Drummond announced today that Eleos has completed the pDSI-Risk Certification program for predictive decision support interventions (pDSI). The review confirms that Eleos’ AI capabilities implement practical safeguards for Intervention Risk Management (IRM) and Source Attribution disclosure, aligned with expectations under ASTP/ONC’s standards for pDSI solutions supplied to health care organizations.
The certification evaluates whether AI-enabled health IT can reliably identify and mitigate risks, document how recommendations are generated, and disclose intended uses, risks, and governance activities in a way that is understandable to users. Eleos demonstrated these controls across representative and real-world usage scenarios, with an emphasis on clarity, traceability, and operational governance. Details of Eleos’ approach are highlighted in this case study.
“As our work scales to larger enterprises and regulatory stakeholders, expectations for assurance grow,” said Raz Karmi, Chief Information Security Officer at Eleos. “This certification reflects the way we operate: decisions are explainable and risk is managed as part of routine operations, not as an afterthought.”
“pDSI-Risk translates principles like transparency and accountability into concrete practices,” said John Valukevich, Director of Programs at Drummond. “Eleos showed a strong grasp of governance and controls that support knowledgeable and informed use of AI in clinical workflows.”
Drummond’s pDSI-Risk Certification offers a path for modular or standalone AI solutions, including products outside the ONC Health IT Certification scope, to demonstrate alignment with risk, governance, and transparency expectations. By validating intervention risk assessment, monitoring, documentation, and source attributes disclosures, the program helps organizations communicate how their AI behaves and how it is governed over time.
Read the original press release