Model behaviour: A systems approach to alcohol-related harm

Project title: Simulation modelling of alcohol consumption and the effectiveness of harm-reduction policies

Project start date: April 2014

Project end date: June 2016

This project is finished. Click on the image to read the Findings Brief.
What is dynamic simulation modelling? Watch an explainer video here.

What is the issue?

Excessive alcohol consumption causes harm across society, from the effects of binge drinking, including alcohol poisoning, violence and road traffic accidents, to long-term high overall alcohol consumption which contributes to chronic health problems, including heart disease, cancer and stroke.

There are many ways to reduce excessive alcohol consumption, including increasing taxes on alcohol sales, restricting hours of sales, drink driving regulation, and education and social marketing.

However, alcohol-related harm is a complex problem, with many inter-related causes, so we needed to explore what combination of interventions is likely to produce maximum community-wide impact.

How did the project address the issue?

The Prevention Centre developed a simulation model of alcohol use in NSW that could be used to forecast the effectiveness of a variety of approaches to reducing alcohol-related harm, both individually and in combination. The model addressed both binge drinking and high average consumption that leads to chronic disease.

Simulation modelling brings together a variety of evidence sources, such as research, expert knowledge, practice experience and data, to capture the complexity of a problem. That model is then used to simulate various policy scenarios to see which is likely to have the most effect.

The key steps of the project were to:

  1. Compile research on the causes of the problem and the intervention options.
  2. Bring together a small group of policy makers, researchers, practitioners and other to collaboratively build the model, drawing on research evidence, their expert knowledge and data.
  3. Test and validate the model.
  4. Examine the effectiveness of alternative policy options to inform policy design and facilitate broader consensus for action to address the problem.

What were the outcomes?

The project produced a decision support tool that allows policy makers and other stakeholders to explore the impact of a broad range of potential intervention scenarios over time.

Other outcomes of this project were:

  • It collaboratively brought together many valuable sources of information in a systematic way to produce a practical tool that policy makers can use to guide their decisions on how best to address the complex problem of alcohol-related harms.
  • The process enhanced appreciation of the value of modelling for informing policy and practice.

Relevance for practice

Excessive alcohol consumption is a cause of chronic health problems and Prevention Centre partners have confirmed that alcohol policy is of interest.

The model explored which combinations of interventions will have the greatest impact for dollars spent when addressing both binge drinking and long-term high average alcohol consumption, informing policy decisions about alcohol-related harm.




Project lead

Project team

This project was funded by the NHMRC, Australian Government Department of Health, NSW Ministry of Health, ACT Health and the HCF Research Foundation.


NHMRC, Australian Government Department of Health, NSW Government Health, HCF Research Foundation, ACT Government Health, hosted by Sax Institute

This project has now been completed.The model simulated a large range of possible combinations of interventions. Several findings are of particular interest:

  • Combining different interventions has unanticipated, synergistic effects. For example, combining a policy of 3 am licensed venue closing time plus 1 am lockout with an expansion of treatment service coverage to 20% of heavy drinkers suggested a 33% reduction in acute alcohol-related harms. This population impact was greater than the sum of each policy modelled individually. This understanding will help policy makers to develop suites of interventions within the constraints of limited government resources.
  • Some interventions can have unintended consequences. For example, when we tested the impact of raising drink prices in licensed venues, the model unexpectedly showed an increase in consumption and violence. This is because individuals in the model shifted their behaviour to drinking more at home parties or ‘preloading’ due to the higher prices in pubs and clubs.
  • With sustained investment, the effect of interventions can grow stronger over time – though the full impacts may not be seen within a policy cycle.


  • Atkinson JA, Prodan A, Livingston M, Knowles D, O’Donnell, Room R, Indig D, Page A, McDonnell G, Wiggers J. Impacts of licensed premises trading hour policies on alcohol-related harms. Addiction. 2018; doi: 10.1111/add.14178. [Epub ahead of print]


  • Atkinson J and Page A. Dynamic simulation modelling to inform strategic planning for prevention of alcohol-related harms and suicide. Intelligent Decision Support Systems in Healthcare symposium. 21 June 2016. Sydney
  • O’Donnell E. Dynamic simulation modelling: Supporting decisions to reduce alcohol-related harms in NSW.  Emerging Health Policy Research Conference. Menzies Centre for Health Policy. Sydney. July 2016.
  • Knowles D, Prodan A, McDonnell G, Atkinson A. Developing decision support tools in partnership with State Health Departments to inform prevention policy. AnyLogic Conference. 16 November 2016. Nashville, US.