The Swedish RAND-36: psychometric characteristics and reference data from the Mid-Swed Health Survey

Abstract

Background: This study aims to evaluate data quality, scaling properties, and reliability of the Swedish RAND-36 in a general population sample and to present reference data for the Swedish population. Methods: Testing of data quality, scaling assumptions and reliability followed methods recommended for the International Quality of Life Assessment Project, previously used for psychometric testing of SF-36 and RAND-36. Data were collected via regular mail for a random stratified sample of the general population in a Swedish county. Weighted means for RAND-36 scores were used and differences by sex, age, education, and occupational groups were tested. Results: The response rate was 42%, and the sample comprised 3432 persons (45% men, 55% women) with a median age of 56.9 years. The internal consistency reliability was satisfactory, with Cronbach’s alphas > 0.80 for all eight scales. The percentage of missing items was low, ranging between 1.3% and 3.2%. No floor effects (≥15%) were noted, while ceiling effects were observed for physical functioning, role-functioning/physical, pain, role-functioning/emotional, and social functioning. Item–scale correlations were satisfactory (r ≥ 0.40). Correlations among the physical health scales were strong (range 0.58–0.68) as were the correlations among the mental health scales (range 0.58–0.73). Men reported significantly better health-related quality of life (HRQoL) on all scales, although the gender differences were small. Comparisons among age groups showed approximately equal scores among those 20–29, 30–39, and 40–49 years, while significant decreases in physical health were observed in the older age groups. Substantially worse physical health scores were observed in the oldest age group (80+). Significant differences among age groups were noted also for the mental health scales; however, better energy/fatigue and emotional well-being scores were seen in the older age groups, except for the oldest (80+). Those with university education reported significantly better scores on all scales compared to those with mandatory education. Conclusions: The study suggests that the Swedish version of RAND-36 is an acceptable and reliable instrument for measuring HRQoL in the general population. The study provides reference data that can be used for norm-based comparisons.