Case Study: The power of patient matching in real-world data

The case study addresses the challenge of matching patients across de-identified datasets and the value of accurate identity resolution for healthcare research and analytics. Kythera Labs, a healthcare data transformation company, implemented Datavant's identity-resolution technology, which uses machine-learning models trained on EHR and consumer data alongside referential data to improve matching accuracy. The implementation narrowed the number of distinct patients by 96 million (22%) of the dataset, improving data quality and comprehensiveness by eliminating duplicate patient records across combined datasets without compromising privacy.

Datavant is the tokenization and linkage infrastructure behind most of the US real-world-data ecosystem — it de-identifies and connects patient records across health-data partners without exposing PII.

This sits in the Data Providers stage of the AI use-case lifecycle: Named examples from the real-world-data, claims and competitive-intelligence providers pharma AI systems are trained on, benchmarked against, or run inside of.