Published online by Cambridge University Press: 13 March 2014
Neuropsychiatric symptoms (NPS) are highly prevalent in dementia. The recently developed Neuropsychiatric Inventory – Clinician rating scale (NPI-C) includes clinical judgment and new symptom domains. Our objective was to evaluate NPI-C reliability and to compare caregiver and clinician ratings across the range of mild to severe cognitive impairment.
This is a cross-sectional observational study. Participants were geriatric memory clinic patients and nursing-home residents (n = 30) with an established diagnosis of dementia or Mild Cognitive Impairment (MCI). A psychiatrist (MK) interviewed caregiver–patient dyads using the NPI-C. Neuropsychological tests and Mini-Mental State Examination (MMSE) were used to assess cognitive impairment. Two NPI-C caregiver interviews were videotaped and rated by psychologists and geriatricians. Intra-class correlations (ICCs) were used to examine inter-rater agreement. Correlation coefficients were calculated to evaluate caregiver and psychiatrist NPI-C ratings. Disagreement between caregiver and clinician was expressed in delta scores and examined across the range of mild to severe cognitive impairment, using Levene's homogeneity of variances tests.
Inter-rater agreement on ratings of two caregiver videos was high (ICC = 0.99–1.0). Clinician–caregiver concordance on NPI-C total severity ratings was high (r = 0.77). Variability in clinician–caregiver concordance was associated with cognitive impairment: MMSE (P = 0.02), CAMCOG-R (Cambridge Cognitive Examination-revised) total scores (P = 0.02), CAMCOG-R Memory scores (P = 0.04) and Language scores (P = 0.01).
The NPI-C is a reliable measure of NPS in patients with MCI or dementia. Clinician–caregiver agreement on NPS severity may vary with cognitive impairment, underlining the importance of clinician-based measures of NPS.