University of Iowa Hospitals and Clinics log more than a million patient visits each year, collecting a trove of information about health problems and treatments. But accessing this data for research purposes isn’t necessarily easy. 

A partnership between ITS Research Services, the Institute for Clinical and Translational Sciences (ICTS), and Health Care Information Systems is addressing one facet of the challenge. 

The Iowa Health Data Resource Data Enclave went live in July 2022. It connects Iowa’s existing big-data tools to a new clinical data warehouse. 

It’s part of a larger initiative to help researchers tap data from patient records securely and effectively, attract more research funding, and make Iowa a national model. 

A piece of a bigger puzzle 

“Accessing clinical data for research is a big issue for every academic medical center,” says Boyd Knosp, associate director for biomedical informatics operations at the ICTS and associate dean for IT in the Carver College of Medicine. He’s interviewed dozens of colleagues at institutions around the country to identify specific challenges.  

Knosp’s findings shaped a winning pitch for P3 funds to build the Iowa Health Data Resource (IHDR), a campus-wide initiative with especially strong support from Iowa’s health colleges. 

Creating a secure data enclave was one of three main goals for the project. Others included developing transformative data sets and building a cadre of faculty and staff “data liaisons” who can help colleagues design data-centric projects. 

“Clinical data isn’t collected for research—that’s a secondary use,” Knosp says. “We realized we needed clinicians and health scientists who understand data and can help bridge the gaps.” 

IHDR health data liaisons receive partial salary support from P3 funds, plus training on data collection and study design. 

Accessing big-data tools in the Data Enclave 

The new Data Enclave offers a previously unavailable connection between protected health datasets and UI computational resources like the Argon high-performance computing cluster and the Interactive Data Analytics Service (IDAS). 

It provides a secure storage system where data extracted from health records can be stored and processed, but not downloaded. The enclave’s architecture lets Argon and IDAS access approved data extracts for studies involving machine learning and predictive analytics. 

This connection helps College of Nursing investigators, for example, assess symptom variability advanced-cancer cases and classify safety criteria for patient mobility. It lets College of Pharmacy researchers use machine learning to detect eyelid tumors from clinical photographs, or a Carver College of Medicine team perform retrospective modeling of neonatal intensive care. 

Research teams work with their IHDR health data liaisons to define data needs. The resulting datasets are securely transferred to the Data Enclave. 

“On the technical side, our storage specialists John Saxton and Brendel Krueger were able to recycle architecture and support from our existing Large Scale Storage service,” says Genevieve Johnson, self-service integration specialist for ITS Research Services. “It was a cost-effective, time-effective solution.” 

Collaborating to use existing tools is a win for everyone, especially investigators.  

“This collaboration is about more than technology,” says Joe Hetrick, director of ITS Research Services. “It’s also about the expertise, processes, and training the university has intentionally and consistently invested in.” 

Mike Frangi, ITS project manager, and Heath Davis, IHDR assistant director, played central roles in the project. Additional partners included Health Care Information Systems and the ITS Project Management Office and Enterprise Infrastructure team.  

More data, more opportunities 

Knosp expects the IHDR and services like the Data Enclave to keep expanding resources available to research teams. 

Creating transformative datasets—for example, collecting a decade of mother and child health data, extracting nursing-intervention data, or cross-referencing patients who’ve received both dental and medical care at the university—is helping to grow the IHDR and laying groundwork for future data-driven research programs. 

Eventually, these and other resources might be able to tap non-UI data or include information that’s not yet part of typical patient record. 

“How might we partner with other health systems around the state to leverage their data for research?” Knosp asks. “Or can we link social determinants of health from publicly available databases to our patient record? The more we consider the possibilities, the more questions we can start to answer.”