Subsalt Unveils Secure Runtime for PHI-Compliant Data Analysis

The platform functions by isolating raw data within a controlled space, where users can execute metrics, test hypotheses, and generate heatmaps or scatterplots without accessing patient-level details. Subsalt retains governance over data exports and access rights, ensuring that only approved analytical outputs leave the environment. This release complements Subsalt’s existing digital twin technology, which creates de-identified patient populations for clinical research.

Dr. Emily Payne, a senior privacy expert at Datavant, noted that the certification process involved rigorous risk assessment for attribute and membership disclosure. The system is engineered to align with the stringent "very small" risk threshold required by HIPAA. In addition to Secure Runtime, the company introduced Subsalt MCP, a tool designed to connect AI chat interfaces with custom de-identified datasets. According to Subsalt CTO Luke Segars, these natural-language tools aim to bridge the gap between clinical inquiry and technical data navigation, allowing clinicians to query systems directly through agents.