Neuropsychological detection of cognitive impairment: inter-rater agreement and factors affecting clinical decision-making.

The agreement between neuropsychologists identifying cognitive impairment (CI) in older adults was examined, as were factors influencing the classification process. Twenty four neuropsychologists in 18 study centers classified cases with or without CI after reviewing neuropsychological findings and other relevant information. All cases were participants in the third wave of the Canadian Study of Health and Aging, a study of CI in later life. For 117 randomly selected cases, a second neuropsychologist reviewed the same material and reclassified the cases. Cases given the same (concordant) or different (discordant) classifications were compared with respect to patient and rater characteristics. The inter-rater agreement was moderate (77.7% agreement, kappa = .49). On all measures of cognitive functioning, the concordant group without impairment obtained a higher mean score than the discordant group, and the discordant group obtained a higher mean score than the concordant group with impairment. For 5 out of 8 cognitive measures, the concordant group with impairment differed from the concordant group without impairment and the discordant group, but the latter two groups did not differ significantly. The findings are comparable to others in the field and highlight the need for neuropsychologists to further clarify procedures for identifying subtle, or mild, forms of cognitive impairment.