Computerized adaptive testing (CAT) and Optimal Test Assembly (OTA) methods that used the newly developed Rasch Everyday Activity Limitations (REAL) item bank can accurately assess physical function for patients with rheumatoid arthritis (RA), according to results published in Rheumatology.
The study included participants with RA from the rheumatology clinic at Medisch Spectrum Twente (n=209). Participants used a web portal to fill out the REAL-CAT, Health Assessment Questionnaire Disability Index (HAQ-DI), and the Medical Outcomes Study Short Form 36. The researchers also developed a 6-item short form (REAL-6) using OTA, and participants also completed the REAL-6. The researchers evaluated measurement precision using item response theory methods and evaluated construct validity by testing hypothesized correlations with external measures.
The REAL-CAT model had better measurement precision compared with both the HAQ-DI and REAL-6. However, the REAL-6 still outperformed HAQ-DI in all but the very lowest level of physical function. Both REAL-CAT and REAL-6 had a greater range for which scores were reliable compared with HAQ-DI.
All 3 instruments had good construct validity with >75% of hypotheses confirmed. Only HAQ-DI demonstrated a ceiling effect (23.9%).
“We were…able to demonstrate considerable gains in measurement precision using REAL-derived CAT and OTA measures compared with the HAQ-DI, while at the same time decreasing the average number of items that patients have to respond considerably,” the researchers wrote.
Reference
Oude Voshaar MAH, Ten Klooster P, Vonkeman HE, van de Laar MAFJ. Rasch measurement in rheumatoid arthritis: deriving psychometrically optimal measures from the Rasch Everyday Activity Limitation item bank [published online June 23, 2018]. Rheumatology (Oxford). doi:10.1093/rheumatology/key154