Taldavi Health Index

Routine laboratory data contains more information than conventional interpretation reveals.

Disease emerges through subtle physiological changes long before individual lab values become abnormal. The Taldavi Health Index uses AI to identify these hidden patterns in routine clinical data, generating physiological insights that support earlier, more informed clinical decision-making. It is a pattern-recognition platform, not a diagnostic test.

First Commercial Focus: MASH

Our initial commercial application focuses on Metabolic Dysfunction-Associated Steatohepatitis (MASH), demonstrating how routine laboratory data can reveal clinically meaningful disease patterns beyond conventional interpretation.

Built for Healthcare Partners

The Taldavi Health Index is for organizations seeking to extract greater clinical value from existing laboratory data, including:

  • Pharmaceutical companies

  • Reference laboratories

  • Healthcare organizations

  • Health plans and payors

  • Consumer health platforms

Current Performance

  • 0.89 AUC using the NHANES dataset

  • 37,000-person clinical dataset

  • 80/20 holdout validation methodology

  • AST, ALT, age, and platelets intentionally excluded to prevent reconstruction of the FIB-4 score

  • External validation currently underway

  • A development platform demonstrated across approximately 21 disease applications, with MASH as the first commercial focus

Platform Capabilities

The Health Index platform is designed to deliver:

  • Early physiological drift detection

  • Multi-biomarker pattern recognition

  • Health trajectory mapping

  • Actionable insights that support clinical decision-making

  • A scalable AI platform supporting a growing pipeline of disease-specific models