An English and Spanish quality of life measure for rheumatoid arthritis

Objective. To develop a rheumatoid arthritis-specific health-related qualit y of life instrument, translate the English instrument into Spanish, and te st the scaling assumptions, reliability, validity, and feasibility of both the English and Spanish versions. Methods. The development of the Quality of Life-Rheumatoid Arthritis Scale (QOL-RA Scale) involved literature review, consultations with experts, 40 f ace-to-face interviews, and 5 focus group discussions with multiethnic and multilingual women with rheumatoid arthritis (RA). Translation design facil itated conceptual and linguistic equivalence. Data for the psychometrics ca me from telephone interviews of a sample of 107 Caucasian/English and 80 Hi spanic/Spanish women with RA. The instruments were (a) the Arthritis Impact Measurement Scales 2 (AIMS2), (b) the Lubben Social Network Scale (LSNS), (c) the Center for Epidemiologic Studies-Depression Scale (CES-D), and (d) the QOL-RA Scale. Descriptive statistics, significance tests, Cronbach's al pha technique, correlation, and factor analysis were used. Results. The QOL-RA Scale, an 8-item scale, took 2 to 3 minutes to administ er. Psychometric analysis revealed that the psychometric attributes and con structs of both English and Spanish questionnaires are comparable (i.e., eq uivalent). Both versions demonstrated the following: (a) normal distributio n of the QOL-RA Scale, roughly symmetrical distributions of the items, equi valent means and standard deviations across items, and less than 10% floor and ceiling effects, (b) Cronbach's alpha coefficients of 0.87-0.90, (c) si gnificant correlations of the QOL-RA Scale with the AIMS2 subscales, LSNS, and CES-D, ranging from 0.25 to 0.66 (P less than or equal to 0.01), and (d ) extraction of 2 factors, namely physio-psychological and socio-psychologi cal, that explained 65% to 73% of the variance in the scale scores. Conclusion. The QOL-RA Scale, in both English and Spanish versions, appears to meet the assumptions of a summated rating scale and the criteria of rel evance, reliability, validity, feasibility, and adaptability to several lan guages.