CHI In Silico Health Policy Experiments

From HPSIGWiki

Changing healthcare policy in silico

Geoff McDonnell is a firm believer that changes to the financing and delivery of healthcare, be they big ones like modifications to Medicare, or a smaller one such as changing a hospital admissions policy, should be computer simulated before they’re launched in the real world.

“Modelling the impacts of proposed changes to healthcare before they become policy is smarter, safer, cheaper and quicker,” says Dr McDonnell, an MIT and Harvard-trained physician and engineer.

“The inherent complexity of healthcare systems means that foreseeing the full impact of changes to policy and practice is nearly impossible.”

A Simulation Research Fellow at the Centre for Health Informatics, Dr McDonnell uses modelling software to understand and improve the interaction between structure and action in healthcare and policy. He has successfully used dynamic systems modelling in commercial engineering and healthcare projects in Australia.

“Real world experiments can be extremely expensive and irreversible – often for decades – because they often require significant capital investment and workforce realignment.

“The most important positive changes we can make in dynamic systems like healthcare are often counter-intuitive,” Dr McDonnell says.

“For example, hospitals often try to absorb the spike in demand for acute care hospital beds that occurs in wintertime by increasing the number of acute care beds in the Emergency Department. That’s an intuitive quick fix approach to the problem.

In fact, widening access to a hospital’s front door only slows down the system and exacerbates the initial problem. Dynamic modelling in the UK and Australia has shown that the best solution to this kind of problem is the non-intuitive one of increasing the hospital’s ‘downstream’ post-acute capacity.

“My dream is that in future, we’ll resist the temptation to make changes in healthcare policy until we've modelled their impacts ahead of time.”