Rush and nference build AI data platform

Rush University System for Health and nference said Wednesday they have formed a long-term partnership to build a secure clinical analytics platform hosted by Rush for research on de-identified patient data. The companies said the system will let Rush researchers work with longitudinal, multimodal clinical information that has been prepared for analysis. nference, based in Cambridge, Massachusetts, said it will supply federated software, data harmonization tools and analytics infrastructure for the project. The announcement was made Sept. 2 in a joint statement. ### What exactly are Rush and nference building? The platform described by Rush and nference is a Rush-hosted environment for analyzing de-identified clinical data drawn from multiple formats over time. The companies said it is intended to make structured and unstructured information “research-ready” by converting it into curated datasets for observational studies and prospective research. (prnewswire.com) nference said its role is to architect the platform using federated software and analytics infrastructure. Rush said the system will be hosted within its own environment, a detail the companies highlighted alongside the platform’s security and real-time capabilities. ### What does “multimodal” data mean in this deal? (prnewswire.com) nference’s own product materials describe its research platforms as combining several kinds of clinical information, including structured records, unstructured notes, imaging, molecular data and physiologic signals. The Rush announcement did not list every data type that will be included at launch, but it said the collaboration covers de-identified, longitudinal and multimodal clinical data. (prnewswire.com) The company also says its broader network includes radiology images, pathology slides, ECG data, echocardiogram videos and clinical notes. Those materials describe the type of data environment nference has been building with health-system partners, though the Rush release did not specify which of those assets will be available first in Chicago. ### Why are the companies emphasizing that Rush will host it? Rush and nference said the platform will be hosted by Rush rather than run as an outside repository. (prnewswire.com) In the announcement, the companies paired that point with repeated references to de-identified information and secure analytics, signaling that the data will remain inside a Rush-controlled setting while being prepared for research use. (nference.com) The joint statement said the goal is to help Rush researchers analyze data more effectively while protecting patient privacy through de-identification. The companies did not disclose financial terms, a launch timetable or a target number of researchers who will use the system. ### How does this fit into Rush’s broader AI push? (prnewswire.com) Rush has already announced other AI-related partnerships. In July 2025, Rush and Fabric said they were working together on Rush Connect and Rush Connect+, digital-first care access products that used AI and virtual-care technology to expand access and reduce wait times. (prnewswire.com) The new nference partnership is focused on research infrastructure rather than front-end patient access. Wednesday’s release said the platform is meant to support AI-enabled clinical research and patient care by giving researchers curated datasets that can be used in real time. ### What comes next for the partnership? (prnewswire.com) The companies said the platform is expected to support increasingly advanced multimodal and agentic AI research as the partnership develops. In the same statement, Rush and nference said the system is intended to continuously transform de-identified clinical information into research-ready assets for future studies. (prnewswire.com) Rush researchers are the named first users in the announcement, and nference said it will provide the software, data harmonization and analytics layer as the build-out proceeds. The Sept. 2 release did not give a go-live date, but it identified observational studies and prospective research as the immediate use cases. (prnewswire.com)