As a premier research institution, the University of North Carolina at Chapel Hill connects its researchers’ valuable work
to companies in the private sector. Digital health is one of the university’s top domains, and research in that field is beingtranslated into technology that can be used by startups. “Our goal is to make the process as easy as possible so we can move more quickly to the point where research can be commercialized,” says Bob Dieterle, managing director of UNC’s First in Venture Studio at the Eshelman Institute for Innovation.
“For that to happen, you need a technology stack at the university that can be a pipeline.” Working with Amazon Web Services (AWS) has helped UNC build that pipeline with cloud technology. The institution moved researchers into secure, cloud-based workspaces powered by AWS and streamlined the process for developing and transferring
algorithms and applications to startups. To do that, UNC and its technology partners had to break down traditional silos and rethink the way research is conducted and secured. “The majority of universities aren’t using the latest cloud technologies,” Dieterle says. “It’s about thinking outside the box a bit.”
UNC’s previous status quo
Healthcare relies heavily on research that draws from clinical data while protecting patient privacy. UNC researchers previously used complex desktop computer platforms to analyze data, develop algorithms, and create models. The on-premises platforms lacked other capabilities to accelerate commercialization and
clinical research studies, including analytics, dashboards, and workflow tools. “Projects were taking a lot of time to move forward,” Dieterle says. “Faculty would need a software developer to create a model app to deploy for research. Or if a startup company was licensing the technology, it would take them six to eight months and a lot of money to deploy.”
New research workspaces
Dieterle had a simple message for UNC researchers: Cloud-based technologies
would allow studies to be developed and deployed faster, which would mean more research for commercialization.
UNC reached out to a wide range of cloud providers. Dieterle found that many vendors were siloed — with separate teams for research and commercial offerings, or higher education and business. AWS had a different response. “They said, ‘Give us four months,’” Dieterle says. “Two months later, they had enough consensus to pilot this model.”
UNC’s goal was to move researchers in digital health and other domains into secure, cloud-based research workspaces. Modern application architecture and low/nocode tools created a “drag-and-drop” approach to building workflows that once took expertise in computing languages like Python. “UNC researchers are now able to do more with their lab personnel and don’t have to find specialized skills,” Dieterle says.
The partnership also transformed the institution’s cybersecurity. An AWS consulting partner, CloudHesive, worked with UNC to replace on-premises security tools with cloud-based technology. Cloud security ensures that sensitive patient data for clinical studies complies with the Health Insurance Portability and Accountability Act (HIPAA) and other standards.
To introduce UNC researchers to the cloud platform, Dieterle and his team created “immersion days” during which researchers learned how the technology could accelerate their use of artificial intelligence and machine learning. Following training sessions, the team gave researchers the opportunity to experiment by provisioning them with AWS accounts. Grants of up to $5,000 in credits from the Eshelman Institute, along with monthly office hours from AWS staff and outside consultants, helped researchers “take the momentum of what they learned in hands-on classes and start to port their current research to the cloud,” Dieterle says.
Accelerating the pipeline
Through the work with AWS, UNC’s First In Venture Studio has advanced its role as a catalyst for commercializing the university’s digital health research. The studio has trained about 40 individual researchers and labs. Nine projects have entered the commercialization pipeline, with more in the works, according to Dieterle.
Thanks to the simplified process of building out algorithms, workflows, and visualizations, many research teams are reaching key milestones faster. For data collection, some principal investigators are hiring undergraduates, provisioning them on AWS, and giving them assignments to label data. “What would have taken months is now taking weeks,” Dieterle says.
At the other end of the pipeline, porting data and tools to startups no longer involves thinking through security protocols to make sure patient data remains secure as it leaves university systems. “You don’t have to worry about HIPAA compliance when you have cloud security right out of the box,” Dieterle says.
UNC and AWS are also helping startups take full advantage of the university’s research. For example, one member of Amazon’s cloud team is working in a fractional role with a UNC startup partner that offers an opioid solution for local governments. This will help the startup develop infrastructure to run pilots in multiple municipalities and ultimately commercialize the solution.
Providing technology support to new companies is critical to “de-risking” startups intent on commercializing UNC research, Dieterle says. “Once we launch a startup, we want to increase its odds of success,” he says. “This approach does that.”
Moving ahead
Dieterle says research institutions need experience from the private sector, strong stakeholders within the university, and effective technology partners to break down longstanding silos.
“Over time you’ll see more and more institutions change as the early adopters blow away those that aren’t getting things done,” he says. “All groups benefit more when they don’t work in silos.”
This piece was written and produced by the Center for Digital Education Content Studio, with information and input from AWS.