Convergent validity of the Cognitive Performance Scale of the interRAI acute care and the mini-mental state examination

Abstract

Objective: The Cognitive Performance Scale (CPS) is generated from five items of the interRAI/ Minimum Data Set instruments, a comprehensive geriatric assessment method. CPS was initially designed to assess cognition in residential care, where it has shown good psychometric performance. We evaluated the performance of the interRAI Acute Care in identifying cognitive impairment among patients hospitalized on acute geriatric wards.

Methods: An observational study was conducted on two geriatric wards. Trained raters independently completed the interRAI Acute Care and the Mini-Mental State Examination (MMSE) in 97 inpatients (85 ± 5 years; 67% female). The level of agreement between CPS and MMSE was explored using comparisons of means, agreement coefficients, and diagnostic accuracy.

Results: Cognitive impairment was present in 61% of the participants. Average MMSE scores were significantly different between groups with low CPS scores compared with those with high CPS scores (p <0.05). CPS explained only 48.8% of the variability in MMSE. Agreement in defining cognitively impaired subjects was moderate (percentage observed agreement, 68%; κ = 0.41). With MMSE score less than 24 as a gold standard, diagnostic accuracy of CPS was moderate (area under curve = 0.73), with low sensitivity, but excellent specificity. When lowering the MMSE cutoff to less than 18 and focusing on patients with severe cognitive impairment, CPS agreement coefficients and sensitivity increased but specificity decreased. Using education-adjusted MMSE cutoffs did not substantially affect the results.

Conclusion: CPS can be used for coarse triage between intact and severe cognitive impairment. Although promising results have been obtained in residential and community settings, our results suggest that CPS fails to differentiate across different levels of cognitive impairment in hospitalized geriatric patients.