OBJECTIVES: The Generating Evidence in Oncology (GEO) research environment integrates multiple real-world administrative claims and clinical datasets to provide a unique platform for oncology research. This study described GEO patient demographics and contrasted them to the American Community Survey (ACS) and the US Cancer Statistics Public Use Database (USCS), which includes cancer incidence data from 50 states and territories. The study focused on non-small cell lung cancer (NSCLC).

METHODS: Patients with incident NSCLC from 01/01/2015 to 12/31/2019, aged ≥18 years old, without other primary cancers, and with ≥6 months of health plan enrollment before diagnosis date were identified from two GEO data containers.

(1) HIRD+CCQP: contains closed claims data from the HealthCore Integrated Research Database (HIRD) of commercially-insured and Medicare members and clinical data from a Cancer Care Quality Program (CCQP) where medical oncology practices submit information as part of incentive and utilization management.

(2) HIRD+BHI+ION: contains closed claims data from HIRD and Blue Health Intelligence (BHI) and electronic medical records from IntrinsiQ (ION).

Patient demographics and all-cause mortality were described in each container and contrasted with the ACS and USCS.

RESULTS: 6,223 (HIRD+CCQP) and 1,176 (HIRD+BHI+ION) patients with NSCLC were identified. Both GEO data containers showed similar patient demographics: mean age at first diagnosis was 63, 71% were White, 74% (HIRD+CCQP) and 86% (HIRD+BHI+ION) resided in the South/Midwest. The all-cause mortality rate was 30.7 (HIRD+CCQP) and 24.5 (HIRD+BHI+ION) per 100 person-years. Compared to national databases, patients in GEO data over-represented the working group aged <65, had better socio-economic status, and appeared to have more advanced disease.

CONCLUSIONS: The linked GEO research environment provides a sophisticated resource to enhance understanding of real-world oncology patient outcomes. The commercially insured population is younger but more diverse than the national NSCLC population and most relevant when used for understanding working-age patients.