Proxy reporting in five areas of functional status. Comparison with self-reports and observations of performance

Proxy ratings of functional status were compared with subject self-reports in five domains relevant to the study of older persons and with observations of subject performance in two areas (physical and instrumental functioning). Data were derived from 233 proxy-subject pairs evaluated in a prospective study of hip fracture patients aged 65 years or more in Baltimore, Maryland (1990-1991). Agreement between proxy and subject reports was highest for a summary measure of instrumental functioning and lowest for a measure of depression. Proxies tended to report more disability than did subjects, although bias varied by function. Patterns of agreement for proxy reports versus observations of performance compared with patterns for proxy reports versus subject reports were lower for measures of instrumental functioning, and bias was generally more extreme for instrumental and physical functioning measures. The authors conclude that agreement and bias differ by functional domain, by the way summary measures are created and scored, and by the criterion against which proxy reports are compared.