The KOOS-PS is a 7-item measure of physical functional derived from the items of the Function, daily living and Function, sports and recreational activity subscales of the KOOS. As with the KOOS it is intended to elicit people’s opinions about the difficulties they experience with activity due to problems with their knee.

Why was a short measure developed?
The need for a short measure was identified as part of an OARSI (Osteoarthritis Re­search Society International) and OMERACT (Outcome Measures in Clinical Trials)initiative to develop criteria for the failure of medical therapy, more specifically disease modifying drugs for osteoarthritis (DMOADS), such that individuals would be candidates for surgical intervention such as joint replacement.

The components of the criteria were determined by the working group to include pain, physical function and structural damage [1].

Given the overall objective, it was critical that each component be parsimonious and cross-culturally valid to enable use of the measure in multi-centre, cross-country studies and clinical settings.

As a result, the decision was made to attempt to shorten the KOOS and to combine the items from the Function, daily living and Function, sports and recreational activity subscales.

We chose to use an item response theory approach, specifically the Rasch mathematical model, to develop the short measure. Rasch is a mathematical process that converts ordinal data (such as the responses in the KOOS-PS of none, mild, moderate, severe, extreme) to interval level data [2].

Ordinal data have logical order but the distances between question responses like ’none’ and ’mild’ and ’mild’ and ‘moderate’ may be different. In contrast, interval level data has order and the distances be­tween scores are equal. The advantage of interval level data is that we can accurately quantify physical disability to be able to compare individuals and groups and to know how much they have changed.

On a measure that has interval level scaling properties, we know that people who change from 10 to 20 points and those who change from 60 to 70 points have all changed exactly ten points and that this ten points is equal along the scale range. On an ordinal scale, although the questions are often summed, it is often assumed but not confirmed that this scenario would represent 10 points that are of equal distance along the scale.

The most basic form of the Rasch model, based on a dichotomous response scale, is that the probability of a person endorsing an item is a logistic function of the difference between the person’s ability and the difficulty of the item. This can be expressed as a logit model:

where ln is the logarithm function, p is the probability of person n endorsing item i, φn is the level of functional ability of person n, and bi is the difficulty of item i. The item and person estimates are expressed as logits which allows for linear transformation of the raw score. As an item estimate is based on responses to the other items, the model is able to accommodate missing responses to an item for a given respondent. In the case of multiple response options such as with the KOOS, item and person score estimates are similarly estimated by extension of the model [3,4].

The Rasch model provides a number of advantages over classical test theory methods. Specifically, it provides scores that have the requisite interval level measurement properties for inferential statistical procedures; estimates place items in a hierarchy of difficulty such that range of difficulty represented by the measure is known; and, the standard error of an item is independent of the SE of other items such there is improved accuracy and stability of the performance of the items across different samples [3,4].

The final items included in the short measure were based on the data from the samples and countries described below meeting the criteria for fit to the Rasch model [5].

What is the purpose of the KOOS-PS?
The KOOS-PS can be used as part of the criteria, once studies are completed, to determine failure of DMOADS. The KOOS-PS (with only 7 items instead of the 22 of the KOOS Function, daily living and Function, sports and recreational activity subscales) can be used when fewer items are advantageous to limit responder burden and/or cost.

For example, the short questionnaire may be preferable when several questionnaires are administered simultaneously, in registries, or in special populations such as the elderly.

For what populations is the KOOS-PS relevant?
The initial development of the short measure included people with a range of knee problems from multiple countries [5].

Individuals were from community and clinical samples (e.g., people many years post meniscetcomy, following tibial osteotomy or anterior cruciate repair and just prior to having their knee replaced) and were from Sweden, Canada, France, Estonia, and the Netherlands. These individuals ranged from 26 to 95 years.

What are the measurement properties of the KOOS-PS?
The internal consistency of the KOOS-PS was 0.90 and it represents a unidimensional construct by virtue of the fit of the data to the Rasch model [5].

How do I use and score the measure?
The measure is intended to be self-completed and, as with the full version of the KOOS, can be completed as a mailed survey or in the waiting room.

The measure is scored by summing the responses to the 7 items of the KOOS-PS. The interval score from 0 to 100 with zero representing no difficulty is obtained by using the published conversion chart [5].

What studies are currently using the KOOS-PS?
The measure is currently under study in multiple countries to determine the threshold at which individuals are considered candidates for joint replacement as part of an initiative. The first phase of this work was cross-cultural validation of the measure through a for­mal process of forward- and back-translation with reconciliation of language and con­ceptual meaning differences where no culturally adapted version of the KOOS existed. The second phase includes completion of the questionnaire by patients referred to an orthopaedic surgeon for consideration of knee replacement. These data will be com­pared to the surgeon’s decision about the recommendation
for surgery. This study is being conducted in Australia, Canada, the Czech Republic, France, Germany, Italy, the Netherlands, Spain, Sweden, Britain and the United States.

Additionally, the KOOS-PS is being evaluated in people undergoing total joint replace­ment to further test its reliability, validity and responsiveness in Canada.

References

1. Gossec, L., et al., OMERACT/OARSI initiative to define states of severity and indication for joint replacement in hip and knee osteoarthritis. Journal of Rheumatology, 2007; 34(6): 1432-35.

2. Tennant A. Conaghan PG. The Rasch measurement model in rheumatology: what is it and why use it? When should it be applied, and what should one look for in a Rasch paper? Arthritis & Rheumatism. 2007; 57(8):1358-62.

3. Rasch, G., Probablislistic Model for Some Intelligence and Attainment Tests. Reprint ed. 1960, Chicago: University of Chicago Press.

4. Andiel, C., Rasch analysis: A description of the model and related issues. Canadian Journal of Rehabilitation, 1995; 9(1): 17-25.

5. Perruccio AV, Lohmander, LS, Canizares M, Tennant A, Hawker GA, Conaghan PG, Roos EM, Jordan JM, Maillefert JF, Dougados M, Davis AM. The Development of a Short Measure of Physical Function for Knee OA - KOOS-Physical Function Short-form (KOOS-PS) – An OARSI/OMERACT Initiative. Osteoarthritis and Cartilage, 2008; 16: 542-550.