Improving Patient Reported Measures for Knee Function

Abstract

Knee injuries account for nearly 15% of all sports-related injuries. Patient-reported outcomes (PRO) such as the Knee Injury Osteoarthritis Outcomes Score (KOOS) and the International Knee Documentation Committee Subjective Knee Form (IKDC-SKF) have been used to evaluate changes in function following knee injuries. These PRO measures, however, contain significant ceiling effects which limit their clinical utility when used to monitor athlete function. Therefore, the purpose of this project was to evaluate and improve PRO measures for athletes through a series of studies.

In the first study, the measurement properties of three PRO (e.g., KOOS, IKDC-SKF, and Marx Activity Rating Scale) measures were examined using the Rasch partial-credit model. Among the 65 items examined in the first study only 30 yielded acceptable model-data fit. Furthermore, measurement precision decreased as function increased. To improve functional assessment a new PRO measure was developed by combining the acceptable items from the first study with 12 new items.

In the second study the psychometric properties of the new PRO measure were evaluated. Thirty-two student-athletes were recruited to pilot test the Patient-Reported Knee Outcomes Assessment Tool (PRKOAT). Feedback regarding readability and relevance were used to modify the instrument reducing the total number of items to 32. the revised instrument was calibrated using a sample of 203 student-athletes. Rasch partial-credit model analysis indicated that majority of the items (n = 27) had acceptable model-data fit. Poor fitting items were removed.

Item difficulty estimates ranged from -4.74 - 1.89 logits. Person ability estimates ranged from -3.24 - 2.29. Injured athletes (mean score 39.25 14.00), scored significantly higher (lower knee function) than non-injured athletes (11.93 10.78) (t188 = 12.89; p < .01) providing evidence of known-group difference validity. Among the participants sampled in this study only 3.9% obtained the best score (i.e., floor effect), while none of the participants obtained the worst score (i.e., ceiling effect).

A major advantage of this study was the use of advanced measurement theory. Compared to commonly used PRO measures the PRKOAT has fewer ceiling effects in athletic populations. Clinicians should consider using the PRKOAT to evaluate PRO throughout the rehabilitation process following knee-related injuries.