The Impact of Digital Inequities on Esophageal Cancer Disparities in the US

Abstract

Background: There is currently no comprehensive tool that quantifiably measures validated factors of modern technology access in the US for digital inequity impact on esophageal cancer care (EC).

Objective: To assess the influence of digital inequities on esophageal cancer disparities while accounting for traditional social determinants.

Methods: 15,656 EC patients from 2013-2017 in SEER were assessed for significant regression trends in long-term follow-up, survival, prognosis, and treatment with increasing overall digital inequity, as measured by the Digital Inequity Index (DII). The DII was calculated based on 17 census tract-level variables derived from the American Community Survey and Federal Communications Commission. Variables were categorized as infrastructure access or sociodemographic, ranked, and then averaged into a composite score.

Results: With increasing overall digital inequity, significant decreases in the length of long-term follow-up (p < 0.001) and survival (p < 0.001) for EC patients were observed. EC patients showed decreased odds of receiving indicated surgical resection (OR 0.97, 95% CI 0.95-99) with increasing digital inequity. They also showed increased odds of advanced preliminary staging (OR 1.02, 95% CI 1.00-1.05) and decreased odds of receiving indicated chemotherapy (OR 0.97;95% CI 0.95-99).

Conclusions: Digital inequities meaningfully contribute to detrimental trends in EC patient care in the US, allowing discourse for targeted means of alleviating disparities while contextualizing national, sociodemographic trends of the impact of online access on informed care.

Keywords: broadband service; esophageal cancer; internet access; social determinants of health; technology infrastructure.

Figure 1

Distribution of total DII ranked scores across the US. Ranked digital inequity scores were assigned per county in the (A) total composite DII, (B) infrastructure access and usage, and (C) sociodemographic categories.

Figure 2

Relative decreases in months surveyed with increasing DII scores. (A) Percentage decreases from lowest to highest DII quintiles based on mean months surveyed for total DII score and subcomponent DII theme subscores. EC patients were assigned DII scores and split into relative quintiles. (B) A linear regression across all the represented values (i.e., not the mean values) in each of the boxplot quintiles was performed to assess for continuous trend significance of the surveillance period for increasing the total DII. Boxplots = median, IQR, 1.5*IQR; mean months surveyed per quintile = maroon diamonds; outliers = black dots; p-value for regression.

Figure 3

Relative decreases in months survival with increasing DII scores. (A) Percentage decreases from lowest to highest DII quintiles based on mean months survived for total DII score and subcomponent DII theme subscores. EC patients were assigned DII scores and split into relative quintiles. (B) A linear regression across all the represented values (i.e., not the mean values) in each of the boxplot quintiles was performed to assess for continuous trend significance of the survival period for increasing the total DII. Boxplots = median, IQR, 1.5*IQR; mean months survived per quintile = maroon diamonds; outliers = black dots; p-value for regression.