Dw. Oconnor et al., CROSS-NATIONAL INTERRATER RELIABILITY OF DEMENTIA DIAGNOSIS IN THE ELDERLY AND FACTORS ASSOCIATED WITH DISAGREEMENT, Neurology, 47(5), 1996, pp. 1194-1199
Thirteen researchers from five centers in Australia, Germany, the Neth
erlands, United Kingdom, and United States applied DSM-III-R and Clini
cal Dementia Rating (CDR) syndrome-level dementia criteria to written
vignettes of 100 elderly people identified in clinics or community sur
veys. Subjects ranged in type from cognitively intact to severely deme
nted and many were also frail, partially sighted, or deaf. This paper
concerns reliability within and between centers, and the relationship
between reliability and factors such as diagnostic criteria, dementia
severity, and respondents' clinical characteristics. Within-center int
errater reliability was high, more so for ''yes-no'' DSM-III-R diagnos
es than the multi-level CDR. Between-center rates were lower but still
moderate to good. Concordance was lower for intermediate dementia lev
els than for no dementia and severe dementia. Physical disability made
an additional contribution to uncertainty but deafness, poor vision,
anxiety, and depression had no discernible effects. Reliability levels
are likely to be lower in representative aged populations than in car
efully selected clinical groups.