Differences in self and proxy assessments regarding motivation, quality of life and care dependency of people with dementia considering a technology-based intervention

Background: Self-reported assessments of people with dementia (PwD) are sparsely used in research, which may be due to the progressive impairment of cognitive functioning and higher time expenditures for assessments. Proxy assessments by healthcare professionals or relatives are often used instead. The aim was to investigate whether a technology-based intervention significantly influences self and proxy assessments of PwD’s quality of life (QoL), motivation, and care dependency. Method: Data from the RCT of the project AMIGO were used. 32 community-dwelling people with mild to moderate dementia and their relatives received either a humanoid social robot combined with a tablet-based dementia training (IG) or a tablet-based dementia training only (CG). Self and proxy versions of three standardized measurements (AES, DEMQoL, PAS) assessing motivation, QoL, and care dependency were used and analyzed using descriptive statistics and Wilcoxon signed-rank tests. Result: Compared with proxy assessments by relatives, PwD rated themselves better in all outcomes, regardless of group allocation or time points. At baseline, the median differences (MD) differed significantly between self and proxy assessments in IG and CG participants regarding PwDs’ motivation (16.0, p<0.001; 9.0, p<0.001) and care dependency (11.0, p = 0.014; 7.0, p = 0.002), while for QoL this was only the case in the CG (11.0, p = 0.035) and not in the IG (5.0, p = 0.139). After intervention, all outcomes in the IG and CG were significantly differently rated by PwD and relatives demonstrated by the MD for motivation (8.0, p<0.001; 12.0, p = 0.0101), QoL (16.0, p = 0.0023; 16.5, p = 0.030) and care dependency (12.0, p = 0.006; 4.5, p = 0.019). Conclusion: There are significant differences of assessments between relatives and PwD, which could be due to a reduced illness insight of PwD, different foci of PwD and relatives on scale domains (e.g., psychosocial, physical), or an influence of relatives' attitudes and burden. However, QoL in the IG was not perceived differently at baseline, but after the intervention. This finding as well as the significant increase of self-reported QoL, but the non-significant proxy evaluation in the main study, highlight the PwDs’ ability to assess themselves and the necessity to involve them in research evaluations.