SPSS Analytics Partner | Case Study – Discovery Health

Case Study – Discovery Health

Johannesburg, South Africa

Solution components

  • IBM® SPSS Statistics
  • IBM SPSS Modeler
  • IBM SPSS Text Analytics for Surveys
  • IBM PureData™ System for Analytics, powered by IBM Netezza®

Discovery Health


Predictive analytics used to craft preventive programs that keep members healthier and costs lower


Founded in Johannesburg in 1992, Discovery Health is South Africa’s largest healthcare insurance provider, managing 14 medical plans, including the Discovery Health Medical Scheme, South Africa’s largest, covering more than 2.6 million lives. Discovery employs more than 5,000 people.


The Opportunity

For Discovery Health to maintain its cost advantage over competing health insurance providers, it needs to constantly refine the clinical risk management capabilities that are the source of its strength. That means tapping new sources of data and applying innovative models to pinpoint where the health risks are and how best to manage them proactively.


What Makes It Smarter

Discovery’s new predictive risk management solution extracts deep and accurate insights from clinical, demographic, billing and even unstructured member data to point out chronic health risk patterns within its member base and provides planners with the granular, analytical guidance to develop the most effective preventive programs.


Real Business Results

By improving the predictive accuracy of its clinical risk models, Discovery can better target preventive programs to at-risk subscribers, thus reducing the overall cost of care and helping minimize premiums. The solution reduced the time required to run predictive analytics models by more than 99 percent, from nearly a day to minutes, yielding results faster and giving risk analysts the flexibility to develop and test new models that further improve predictive accuracy. The solution has also enabled administrators to identify and recover more than USD25 million resulting from a combination of fraudulent claims and billing errors.

“We can now predict which of our members are most likely to require procedures and adjust our health plans to serve them more effectively and offer better value.”
— Matthew Zylstra, actuary, Risk Intelligence Technical Development

© Copyright IBM Corporation 2014