INTRODUCTIONThe objectives of this study were to assess the utility and validity of the self-administered SF-36 and the effect of visual or cognitive impairment on these measures in an older population.

MATERIALS AND METHODSAttempt rates, completion rates and internal consistency (Cronbach alpha) were compared within the second cross-sectional, population-based Blue Mountains Eye Study (n = 3509, mean age 66.7 years, 57% women).

RESULTSThe SF-36 was attempted by 3162 (90.1%) participants, of which 2470 (78.1%) completed all items and 2873 (90.9%) completed sufficient items for calculation of all dimensions. In a multivariate model adjusting for age, sex, and presenting visual and cognitive impairments, women (P = 0.011) and participants with visual or cognitive impairments (P < 0.0001) were less likely to attempt the questionnaire. Completion rates were significantly lower with increasing age (P < 0.0001), in men (P < or = 0.0005) and in those with cognitive impairment (P < 0.0001). A high level of internal consistency (Cronbach alpha > 0.85 for all dimensions) and construct validity was demonstrated, the latter distinguishing between those with and without medical conditions, disabilities or recent hospital admissions (P < 0.01). As the prevalence of visual or cognitive impairment was relatively low in this population, we found no apparent effect of these impairments on the validity of SF-36.

CONCLUSIONSAttempt and completion rates, but not internal consistency and construct validity, of the SF-36 were influenced by age, gender, and presenting visual and cognitive impairments. The overall high attempt and completion rates, internal consistency and construct validity suggest that the self-administered SF-36 is a suitable healthrelated quality of life (HRQOL) measure in similar older community-living populations.