Abstract

Objectives

To estimate the excess formal social costs or direct non-healthcare costs of dementia-related neuropsychiatric symptoms (NPS).

Methods

The presence of dementia, NPS, antipsychotic and antidepressant use, somatic and psychiatric comorbidities, and formal social benefits were studied in a regionwide cohort of all 60-year-old and older individuals. A random forest-based algorithm identified NPS, and 2-part regression models and entropy balance were used.

Results

Of the 215 859 individuals, 7553 (3.50%) had dementia, 74 845 (34.7%) had some NPS, and 20 787 (9.63%) received long-term care benefits. Notably, nearly two-thirds (63.9%) of people with dementia received benefits. The probability of having social costs varied markedly with age (odds ratio [OR] 12.28 [10.17–14.82] for >90-year-olds category), and the presence of dementia (OR 7.36 [6.13–8.84]) or NPS (OR 3.23 [2.69–3.88]). NPS (relative change [RC] 1.39 [1.31–1.49]) and dementia (RC 1.32 [1.24–1.41]) were associated with higher average benefit costs. Low socioeconomic status was significantly associated with both a higher probability of receiving benefits (OR 1.52 [1.38–1.68]) and higher costs of their provision (RC 1.18 [1.15–1.21]).

Conclusions

The burden of caring for NPS is greater than that indicated by the literature as these symptoms multiply the social costs of dementia by more than 3, owing to the greater use of residential care and formal coverage reaching more patients than that indicated by the literature. The greater presence of dementia and NPS in the population of lower socioeconomic status indicates an inequality in health attenuated by greater use of social benefits.