Estimating the burden of dementia and parkinsonism through a novel identification algorithm based on healthcare administrative data

Abstract

Introduction: Neurological disorders (ND), particularly dementia and parkinsonism, are major public health challenges in aging populations. Estimating their prevalence and incidence is essential for healthcare resource planning and targeted interventions. This study aims to estimate the burden of these conditions in the Marche region of Italy, using a novel identification approach applied to administrative healthcare data.

Methods: A cross-sectional study was conducted using administrative databases from the Marche region (2016-2021), including drug prescriptions, hospital discharge records, and chronic condition registries. The TREND protocol was used to enhance case identification. Individuals aged 40 and older were included. Age- and sex-adjusted prevalence and incidence rates were calculated for dementia, parkinsonism, and their co-occurrence. Geographic Information Systems (GIS) were used to analyze spatial distribution.

Results: In 2021, age-adjusted prevalence rates were 7.1‰ for parkinsonism and 31.2‰ for dementia among individuals aged 40 and older, rising to 22.6‰ and 65.8‰, respectively, in those aged 65 and older. Five-year incidence rates were 1.7‰ for parkinsonism and 6.9‰ for dementia. Dementia was more common in women, while parkinsonism predominated in men. GIS revealed higher parkinsonism in southern areas and higher dementia in central and inland areas of Marche. Including antipsychotic and antidepressant prescriptions improved dementia case detection sensitivity.

Discussion: This study demonstrates the value of administrative data and the TREND protocol in improving case identification for neurodegenerative diseases. The observed geographical patterns provide insight for regional healthcare planning in the Marche region. The analysis of antipsychotic and antidepressant use underscores the clinical complexity and healthcare needs of affected individuals. The methodology is scalable and supports reproducible, data-driven strategies for public health policy in aging populations.

Keywords: dementia; healthcare administrative databases; incidence; neurological disorders; parkinsonism; prevalence.

Figure 1

(A) Distribution of prevalence of parkinsonism, dementia and parkinsonism with dementia grouped by age (≥40 years) and gender in Marche Region in year 2021. (B) Trends in the prevalence of Parkinsonism, Dementia, and Parkinsonism with Dementia from 2016 to 2021. Solid lines represent the age-adjusted prevalence rates (per 1,000) based on the WHO 2000–2025 standard, while dashed lines represent the crude prevalence rates (per 1,000). (C) Five-year incidence rates of Parkinsonism, Dementia, and Parkinsonism with Dementia. Dark purple bars represent age-adjusted incidence rates (per 1,000), while light purple bars represent crude incidence rates (per 1,000). The left y-axis represents the age-adjusted prevalence/incidence rates, and the right y-axis corresponds to the crude prevalence/incidence rate.

Figure 2

Histograms showing the annual trends in the number of patients (blue line), ATC prescriptions (pink bars), and exemptions (light blue bars) over the years 2016–2021 for (A) individuals with parkinsonism, (B) dementia and (C) parkinsonism concomitant with dementia. The left y-axis represents the number of patients, while the bars display the corresponding numbers for ATC prescriptions and exemptions. Numeric values above the line represent the total number of patients per year.

Figure 3

Distribution of the ATC-classified pharmacological treatments for (A) parkinsonism and (B) dementia.

Figure 4

Geographical distribution of total cases, prevalence, and incidence of parkinsonism, dementia, and parkinsonism with dementia in the Marche Region, year 2021. All data are per 1,000 inhabitants. Green circles indicate the availability of beds in nursing homes for dementia care, with larger circles representing a greater number of beds.