No single data source tells the whole story. Claims data captures diagnoses and procedures but may miss clinical nuance. EHRs capture lab values and physician notes but are limited to a single health system. Linking these sources at the patient level – using privacy-preserving methods such as tokenization – enables analysts to build more accurate and complete datasets for research, targeting and measurement. The integrity of any linked dataset depends on the quality of the matching methodology and the comprehensiveness of the underlying sources.