Aims: The number of people with dementia is increasing and assessment services need to be expanded to deal with this. There is no consensus on how to calculate the augmentation of services required. We aimed to use linked data from specific geographical regions to calculate dementia incidence and using these locally derived data determine the number of assessments required for these regions.
Methods: We obtained linked data from all people in Western Australia (WA) over the age of 60 years for the period 1989-2014. Administrative datasets including hospital, emergency department, cause-of-death, Aged Care Assessment Program and mental health services. Incidence and prevalence rates of dementia were calculated for four regions of WA based on Local Hospital Networks (LHNs). We calculated the number of assessments required using a diagnostic rate of 63% and that the services would have a penetrance of 29%. Sensitivity analyses were performed for varying penetrance and diagnostic rates.
Results: There were 689,477 people included in this study of whom 51.2% were female. The number of assessments per week required for the four LHNs varied from 7.9 to 14.5. Even assuming 50% penetrance and a diagnostic rate of 40%, the number of assessments varied from 21.3 to 39.3 across the LHNs. The majority of the assessments were for people over the age of 75 years.
Conclusions: We conclude that linked administrative datasets can provide reasonably robust estimates of dementia incidence within relatively small geographically defined areas that can be aligned to health service provider requirements.