U of A launches safe data-sharing and analytics platform to boost AI, machine learning and digital health research

DARC helps researchers analyze larger amounts of health data faster and securely.

Laura Vega - 7 July 2021

The University of Alberta has launched a new health-data management and analysis platform to increase capacity in research areas such as big data, precision health and machine learning.

The Data Analytics Research Core (DARC) was developed in the U of A Faculty of Medicine & Dentistry (FoMD) to provide high-performance computing and sensitive data storage in a secure environment, compliant with health information privacy regulations and requirements from the provincial health system to share this information. Through this integrated platform, researchers can grow their capacity for big data analysis and artificial intelligence, e-health innovation and training for graduate students in health-data analytics.

In the past, researchers have had to look for their own options to store and use health data, which had to be vetted for compliance by the information custodian—in most cases, Alberta Health Services. With DARC, they will be able to access a secure environment specifically developed to meet the needs of the health information custodian for data sharing and pave the way for an Information Management Agreement.

“That was a key driver of this initiative. If we could create an environment that helped to facilitate data sharing because key requirements in health data research computing were being met consistently, the custodian's ability to share more data would improve,” says project lead Lawrence Richer, vice-dean of research and Alberta Health Services Chair in Health Informatics Research.

The other issue addressed by DARC, explains Richer, is that artificial intelligence and machine-learning projects require a much larger computing capacity than individual desktops and small servers offer. With approximately 40 users currently registered to use the system, Richer and the DARC steering committee working to develop this platform have seen the success of its larger capacity in real research projects.

“The early evidence is that it’s working,” he adds. “We had a scenario very recently where Alberta Health Services was able to transfer terabytes of diagnostic imaging data to the DARC environment to allow for research computing while preserving privacy. That was something they had never done before.”

To consolidate platforms to manage data, DARC features SAS Viya, a scalable platform for statistics, data visualization, predictive modelling and machine learning. SAS Viya can expand capabilities for these analytics and train learners so they are familiar with the latest version of this platform, which is already widely used in Alberta’s health system.

For Richer and his team, this initiative could help propel research beyond the FoMD, generate new ideas for projects that would have been too difficult without the right infrastructure, and eventually improve health care.

“I think the potential to benefit all researchers is there because we can do research computing on any data,” says Richer. “The focus has been health information, data that is governed by the Health Information Act. But when you think about that data, it’s artificial intelligence and diagnostic imaging, risk prediction, machine learning, traditional health outcomes, clinical trials. It’s all those areas.”

“Let's say you're doing a genetics research project on the genomics of a certain population with a certain disease. You have 100 participants and they all consent to using not only their genomic information, but information from their health record. Now you'd have a place to merge that data. So now you're mixing translational research with more traditional health outcomes, and you have a safe platform to bring different data sources together.”

DARC is available for U of A investigators and research groups across campus. Find more information and requirements for enrolment here.

DARC is supported by the Faculty of Medicine & Dentistry along with funding from the Alberta Health Services Chair in Health Informatics Research.