Agreement in the clinical diagnosis of dementia: evaluation of a case series with mild cognitive impairment

Neuroepidemiology. 1994;13(3):89-96. doi: 10.1159/000110364.


Interobserver agreement in the clinical diagnosis of dementia among four neurologists was evaluated. The physicians, masked to the original diagnoses, independently reviewed the clinical records of 50 outpatients consulting either the 1st University Neurology Department of Milan or the Neuropsychology Unit of the Medical Center of Veruno (Novara) for suspected cognitive impairment, during a 6-month period. The records contained patients' medical and neurological history, results of neuropsychological testing, laboratory tests, cerebral computed tomography and other investigations. For each patient, the raters had to provide both a diagnosis concerning the presence or absence of dementia and to assign an analytical diagnosis to all the dementia cases. The kappa statistic was used as a measure of interrater reliability. The level of agreement on the primary diagnosis of dementia was moderate (kappa = 0.49); with respect to the nosological diagnoses, the kappa values ranged from 0.16 for depression to 0.80 for multi-infarct dementia.

MeSH terms

  • Aged
  • Dementia / diagnosis*
  • Female
  • Humans
  • Male
  • Middle Aged
  • Neuropsychological Tests*
  • Observer Variation*
  • Reproducibility of Results