Novartis Reimagining Local Laboratory Reference Range Management
The Novartis Clinical Data Operations team was the recipient of the SCDM 2026 Innovation Award for Innovation in Health Technology Solutions, recognizing its transformative approach to simplifying the LNR (Local Lab Normal Reference Range) dataflow while enhancing the experience of investigational sites and Clinical Research Associates through an internally developed AI-enabled solution.
Solution overview
The solution applies advanced artificial intelligence and metadata-driven automation to convert complex, unstructured local laboratory reference range information into standardized, structured, and loadable data. Designed to support CRAs, data managers, local laboratory administrators, study teams, and clinicians, the solution streamlines the end-to-end workflow—from document interpretation and analyte identification to validation, mapping, and downstream system readiness.
The challenge
Local laboratory reference ranges are essential for the accurate interpretation of safety laboratory results. However, they are often received from investigational sites in highly variable formats, including spreadsheets, PDFs, scanned images, translated documents, and native laboratory reports. Differences in language, terminology, structure, units, and completeness make processing labor-intensive and heavily dependent on human interpretation.
Traditional workflows require teams to manually review, transcribe, validate, and reconcile this information before entering it into clinical systems. The resulting administrative burden can delay the availability of safety-ready laboratory data, increase the potential for inconsistencies and rework, and limit the time clinical teams can devote to site engagement, quality oversight, and patient-focused activities.
The innovation
The solution introduces specialized AI agents that extract, interpret, validate, and manage reference range information across diverse source formats. Optical Character Recognition (OCR) and Generative AI accelerate the interpretation of large and complex files, while integration with a governed metadata repository enables the solution to identify laboratory analytes and map them consistently to approved standards.
The solution can distinguish between new and follow-up LNR artifacts, identify incremental changes requiring review, and highlight missing values, anomalies, or potential quality issues. Rather than replacing human expertise, the agent augments it—automating repetitive activities, surfacing exceptions, and enabling users to focus their judgment where it adds the greatest value.
What differentiates the solution is its combination of intelligent document processing, governed metadata, workflow integration, and human oversight within a scalable framework designed for global clinical operations.
Impact and results
The agent currently saves an estimated 6,000 hours annually by reducing manual effort associated with transcription, quality control, data entry, and follow-up with local laboratories, CRAs and investigational sites. As adoption expands and automation matures, annual savings are projected to reach up to 10,000 hours.
Beyond efficiency, the solution improves consistency in analyte mapping, strengthens traceability and governance, reduces rework, and accelerates clinician access to safety-ready laboratory data. It also reduces the administrative burden on CRAs, allowing greater focus on site partnership, oversight, and patient-centric trial execution.
Looking ahead
The next phase of the solution includes an intelligent search capability that will help teams identify and reuse previously processed reference range artifacts based on site, location, effective date, and related attributes. Novartis is also exploring collaboration with technology providers and industry peers toward a common, reusable reference range dictionary.
Our ambition is to help advance a more connected, standardized, and intelligent approach to local laboratory metadata management—strengthening clinical data quality, improving site experience, and supporting faster, safer, and more patient-focused clinical research.
