OBJECTIVES: Health insurance claims have become the mainstream foundation for real-world evidence in recent decades. Yet, the lack of clinical details is a well-known limitation of claims databases. Thanks to modern technology of tokenization, the linkage between claims and lab results is now more accessible and with patient privacy preserved. Using a recently developed lab dataset as an example, this study explored the data capabilities of lab results and evaluated their potential to enhance claims databases.

METHODS: In a cross-sectional approach, this study assessed data availability of the Komodo Lab Results (KLR), characterized patient-level laboratory testing pattern, and quantified cohort overlap via linkage to administrative data and claims in the Komodo Research Dataset (KRD) from January 2016 to November 2023.

RESULTS: KLR has 33 lab tests and over 3.4 billion results recorded for over 59 million unique patients in the US, including blood cell/platelets, neutrophils, immunoglobulins, EGFR, and INR. Specifically, over 46 million patients had serum potassium values, with 65% repeatedly tested and a median 31.2 months since the first to last values. Equivalent estimates respectively for LDL-cholesterol and HbA1c were 34 and 28 million, 60% and 58% repeatedly tested, and 35.1 and 30.1 months. On average, 97.1% of patients also contributed to the KRD. Repeat testing rate was within 1% difference for all included tests between overall KLR patients and the subset overlapped with KRD.

CONCLUSIONS: Lab datasets contain a wealth of structured clinical information. Together with the longitudinality strength of claims databases, lab results are a valuable resource for researchers seeking contextualization of patient health over time and the granularity of outcomes assessment.