OBJECTIVES: Electronic medical records (EMR) are a common source of real-world data in oncology research. However, capturing exposure to certain oncological treatments, such as oral anti-cancer agents, can be challenging for EMRs. This study aimed to validate a curated EMR dataset for exposure to oral anti-cancer agents compared with pharmacy claims from an open administrative claims database.

METHODS: In this retrospective study, curated EMR data from the U.S.-based ConcertAI Patient360^TM database was compared with dispensed pharmacy claims from an administrative claims dataset. Patients included in the analysis were aged ≥18 years with a confirmed diagnosis of melanoma at any stage and were present in both datasets. The two datasets were linked using patient-level data tokenization. All exposures to oral anti-cancer agents (encorafenib, binimetinib, dabrafenib, trametinib, vemurafenib, cobimetinib, lenvatinib, imatinib, larotrectinib, entrectinib) were captured from January 2014 to March 2022. The evaluation of the EMR dataset included sensitivity and specificity for exposure to any oral anti-cancer agents and sensitivity for exposure to specific oral anti-cancer agents.

RESULTS: Overall, 5,839 melanoma patients met the criteria for this study and 995 patients (17.0%) had exposure to any oral anti-cancer agent in at least one of the datasets. The concordance of exposure to any oral anti-cancer agents at the patient level was moderately high at a rate of 87.3%. The curated EMR dataset had a sensitivity of 92.4% and specificity of 87.1% for capturing any oral anti-cancer agent exposure compared with the claims dataset. For specific oral agents, the EMR data set had a sensitivity of 88.8% for capturing exposures found in the claims dataset.

CONCLUSIONS: Deeply curated EMR data adequately captures exposure to oral anti-cancer agents compared to an open administrative claims dataset. Augmenting EMR data with administrative claims pharmacy data would increase the number of patients with a documented exposure to oral anti-cancer agents.