As a proof-of-concept, we have developed Australia’s first national system dynamics model of chronic disease burden. This is a what-if tool that, when complete, will enable users to simulate policy outcomes before they are implemented in the real world.
The model computes Disability-Adjusted Life Years (DALYs; a summary measure that measures the years of healthy life lost from death and illness), healthcare costs and productivity costs by six key risk factors – tobacco, harmful alcohol consumption, physical inactivity, high BMI, overall dietary risk and high blood pressure.
It requires further development before it can be used as a decision-support for specific policy interventions. When it is finished, it will demonstrate how different risk factors interact with each other, and the impact of a range of interventions on both risk factor prevalence and health and economic burden.
Project co-lead and model architect, Mark Heffernan, said the completed model may be used as a portfolio management tool for policy makers to test how interventions would impact multiple risk factors at the same time, and what the combined effects would be on disease.
“We currently don’t know what the effect on chronic disease is of interventions targeting multiple different risk factors at once. For example, will focusing investments on the reduction of alcohol consumption, or obesity prevalence, or smoking, or increasing physical activity, or some combination of these strategies deliver the greatest reduction in chronic disease burden in Australia? If people drink too much alcohol, smoke and are obese – how will addressing one of these risk factors alter their overall risk of cardiovascular disease?” he said.
“This tool will demonstrate what combination of things we need to do to achieve various reduction targets in chronic disease.”
Policy makers are being invited to experiment with the proof-of-concept model as a demonstration of what will ultimately be possible once the model is complete. The model has an online interface, enabling users to simulate policy experiments and explore their costs and benefits over time.
It demonstrates how dynamic systems models can forecast health burden outcomes such as the prevalence of risk factor behaviour, change in prevalence over time, years of life lost due to risk factor-related disability, and key economic indicators such as labour-related Gross Domestic Product (GDP). Uniquely, dynamic systems models can account for population dynamics, behavioural dynamics, service dynamics, the variation in the intervention impacts over time and the non-additive effects of combining interventions that challenge traditional decision analytic methods.
Forecasting to the year 2051, it indicates there will be a consistent increase in risk factors for chronic disease such as alcohol consumption, high blood pressure, high BMI and physical inactivity. However, the tobacco-related burden will decline due to legislated future price rises, declining tobacco use prevalence and a constant net migration rate.
A number of hypothetical interventions have been modelled, and their impact on DALYs calculated both independently and in combination with other interventions.
The model showed that around 10 years of implementation of any of these prevention interventions is required before any reduction in the growth of DALYs can be seen.
The model has been validated against the Global Burden of Disease Study 2016, using data extracted for 2016 and 2011 published in 2017 by The Lancet.
“We have been able to establish that the dynamics of preventable chronic disease burden can be modelled with some validity,” said project co-lead Dr Geoff McDonnell.
“Prior to this model, only individual risk factors and several outcomes of interest have been successfully modelled using system dynamic techniques in the health policy sphere in Australia.”
The proof-of-concept model was built using a participatory approach, with input from policy partners including ACT Health, Cancer Council Victoria, the Obesity Coalition and the Australian Health Policy Collaboration.
This process has shaped the development of the next phase of the project, which will again undertake a participatory process to develop the model into a customised, validated, robust decision support tool.