Is There Bias in the Assessment of Contraindications for Resection? Disparities in the Surgical Management of Early-Stage Esophageal Cancer

TL;DR

The study finds that decisions to avoid surgery for early-stage esophageal cancer vary widely between hospitals and are influenced by patient demographics and institutional factors.

Contribution

This study reveals significant disparities in surgical contraindication assessments for esophageal cancer patients across U.S. institutions.

Findings

Contraindication status is associated with age, race, socioeconomic status, and insurance type.

Observed-expected ratios for contraindication status varied over 1000-fold between institutions for cT1N0M0 patients.

Underserved minorities are at higher risk of being labeled contraindicated for surgery.

Abstract

Background: Resection is considered the standard of care for patients with localized esophageal cancer who are “physiologically fit”. Patients who do not meet this standard are considered contraindicated to receive surgery. We hypothesized that among patients with non-metastatic esophageal cancer, the consideration of contraindication status would vary based on clinical and demographic factors and would vary between institutions. Methods: We identified patients with non-metastatic gastric and esophageal cancer in the National Cancer Database (NCDB) from 2004 to 2018. Patients were categorized into three groups based on surgical treatment: surgical resection (including endoscopic mucosal resection), resection contraindicated, and refusal of resection based on the coding of the “reason for no surgery” data element. Demographic, clinical, and institutional characteristics were compared between the groups using bivariate and multivariate techniques to identify factors associated with contraindicated status. A subgroup analysis of cT1N0M0 patients was also used to assess every institution in the NCDB’s observed–expected ratio for contraindication status. Results: In total, 144,591 patients with non-metastatic disease met inclusion criteria: 124,972 (86%) underwent resection, 13,793 (10%) were contraindicated for resection, and 5826 (4%) refused resection. Contraindication was associated with age, non-Hispanic Black race, socioeconomic status, Charlson–Deyo score, insurance type, institution characteristics, clinical T-stage, and clinical N-stage. There were 9459 patients who were cT1N0M0 and had no co-morbidities. In this cohort, there were more than 1000-fold differences between individual programs regarding observed–expected ratio of contraindication status when adjusting for clinical and demographic characteristics. Conclusions: Variation in the assessment of contraindication status varies dramatically between institutions. Underserved minorities, including age, race, and insurance type, are risk factors for being considered contraindicated. These findings highlight the disparities that exist regarding surgical care of non-metastatic esophageal cancer in the United States.

Figures