Abstract

Introduction

To encourage use of the Electronic Medical Record Search Engine (EMERSE) network functionality (Figure - see poster) across institutions, this work aimed to develop a generalizable network participation framework, taking into consideration regulatory, security, and compliance requirements. Collaborating among 5 institutions already utilizing EMERSE— U of Michigan, Columbia U, U of Cincinnati, U of Kentucky, and the U of Virginia—we explored barriers preventing sites from turning on the networking functionality. Notably, the network feature only provides obfuscated counts, and no protected health information (PHI) is transmitted across the network, even when underlying documents might contain PHI.

Methods

The study engaged stakeholder groups at each site, including Institutional Review Boards (IRBs), information technology (IT) security teams, regulatory compliance teams, legal advisors, and data governance bodies, to gather feedback on concerns or approvals related to the EMERSE networking features. Participating sites received preparatory resources, including the EMERSE Networking Guide2 (covering infrastructure, architecture, and configuration) and a Terms of Use document (outlining the roles and rules of network membership), to facilitate their internal discussions. Common themes were derived from meeting notes.

Results

Five main barriers to implementing EMERSE’s networking functionality emerged: 1. Diverse Consultative Requirements: There was significant variability across sites regarding the groups required for approval, with universal involvement from IT security and compliance, but varied necessity for Data Governance, IRB, and chief research information officer (CRIO) consultations. 2. Protracted Approval Timelines: Sites struggled with the slow pace of obtaining necessary approvals, often delayed by other institutional priorities or infrequent meetings of key groups. 3. Lack of Structured Approval Processes: Most sites lacked a clearly defined process or an accountable entity to authorize the networking functionality, resulting in uncertainty in moving forward and additional delays. 4. Excessive Caution in Data Sharing: Despite robust documentation and comparable existing networks like ENACT, some sites exhibited heightened apprehensiveness towards data sharing across EMERSE instances, even when only sharing obfuscated patient counts. 5. Limited Awareness and Demand: Local institutional data use was a higher priority than network sharing, although this may have been due to a lack awareness of the networking feature, especially when not advertised. Sites may wait for increasing user demand before taking the time to pursuing governance and approval processes.

Discussion and Conclusions

Contrary to initial assumptions, this work revealed no flaws within the technical documentation, Terms of Use, technical design, or system architecture. The reluctance/inability to activate the EMERSE network feature largely arose from institutional and procedural barriers, diverse consultative requirements, cautious data sharing attitudes, and limited demand, rather than concerns about the software itself. Socio-technical and bureaucratic barriers remain impediments to the broader adoption of research tools.