GE Healthcare: Corvix

What can others learn from the successes and failures of the way you’ve unlocked the value of big data?: 

**MadPow created three Corvix applications** Corvix India: Explore public health statistics at the Country and State level, place hospitals in Districts within a State with Cardiovascular Disease capacity, group these new hospitals in multiple scenarios, model the health and financial outcome over time, compare the outcome of the scenarios. Corvix Saudi Arabia: Explore 10 years of over two dozen public health statistics at the Country and Region level, view and filter existing hospitals by specialty. Corvix India Oncology: Explore statistics at the Country, State and District level, view and filter oncology facilities by type and existing oncology screening and treatment equipment, view potential distribution of FDG (fluorodeoxyglucose) to support radiology screening and treatment. **The responses below combine lessons learned from all three projects.** The combination of Maps, Models and Games can be a very powerful tool for directing healthcare resources. With data on maps, two aspects of data can be very powerful. The first is having the same data at progressive levels of granularity: Country, State/Region, District/County. The second is having this same data across periods of time. Progressive granularity allows the application to support continuous zoom of the data through the map. A uniform time series allow the application to reveal chronological changes in geographic context.

What were your expectations of the value hidden in your data, and how did they influence the design of your solution?: 

We use progressive disclosure to support exploration. Our audience is concerned about a country/state/region, so a map-based visualization of statistical data makes the comparison of statistics easy to read. The use of two dimensions — color and size — draws the user’s attention to specific areas. Direct interaction with symbols in those areas — markers for transportation and healthcare facilities — reveals relevant data in the appropriate context.

Conversely, how did the design of your solution affect your understanding of the potential value of your data?: 

We wanted the exploration process to make specific targets stand out: an underserved district, a city with appropriate needs and resources and a lack of existing facilities. It is up to the user to define those characteristics, so we designed the application to expose these differences.

Describe the aspects of the design of your solution that do the most to expose meaning in data that would otherwise be harder to discern.: 

It was a challenge to display the relationship in a District between health statistics, the size of the cities and transportation/logistics data to support radiology screening and treatment. These sets of data are independent but when viewed together reveal important potential for impacting healthcare improvement.

How might your solution be extended or adapted to address additional types of data and other questions?: 

**The Corvix model can be extended in several directions:** Maps: The same system can be used to visualize public health statistics for any geographic region. A richer set of data objects can be developed to include States, Districts, Cities, Hospitals, Transportation hubs, etc. Configuring Hospitals: The current work has been limited to a configuration model for a single form of specialty care. Additional forms of specialty care can be added. Models: The current system demonstrates how agent-based modeling simulation (ABMS) can be used to predict impact. Additional work need to be done testing how public health statistics can be used as input to ABMS to produce meaningful predictive models. Gamification: Our initial work shows how the comparison of multiple scenarios can aid exploration. Additional work needs to be done to integrate goals into the modeling process, promoting game mechanics to support problem solving.