An AI model that creates “digital twins” of injured people could help insurers assess the effectiveness of various policy and parameter changes, according to its developer.

University of Melbourne psychiatry associate professor Jason Thompson says the tool can create a “synthetic population” of more than 20,000 injured people and observe the impact of changes in operational models.  

“The idea is that the digital twins act as artificial societies and injured client populations that injury insurance schemes can then test policies and operational models on and see what the impact on system performance might be in the coming periods, rather than testing them in a much higher-risk real world,” he told insuranceNEWS.com.au.  

“They have names, ages, injuries, goals, beliefs, treatment histories, attitudes, costs, selected insurers, levels of health and satisfaction with services, preferred healthcare professionals – anything you might expect in the real world.

“We can then trace how tens of thousands of people move through the system under different scenarios and what the effect of their interaction with different aspects of the system are on overall scheme performance metrics.”  

The tool examines how changes such as increased early intervention and improved access to services affect outcomes in three main areas: financial sustainability, health outcomes and client satisfaction.

Professor Thompson says the model can help strategy managers run through a significant number of parameters and assess which delivers the best results.  

“We can implement policies in the model that cycle through these options very fast ... the result is also not only how different combinations of operational settings affect key performance metrics, but why.

“We can pinpoint where in the model the problems are being generated and check if this is also true of the real system.”

He says his team has designed similar platforms for nearly a decade, attracting funding from the Australian Research Council, National Health and Medical Research Council and Victoria’s Transport Accident Commission.

He says as the tool continues to be developed, it could be used in other insurance areas.