ABSTRACT
Background
Reports of neuropsychiatric symptoms (NPS) from informants, often patients’ caregivers and families, and from clinicians are both important for accurate detection of symptoms. However, informant/caregiver report and clinician assessments of NPS often differ.
Methods
We examined agreement between informant/caregiver and clinician report of NPS in the National Alzheimer’s Coordinating Center Uniform Data Set. Participants were age ≥ 50 at baseline with mild cognitive impairment (MCI) or dementia (N = 27,225). At each visit, informants reported NPS using the Neuropsychiatric Inventory questionnaire (NPI-Q). Study clinicians provided clinical assessment per study protocol. Agreement between informant report and clinician judgment was assessed using Cohen’s kappa statistic. Associations between agreement in reporting for each NPS and participant/informant characteristics were examined using random-effects logistic regressions.
Results
At baseline, participants were on average 72.9 ± 9.4 years old, 49% male, 76% non-Hispanic White, with 14.8 ± 3.6 years of schooling. Average follow-up was 4.0 ± 2.5 years. Informants were 63.7 ± 13.2 years old, 31% male, with 15.4 ± 2.8 years of schooling. 60% of informants were spouse/partner of the participant. Informants were more likely than clinicians to report the presence of all symptoms except for hallucinations. Agreement between informant and clinician reports of all symptoms was lower in patients with more severe dementia. Over time, agreement between informant and clinician reports of apathy increased. Agreement between informant and clinician reporting of NPS differed by participant’s sex and race/ethnicity. Informants who had lower frequency of contact and more distant relationships with the participant were more likely to agree with clinicians’ reporting.
Conclusions
Understanding differences between informant and clinician reports of NPS in dementia is essential in obtaining a more complete, accurate picture of behavioral challenges patients face. Considering patient and informant characteristics and dynamics between them would help clinicians better understand potential biases that may affect the accuracy of reported NPS and better manage and treat the symptoms.