Abstract Introduction: While lower socioeconomic status (SES) is associated with worse cancer outcomes, most electronic health records (EHRs) lack documentation of SES. Here we apply an area-level SES measure to an EHR-derived database to evaluate representativeness according to SES. We then examine socioeconomic disparities in the timeliness of healthcare utilization (biomarker testing and systemic treatment) and overall survival (OS) in 3 common cancers. Methods: This retrospective study uses the nationwide Flatiron Health EHR-derived de-identified database of cancer patients who have clinical activity between January 2011-August 2021. Census block group data from the American Community Survey (2015-2019) was used to measure SES per the Yost Index (incorporating income, home values, rental costs, poverty, blue-collar employment, unemployment, and education information). SES quintiles were determined from the US population and then applied to patients based on their residential addresses. Our database included 2,067,644 cancer patients from community practices. Patients with advanced non-small-cell lung cancer (aNSCLC, n=51,596), metastatic breast cancer (mBC, n=18,268), and multiple myeloma (MM, n=8,246) were sampled and followed from advanced, metastatic, or initial diagnosis date, respectively. Differences were evaluated using Χ2 tests for categorical variables and log-rank tests for Kaplan-Meier survivor functions. Results: Compared to the US population, our database of cancer patients has a similar SES distribution (differences <3%), capturing cancer patients living in the most (Q1) and least affluent areas (Q5) of the country. Yet, among the 1.5% of cancer patients who participated in clinical trials, only 15.2% lived in the least affluent areas. aNSCLC and mBC patients living in the least affluent areas were less likely to receive biomarker testing within 30 days of index diagnosis than those in the most affluent areas (59.4% vs. 68.7%; 74.9% vs. 81.0%, both p<.01). Similar patterns were observed in receipt of systemic treatment within 60 days of index diagnosis (aNSCLC: 58.9% vs. 64.9%; mBC: 76.0% vs. 80.1%, both p<.01). No differences in healthcare utilization were observed among MM patients. Patients in the least affluent areas had lower median OS (months) than those in the most affluent areas (aNSCLC: 10.8 [95% CI: 10.4-11.3] vs. 12.2 [95% CI: 11.8-12.7]; mBC: 28.0 [95% CI: 26.6-29.5] vs. 34.5 [95% CI: 33.1-36.9], MM: 57.4 [95% CI: 53.5-61.5] vs. 67.5 [95% CI: 62.5-76.4]; all p<.01). Conclusion: Lower SES was associated with reduced clinical trial participation, less timely healthcare utilization, and worse OS. Making SES available in real-world data can support the development of inclusive clinical trials and inform interventions to reduce cancer care disparities. Citation Format: Jenny S. Guadamuz, Xiaoliang Wang, Jeremy Snider, James Walters, Rebecca A. Miksad, Gregory S. Calip. Socioeconomic disparities in healthcare utilization and overall survival among patients with cancer: Application of area-level socioeconomic status in a nationwide electronic health record-derived database [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 3668.
Read full abstract