered a real change in condition,
when evaluated using the 11-point numerical rating scale version (scores converted
to 0–100) [ 27 ].
All three subscales of the WOMAC are able to detect improvements with exer-
cise and physical therapy, pharmacological interventions, and total joint replace-
ment [ 28 ]. The pain and function subscales have no fl oor and ceiling effects in
patients with hip and knee OA, meaning that these subscales can monitor deteriora-
tion and improvement in a patient’s condition over time. Ceiling effects have only
been noted after total joint replacement. In contrast, the stiffness subscale has dem-
onstrated both fl oor and ceiling effects in patients with knee OA [ 24 ] and may not
be an ideal tool to evaluate change over time.
A major limitation of WOMAC is that it is protected by copyright and trademark,
and therefore requires permission for use. Although licensing and costs are deter-
mined on each individual request, clinicians should be aware that there may be fees
associated with the use of WOMAC with patients. This restricts the accessibility of
WOMAC to all patients with hip/knee OA, despite the multitude of language ver-
sions available.
10 PROMs for Osteoarthritis