Tick tock on the stomach clock: Factors affecting time to diagnosis of gastric adenocarcinoma at an urban cancer center with a large East Asian patient population.

Abstract

e18508 Background: Gastric cancer is the fifth most common cancer in the world and disproportionately affects the East Asian population. Various factors contribute to this disparity such as genetics and environmental exposure. Given that language barriers can contribute to reduced quality of healthcare delivery, screening rates, and patient safety, it is important to assess the impact of language on the diagnosis of gastric cancer. Tufts Medical Center’s proximity to Boston’s Chinatown make our patient population unique for analysis of these factors. Here we present a single-institution experience of the factors affecting gastric adenocarcinoma time to diagnosis. Methods: We used ICD-10 codes to retrospectively identify patients at Tufts Medical Center with a diagnosis of gastric cancer between January 2011 and December 2021. We excluded patients that did not have a diagnosis of gastric adenocarcinoma or lacked adequate clinical documentation. We recorded demographic, clinical, and treatment characteristics including race (White, Asian, or other), primary language (English, Asian or other), presenting symptoms, enrollment in hospice, and interpreter status (certified, non-certified, none, or not applicable). Certified interpreters included professional medical services for in-person, video, or phone interpreters. Non-certified interpreters included family members, physicians fluent in the language, or others. We assessed the impact of these on the time between initial presentation to the healthcare system and eventual diagnosis of gastric cancer. Results: 80 patients were included in the analysis. Mean days to diagnosis by race were as follows: White (40), Asian (43), other (25), p = 0.20. Mean days to diagnosis based on language were as follows: English (40), Asian (43), other (6), p = 0.07. Mean days to diagnosis based on interpreter status were as follows: certified (19), non-certified (13), none (73), NA (40) (p = 0.05). When pairwise comparison was done, there was no difference between certified and non-certified interpreter (p = 0.40). There was no difference between certified interpreter and no interpreter (p = 0.06). There was a statistically significant difference between non-certified interpreter and no interpreter (p = 0.04). Presenting symptoms and eventual enrollment in hospice did not significantly impact time to diagnosis. Conclusions: We saw a shorter time from initial presentation to diagnosis of gastric adenocarcinoma in non-English speaking patients who used an interpreter at the initial visit compared to those who did not. One should always consider interpreter availability as an important part of reducing health disparities.