Premier President and incoming CEO Mike Alkire.
The incoming CEO of group purchasing organization Premier Inc. says the past year has exposed many issues with the U.S. health system, including problems with the supply chain and over-purchasing, but harnessing artificial intelligence and machine learning may offer a solution.
Mike Alkire, who is president of the company uniting 4,100 hospitals and health systems, oversees the performance services business, which combines AI, data and insights to reduce costs and improve healthcare. This experience, combined with an undergraduate degree in computer science and awareness that his industry is increasingly going digital, has given Alkire an AI-focused mindset as he transitions into his new role.
"If you were going to characterize me going forward and some of the areas of focus that I have it's going to be: how do we get a lot closer to using machine learning and artificial intelligence really to inform the way that we're providing healthcare to patients?" Alkire told S&P Global Market Intelligence.
Premier has been seeking out companies with more digital capabilities — including the $117 million acquisition of TheraDoc Inc. in 2014 and the $400 million acquisition of healthcare technology company CECity.com Inc. in 2015 — and Alkire said these deals have allowed the company to better help its healthcare systems improve performance. When COVID-19 came along, the company used its acquired technologies to help solve major prediction issues.
In June 2020, Alkire, along with Premier's Chief Clinical and Innovation Officer Scott Weingarten and Geisinger Neuroscience Institute director Jonathan Slotkin, built a real-time COVID-19 early warning system.
Using an existing platform with access to the electronic health records, or EHRs, of over 200,000 physicians and healthcare providers, Alkire and his team were able to create an automated surveillance app that can integrate with other companies' existing EHRs and track COVID-19 in real time.
Writing in Harvard Business Review at the time, Alkire, Slotkin and Weingarten said they initially assumed that the U.S. Centers for Disease Control and Prevention had the capability to detect COVID-19 hotspots early, before they realized the CDC's system was not automated.
One of the early features of the app was that it could use machine learning and natural language processing to scan EHRs for key COVID-19 symptoms such as "trouble breathing" and then apply this information to predict which patients might have the virus.
"In tests, we have been able to rapidly identify patients who are presenting with signs and symptoms associated with COVID-19 syndrome," the authors wrote. "By subsequently associating these flagged patients with their COVID-19 test results we believe that the app can be trained to be highly sensitive and specific, identifying infected patients with a relatively low rate of false positives and negatives."
While this technology was initially offered for free to health systems, Alkire — who will take over from CEO Susan DeVore when she retires in July after 12 years at the helm — has been in talks with state and government organizations like the Centers for Disease Control and Prevention and the Federal Emergency Management Agency to build out the tool's infrastructure.
January data from Premier indicates personal protective equipment may soon be in short supply again.
Supply chain solutions
Aside from predicting virus hotspots, Alkire said Premier has been utilizing technology to make predictions about potential supply chain issues like shortages of personal protective equipment, or PPE.
As a group purchasing organization, Premier has approximately 2,800 negotiated contracts with manufacturers in the supply chain, which it uses to negotiate discounts for products and services on behalf of its member hospitals.
Building on lessons learned in New York City, Alkire said Premier created an algorithm that can predict which drugs and PPE hospitals would need as the virus continued to spread. In a January blog post, the company identified several supplies at risk of shortages based on past daily usage data and product lead time including sterile water, pipette tips and PPE.
Since the pandemic first began, hospitals and government entities have been stockpiling drugs and PPE, adding strain to an already-stressed U.S. supply chain. Alkire said a more data-driven approach to this stockpiling, however, could benefit hospitals and patients as well as keep costs from being driven up artificially.
"We do think leveraging machine learning and advanced technologies is going to be critical at a national level to address the next pandemic or the next natural disaster where we're going to need that broad-scale information to help inform, for example, health policy folks as well as folks that are responsible for the health of communities," Alkire said.