Forecasting risk-factor distributions by sex and age, 1990-2050: a parametric estimation framework using GBD summary exposure values, demonstrated for Australia.
- 1Population Interventions Unit, The University of Melbourne, Melbourne, VIC, Australia.
Abstract
Background
This paper presents a method to convert existing Global Burden of Disease (GBD) forecasts of summary exposure values (SEVs) into business-as-usual forecasts of risk-factor distributions.Method
SEVs from the GBD are a relative-risk-weighted summary of risk-factor exposure, ranging from 0 (everyone below the theoretical minimum risk exposure level) to 1 (everyone above the 99th percentile of global exposure). Given the known inputs to calculate SEVs, with one additional assumption of the coefficient of variation of the risk factor, it is possible to generate parametric distributions (e.g. log-normal, gamma) of risk-factor exposures. We demonstrate this method by forecasting SEVs to 2050 by sex, age, and year for body mass index (BMI), systolic blood pressure (SBP), low-density lipoprotein (LDL) cholesterol, and fasting plasma glucose in Australia.Results
Our method generates plausible age- and sex-specific risk-factor distribution forecasts for Australia. Compared with the 2017/18 National Health Survey, fasting plasma glucose and LDL cholesterol agreed reasonably well in younger age groups, with wider discrepancies at older ages. SEV-derived distributions overestimated the mean SBP by ∼3-10 mmHg and mean BMI by ∼1 kg/m2 relative to survey estimates. The method was sensitive to values of the 99th global percentile of risk factors.Conclusion
This method is extensible to any country or demographic group. To facilitate broader use of this method, we recommend further cross-validation studies for other countries and making key GBD data (e.g. 95th percentiles of risk factors) more publicly available.
Related
- Athlete-AI: a multi-modal artificial intelligence framework for personalized athlete monitoring and injury risk forecasting
- TRIDENT-Risk: An AI-driven temporal ensemble framework for proactive high-severity accident risk forecasting in Intelligent Transportation Systems
- Cardiac allograft vasculopathy demonstrated by coronary angiography and histopathologic correlates
- Pulmonary sclerosing pneumocytoma demonstrated heterogeneous tumor growth rate on serial CT: a retrospective cohort study
- Explainable machine learning framework for dynamic line rating forecasting with adversarial robustness evaluation
BibTeX
@article{dhungel2026forecasting,
title = {{Forecasting risk-factor distributions by sex and age, 1990-2050: a parametric estimation framework using GBD summary exposure values, demonstrated for Australia.}},
author = {Bibha Dhungel and Tim Wilson and Stephanie Khuu and Tony Blakely},
journal = {International journal of epidemiology},
year = {2026},
month = oct,
doi = {10.1093/ije/dyag223},
url = {https://europepmc.org/article/MED/42900799},
note = {CC BY-NC},
}
RIS
TY - JOUR
TI - Forecasting risk-factor distributions by sex and age, 1990-2050: a parametric estimation framework using GBD summary exposure values, demonstrated for Australia.
AU - Bibha Dhungel
AU - Tim Wilson
AU - Stephanie Khuu
AU - Tony Blakely
PY - 2026
DA - 2026/10/01
JO - International journal of epidemiology
DO - 10.1093/ije/dyag223
UR - https://europepmc.org/article/MED/42900799
AB - Background This paper presents a method to convert existing Global Burden of Disease (GBD) forecasts of summary exposure values (SEVs) into business-as-usual forecasts of risk-factor distributions.Method SEVs from the GBD are a relative-risk-weighted summary of risk-factor exposure, ranging from 0 (everyone below the theoretical minimum risk exposure level) to 1 (everyone above the 99th percentile of global exposure). Given the known inputs to calculate SEVs, with one additional assumption of the coefficient of variation of the risk factor, it is possible to generate parametric distributions (e.g. log-normal, gamma) of risk-factor exposures. We demonstrate this method by forecasting SEVs to 2050 by sex, age, and year for body mass index (BMI), systolic blood pressure (SBP), low-density lipoprotein (LDL) cholesterol, and fasting plasma glucose in Australia.Results Our method generates plausible age- and sex-specific risk-factor distribution forecasts for Australia. Compared with the 2017/18 National Health Survey, fasting plasma glucose and LDL cholesterol agreed reasonably well in younger age groups, with wider discrepancies at older ages. SEV-derived distributions overestimated the mean SBP by ∼3-10 mmHg and mean BMI by ∼1 kg/m2 relative to survey estimates. The method was sensitive to values of the 99th global percentile of risk factors.Conclusion This method is extensible to any country or demographic group. To facilitate broader use of this method, we recommend further cross-validation studies for other countries and making key GBD data (e.g. 95th percentiles of risk factors) more publicly available.
SN - 0300-5771
SN - 1464-3685
ER -