Transparent Processes and a New Data Classification Enable Greater Integration between the Clinical and Research Missions
The Ohio State University (OSU) is on a journey to modernize its research informatics infrastructure by increasing access to clinical data for researchers by adopting a new data classification, integrating research tools into the electronic health record (EHR) for improved care, and deploying general discovery infrastructure all on a governance process that is reliable and transparent.
RHI: A New Way to Look at Data at OSU
The new data classification – research health information (RHI) – applies to the information collected from a patient’s clinical record and provided to researchers under the Institutional Review Board’s authority. Prior to the establishment of this new data classification, the institution considered clinical research data to be Protected Health Information (PHI) under HIPAA. Following validation by the U.S. Office for Civil Rights (OCR) that a breach of data collected under the purview of a governing IRB and extracted from the EHR as an independent set is not reportable to OCR as long as policy is similarly aligned, OSU established the new RHI classification to align to university protocols for other research data and has leveraged policies and procedures to empower researchers more clearly and transparently.
In creating this distinction and establishing the new RHI classification, OSU experienced a dramatic culture shift and a transformation of its informatics infrastructure. “This is really a risk-mitigation strategy, and it has been a huge culture change because we now have a whole new language to discuss when something is human data or clinical data,” said Timothy Huerta, Ph.D., M.S., Associate Dean of Research Information Technology and Chief Research Information Officer of the College of Medicine and Wexner Medical Center at The Ohio State University. “Until you establish RHI, all your data has to be treated as PHI.”
The shift to what is perceived as a reduced risk level for the Medical Center has come with new opportunities for informatics infrastructure. OSU’s Center for Clinical and Translational Science (CCTS) which is part of NIH’s Center for Advancing Translational Sciences Program, has been engaging with partners from Nationwide Children’s Hospital (NCH), the James Cancer Center, and the Colleges of Optometry, Dentistry and Public Health to collaborate on creating a LifeScale data lake. The resultant data lake is starting with the contents of the Caboodle Data Mart from the OSU Health System, includes linked data from the American Community Survey and the NIH Phenx toolkit on Structural Social Determinants of Health and data from optometry and dental EHRs. Moreover, the dataset will include maternal linkages between mother and baby across organizations to allow analyses that crosses generations. Built in partnership with Microsoft, and leveraging the Azure data platform, the goal of this resource is to support artificial intelligence and machine learning research across both OSU and NCH. This infrastructure will serve as the backbone for the development and testing of in silico approaches to disease and population health management.
Advancing Research in Clinical Decision Support: The Pilot Project
Another breakthrough has been clearly explicating the governance and approval process for engaging research in clinical decision support. The BREASTChoice project, led at OSU by Dr. Clara Lee, is a clinical decision support tool with personalized risk information and solicits patient preferences to support informed choice in breast reconstructive surgery. This project was identified, a priori, as an exemplar that would produce a governance process for research-tool integration; an EHR integration and deployment process for research tools; guidance for researchers seeking to execute research involving EHR integration; and an empirical assessment of the effort required for effective deployment. The result has been clarity around the barriers, boundaries and opportunities for the integration of systems that improve care while at the same time providing an implementation science lens to accelerating adoption.
BREASTChoice was the result of collaboration among the project partners: clinical informatics leadership including the chief medical information officer, clinical IT, research information technology, and researchers at both Ohio State and Washington University in St. Louis.
Engaging Patients as PARTNERS
One piece of OSU’s journey is the PARTNERship for Enhancing Research protocol, which is based on the ORIEN network’s Total Cancer Care protocol. Through PARTNER, the University obtains consent to use both research and clinical data in secondary research, effectively allowing patient data collected in one study to be utilized in other studies through an intermediate honest broker. The CCTS has invested significant resources to create an ecosystem that allows clinical departments to deploy this broad infrastructure to create a holistic research and clinical data repository supported by biospecimen and genomic infrastructure.
Under this decentralized model, data are collected from several OSU departments, which are onboarded and provided a coded limited set of every patient that the department has seen. The result are resources that support more than one researcher or research question. Currently, 14 departments are on a waiting list to use this system.
“The North Star”
Everything the informatics team undertakes is guided by its “North Star,” a strategic plan designed to develop infrastructure capacity for research and to equip researchers with tools and resources to accelerate biomedical discovery.
The plan focuses on four areas: 1) leveraging existing research infrastructure; 2) developing new research infrastructure to advance new areas of research; 3) engaging the research community to utilize that infrastructure effectively; and 4) restructuring research infrastructure to meet the needs of basic, clinical and translational biomedical research more effectively.
“We would not have been able to do this if we did not have a strategic plan that everyone agreed to,” Huerta said. “The research informatics strategic plan was a true partnership across the university and data stakeholders including the CCTS, Wexner Medical Center and the College of Medicine.”