Predictive analytics – the use of historical data to look for patterns and identify trends in order to anticipate future needs and prevent potential problems – is becoming increasingly more common in healthcare today. So much so, that the City of Pittsburgh, the University of Pittsburgh Graduate School of Public Health and Intermedix have partnered with each other to gain a better understanding of how predictive analytics can be used to respond and prepare for disasters in the city.
During a recent emergency preparedness resilience workshop held in Pittsburgh, the city came up with a scenario similar to that of the Donora Smog that killed 20 people and sickened 7,000 in the small town of Donora, Pennsylvania in 1948. Representatives from the city, the University, and Intermedix wanted to know if something happened in modern day Pittsburgh on a larger scale how they would be able to respond to it.
The team, which included city and government leaders, technology and utility companies and first responders, used a simulator created by the University. The simulator, FRED was originally designed to predict the dynamics of infectious disease epidemics and the interacting effects of mitigation strategies, viral evolution and personal health behavior. It has since expanded to include many non-infectious diseases, as well as social and environmental factors that affect health.
By taking the data from the FRED simulator and passing it to Optima Predict, the team was able to understand how many responders are needed for the given scenario, how long it would take them to respond, and when ER departments start to saturate in terms of volume. Armed with this information, Justin Schaper, SVP of Analytics at Intermedix says, “we can try to understand what the effects would be in a disaster and do what-if planning to mitigate those effects. WebEOC allows us to manage and integrate communications through the exercise and Optima helps us understand how to get first responders to and from calls to ensure the best response for the community.”
Justin adds, “we saw the effect on agencies in an emergency situation is non-linear; its not simply a matter of increasing the number of calls coming in – its more complicated than that and analytics around that does require a simulator to be able to truly understand how something as simple as closing a road may have a surprising positive effect on emergency response. In the context of this scenario, having this kind of planning is critical.”
While the workshop itself was part of a broader initiative called ONEPGH, a partnership of the 100 Resilient Cities project pioneered by the Rockefeller Foundation, the software was able to clearly show the impacts to first responders in a disaster scenario with the use of predictive analytics. As Justin says, “close coordination between various stakeholders in this kind of situation is critical. Having clear communication paths and an early understanding about the magnitude of the event communicated to hospitals, utility companies, school systems, responders – the need is very evident.”