Regional Population Ageing, Physician Age Structure, and Physician Supply Data Across Five OECD Countries

Description

This dataset contains harmonized longitudinal regional data on population ageing, physician age structure, physician supply, population density, and economic context across Australia, Canada, France, Hungary, and Japan. Regional observations were harmonized at OECD Territorial Level 2 (TL2). The main analytical panel contains 894 region-year observations from 57 TL2 regions. Country-specific observation periods are Australia (2013–2024), Canada (2001–2024), France (2012–2025), Hungary (2008–2024), and Japan (2006–2024). Observation years are not necessarily continuous for every country. Population ageing is represented by the percentage of the population aged 65 years and older. Physician workforce ageing is represented by the percentage of active physicians aged 55 years and older, constructed from the percentages aged 55–64 years and 65 years and older. The dataset additionally contains physician density, population density, GDP per capita, region-specific temporal slopes, standardized ageing-trajectory measures, trajectory classifications, and population–physician ageing-mismatch measures. Region-specific longitudinal trajectories classify 35 of the 57 regions as showing simultaneous population and physician-workforce ageing and 22 as showing population ageing alongside physician-workforce rejuvenation. The supplied analytical files also permit reproduction of models examining associations between population–physician ageing mismatch and regional physician-density trajectories. The repository provides harmonized and derived data in CSV and RDS formats, analytical tables and figures, R scripts, reproducibility metadata, and a data dictionary. The original OECD source CSV files are not redistributed. Source observations can be obtained from the OECD Data Explorer using the OECD dataflows documented in README.txt. The files can be used for reproducibility, secondary analysis, alternative model specifications, methodological extensions, and comparative subnational research on demographic change and health workforce dynamics.

Steps to reproduce

Regional source data were obtained from OECD Regional Statistics through the OECD Data Explorer and harmonized at OECD Territorial Level 2 (TL2). Four source datasets were used: (1) Ageing healthcare workforce – Regions (DSD_REG_HEALTH@DF_CARE_AGEING), (2) Population by broad age groups – Regions (DSD_REG_DEMO@DF_POP_BROAD), (3) Population density – Regions (DSD_REG_DEMO@DF_DENSITY), and (4) Gross domestic product – Regions (DSD_REG_ECO@DF_GDP). The analysis includes Australia, Canada, France, Hungary, and Japan. Compatible observations were filtered and harmonized using country, TL2-region, and year identifiers. Population ageing was defined as the percentage of the population aged 65 years and older. Physician workforce ageing was calculated as the sum of the percentages of active physicians aged 55–64 years and aged 65 years and older. Physician density was measured as active physicians per 1,000 inhabitants. Population density and regional GDP per capita were retained as contextual variables. No imputation or interpolation was applied to missing physician-ageing or population-ageing observations used to estimate regional trajectories. Region-specific ordinary least-squares regressions against time were used to estimate temporal slopes for population ageing, physician workforce ageing, and physician density. Population- and physician-ageing slopes were standardized within countries to construct the primary ageing-mismatch measure as the standardized population-ageing slope minus the standardized physician-ageing slope. A pooled-standardized alternative was also constructed. Data processing and analyses were performed in R 4.5.2. The reproducible workflow is documented in 04_scripts. Scripts 01_import_and_clean.R through 08_supplementary_tables.R cover data import and harmonization, trajectory construction, primary models, robustness diagnostics, structural-break sensitivity analyses, tables, figures, and supplementary outputs. The master script 00_run_all.R executes the complete sequence. The original OECD source CSV files are not redistributed in this public repository. To reproduce the workflow from the raw-data stage, users should obtain the corresponding source observations from the OECD Data Explorer and place them in a local 01_raw_data directory using the filenames File1.csv, File2.csv, File3.csv, and File4.csv expected by the scripts. Harmonized and derived datasets are provided in 02_clean_data for reuse and reproduction of subsequent analytical stages. Package versions and complete R session information are provided in 05_reproducibility.

Institutions

• Karabük University

  Karabük Province, Karabük

Categories

Licence