Chronic obstructive pulmonary disease (COPD) is a progressive disease with functional decline leading to disability. Dyspnea, the prominent symptom, can be measured using existing measures, but a lack of consensus about standardization of dyspnea measurement remains. We examined the psychometric performance of two item-response theory–based (IRT) measures of dyspnea and related functional limitations (FLs) in patients with COPD and simulated computerized adaptive testing (CAT) of the banks to determine the number of questions required to achieve high precision.
A total of 102 patients completed banks measuring dyspnea and FLs (33 items), from which the 10-item dyspnea and FL short forms were scored as well as other self-report measures of respiratory and physical function and emotional distress. A subset of patients completed the banks 7 to 10 days later. Pulmonary function test results were obtained from medical charts.
The 33-item banks and 10-item short forms had excellent internal consistency (alphas >0.9) and test-retest reliability (intraclass correlation coefficients >0.89). Patients sorted by severity level on the Medical Research Council scale were differentiated by item banks (P 0.01). The banks and short forms were also associated with related measures of dyspnea (e.g., Baseline Dyspnea Index, r = 0.47–0.53), physical function (e.g., 36-Item Short Form Health Survey, r = −0.83 to −0.86) and forced expiratory volume in 1 second (r = −0.32 to −0.35). On average, CAT required 4 and 5 items for accurate measurement of dyspnea and FLs, respectively.
The Functional Assessment of Chronic Illness Therapy–Dyspnea short forms and banks provide options for brief, psychometrically sound measures of dyspnea and/or FLs in COPD.