The Challenge: Difficulty accessing research data
Researchers in the medical community are evolving to work more collaboratively across institutions to share data and improve research outcomes. A healthcare institute, based in St. Louis, provides a research data repository that integrates electronic health records from several hospitals and medical research facilities in the metro area. Using an open source data model (OMOP), the repository is a data warehouse that enables collaboration with other universities and medical institutions, but that data has to be accessible. Leveraging the repository, researchers from across organizations can perform cohort searches, hypothesis testing, correlative studies, and retrospective analysis of defined patient populations. Researchers use this data to inform research projects and request grants.
However, with the current data warehouse, the process to extract that data was slow and painstaking. The institution needed a way to improve the user experience, resolve issues with the system, and ensure it is sustainable, easy to maintain, and extensible for future needs.
The Solution: A more sustainable, user-friendly data warehouse
The research institute partnered with 1904labs to help revamp the repository by enhancing the design of the underlying data warehouse, improving scalability and ease of use.
We rearchitected the data warehouse to create a platform that can be easily extended and maintained. Leveraging technologies that the institute already has and can support in house, the new system is consolidated and reorganized to increase productivity and maintainability. We helped them implement industry best practices for software code versioning control and releases to increase the efficiency of their development team.
Our Process: Human-Centered Design Agile
Through our Human-Centered Design Agile (HCDAgile) process, we collaborated with the institution’s teams and discovered that the current system was not extensible and had to be rearchitected to achieve the functionality and supportability that the organization needed.
We worked with key stakeholders including database administrators, developers, and informaticists to understand and address the pain points for those using the system. We summarized our research in personas, journey maps, story maps, and other artifacts we could document and share what we learned. Throughout the project, we worked with these users frequently and continuously to engage them in the solution, building internal collaboration and buy in, and easing transition.
The Outcomes: Improving healthcare research
An easier way to integrate healthcare data
By automating processes, the redesigned data warehouse reduced manual work for those trying to extract data from it for research projects and requests. New self-service functionality, in designing a system to populate the I2B2 data mart and easy population for future I2B2 data marts, make it quicker and easier for researchers to get the data they need. The system also delivers higher data quality, which is critical to grant approvals and research integrity.
A more functional, operational system for back-end users
We added robust new features to the redesigned data warehouse that made it far more user friendly. Enhancements like error logging controls increase visibility and maintainability. Other new features deliver enhanced traceability, auditability, and reproducibility within the data warehouse. The system also now includes data lineage to improve data quality and facilitate transparency.
A sustainable system to support future needs and growth
The redesigned data warehouse is now easier to maintain, sustain, and extend for future needs, such as easy integration of new data sources. Leveraging technologies that the institute can currently support, the redesign will make it easy to move to more modern data technologies in the future. Aligning with the current technology roadmaps and removing unsustainable technologies also reduced costs. The system is also now documented for easier understanding and onboarding of new users.
Enhancing organization-wide communication and collaboration
Throughout the course of the project, the HCDAgile process evolved the institution’s work processes. Close engagement with their teams helped them learn and apply industry best practices for software development across the organization to increase communication and collaboration. With ownership over the solution from the start, their teams can easily support the data warehouse internally moving forward.