Use of the Barthel Index, mini mental status examination and discharge status to evaluate a special dementia care unit| JOURNAL OF GERONTOLOGY AND GERIATRICS

Abstract

Background. When dementia patients are grouped according to their ordinal Barthel Index and Mini Mental
Status Examination sum scores, it is not clear which portions of the data should be used when valid classifications are developed. Criteria used for classification of patients must be stochastically independent.
Objectives: the relationship between Barthel-Index, Mini-Mental State Examination sum scores and discharge
status was investigated to develop subgroups of dementia patients. The developed classification uses stochastically independent information and can be used to evaluate special care units.
Methods. We used an unrestricted partial credit model to assess the possible scores on the Barthel-Index. We
investigated the individual effects of items on the Barthel sum score by using non-parametric conditional-inference-regression trees. The relationships between Barthel score, Mini Mental Status score, and discharge status, in terms of classifying the dementia patients into subgroups, were investigated using a latent class analysis.
Results. An interval scale Barthel-Index did not yield a significant improvement versus the ordinal Barthel-Index
sum score. Differences in Barthel-Index were meaningful only in the context of three groups using four items.
A classification of dementia patients in latent classes could be developed using three Barthel – and Mini Mental
Status-Groups and the discharge status of patients who were living at home before admission. Three Barthel
– and Mini Mental Status Groups can be combined with the discharge status of those patients who live at
home before admission. A combination of a high Barthel – with a low Mini-Mental-Status-group has the highest
probability to live no longer at home after discharge.
Discussion. The relative frequency of living at home after discharge in different Barthel and Mini Mental Status
subgroups could be compared between different acute hospitals as an indicator of service quality for dementia
patients. A high risk group is identified by a combination of a high Barthel- and a low Mini Mental Status
Examination Group.