Initial findings from patient care data project released
Sept. 25, 2017 — Students and researchers at four universities and Sanford Research have released the findings from their study of real-time patient data.
The data, which is stripped of private information, was made available as part of the Sanford Data Collaborative, which gathered statistics from thousands of patient visits to Sanford facilities and allowed the academic institutions to study it for trends and insights. The study began in February.
“Sharing data and collaborating with regional universities like this is unprecedented,” said Emily Griese, Ph.D., a director at Sanford Research. “But that’s how we’ll move forward as a health care leader – by looking for unique solutions to patient care challenges.”
Sharing de-identified data with other institutions brings broader perspective to issues faced by patients, providers and systems. Sanford Research recognizes that collaboration can bring about solutions while still keeping patient privacy at the forefront.
The projects:
- University of South Dakota (principal investigator: Carole South-Winter, Ed.D.): The team developed a readmission risk algorithm for patients following heart surgery that determines who is at risk and provides insights for care. Previous risk scores did not suggest possible interventions.
- Dakota State University (principal investigator: Yong Wang, Ph.D.): Researchers looked for patterns in how rural and urban patients use various service platforms, including electronic medical records, to search for ways to decrease emergent and urgent care needs.
- University of North Dakota, (principal investigator: Arielle Seyla, Ph.D.): The team developed an algorithm to predict unplanned medical visits for diabetics, taking into account their current disease management behaviors, such as smoking, and other information, and then providing pathways to care.
- South Dakota State University (principal investigator: Surachat Ngorsuraches, Ph.D.): The team developed a patient engagement score using existing patient data. Patient engagement factors into effective management of chronic conditions, but surveys and other tracking methods are time-consuming. This score can help identify and decrease emergency department visits and hospitalizations.
- University of North Dakota, (principal investigator: Jeff Hostetter, M.D.): The team examined how primary care services can affect patients’ use of preventative behaviors and looked to see how that differs with a team-based approach.
- Sanford Research (principal investigator: Susan Hoover, M.D., Ph.D.): The Population Health Group created an algorithm based on current patient data to determine who needs screening for C. difficile. The goal was to decrease unnecessary testing and to develop a platform to be used to decide on ordering the test.
All projects are undergoing validation. Those with clear promise after validation will be implemented this fall. To ensure continued patient privacy, a privacy board was developed and contains legal and Health Insurance Portability and Accountability Act experts and community members.
“The possibilities using big data are endless,” said Allison Suttle, M.D., senior vice president and chief medical officer for health services. “The ability to analyze it in real time, with the expertise of our regional universities, means we’ll better be able to answer key questions and provide more individualized care. We’re able to take something abstract – data – and make it personal, and that’s exciting.”
The Sanford Data Collaborative is a service of Sanford Research and is supported by Sanford Enterprise Data and Analytics. The 2017-18 project RFP will be available in September 2017.
