KEYNOTE: Predictive Modelling in Healthcare; Promise and Peril
Prior to the inception of COVID-19, healthcare budgets in most countries appeared to be out of control, with the U.S. healthcare spending heading to 20% of GDP, and other countries not far behind in terms of rapidity of increases. COVID-19 has caused healthcare authorities, payers and providers to temporarily focus on fighting the pandemic, but once COVID-19 is under control budget pressures will return, greater than ever. Recently, predictive analytics, big data and artificial intelligence have been proposed as solutions that will change healthcare for the better. Certainly, analytics, data science and artificial intelligence are playing a significant role in process automation and improving diagnostic accuracy. Models enable practitioners to identify high risk populations and conditions earlier and to intervene more effectively with patients. But is the optimism for big data and AI in healthcare justified or is it simply hype? Why, if models are so much better, are we not seeing a bending of the cost-curve in healthcare? What will it take for predictive analytics and AI to make a significant impact on the cost and value of healthcare?
We propose three factors that are required to work together to effect transformation: Payment Reform; Predictive Analytics and Behavioral Economics. In the future, more outcomes risk will be transferred to providers and consumers of healthcare services. As risk professionals, actuaries are positioned to be a significant contributor to this transformation.