The impact of digital inequities on gastrointestinal cancer disparities in the United States

Modern‐day internet access and technology usage substantially impacts aspects of surgical care but remain ill‐defined for their associations with gastrointestinal‐cancer (GIC) outcomes. We sought to develop the Digital Inequity Index (DII), a novel, a self‐adapted tool to quantify access to digital resources, to assess the impact of “digital inequity” on GIC care and prognosis.


| INTRODUCTION
Gastrointestinal cancers (GIC) remain among the top five leading causes of cancer-related morbidity and mortality worldwide. 1 Despite the implementation of numerous policy measures to combat disparities, [2][3][4][5] GIC mortality is expected to increase substantially over the coming decades. 6 Social determinants of health (SDoH) such as socioeconomic status, race/ethnicity, and rurality contribute to inequitable access to care and mortality among patients with GIC. [7][8][9][10][11][12][13][14] In fact, prior studies have demonstrated that various SDoH comprises 80%-90% of modifiable contributors to population health outcomes among patients with cancer and other diseases. 15 Recently, the American Medical Informatics Association stated that the digital divide is an important underrecognized SDoH. 16 The internet may play a critical role in disseminating knowledge to patients, as well as creating increased access to specialized healthcare services. In turn, differential access to digital resources may be an important driver in healthcare disparities. Digital inequity refers to disparities in the possession of internet-capable electronic devices, local network infrastructure, or economic broadband service pricing relative to the local income level. Several studies have suggested that digital and internet access may impact cancer prevention and diagnosis, as well as treatment and prognosis. [17][18][19] In addition, access to the internet and the use of digital resources have been associated with mitigating community health disparities among patients with diabetes and obesity. 20,21 While the importance of access to modern information technology has been previously highlighted, the availability of digital resources and the impact of digital accessibility among patients with GIC has not been investigated.

State-and national-level initiatives, including the Digital Divide
Index from the Rural Indiana State database and the Federal Communications Commission (FCC) Connect2Health Broadband Map, have been developed to address digital disparities. 22,23 These tools are limited, however, in that the FCC-Connect2Health is not regularly updated, while the Digital Divide Index has limited geographic scope. 22,23 To date, the impact of digital inequity on outcomes with GIC has not been investigated while accounting for other traditional SDoH such as socioeconomic status, race-ethnicity, and education. Therefore, the objective of the current study was to define a nationwide Digital Inequity Index (DII) to characterize access to digital resources that accounted for other SDoH factors. In addition, we sought to assess how this DII impacted outcomes among patients with GIC.    (Table 1).
County-level geographic distribution of total DII, as well as infrastructure-access DII and sociodemographic DII subcategories, are depicted in Figure 1.   GIC such as small bowel, liver, and biliary tract cancers ( Figure 5).

| DISCUSSION
The American Medical Informatics Association has declared that the digital divide is an important SDoH contributing to healthcare disparities. 16  Geospatial analyses have demonstrated large variations in accessibility to technology and the internet relative to patient health outcomes, especially among patients with a cancer diagnosis. [29][30][31] Several small pilot studies in underserved regions in rural Kentucky and Appalachia reported an association between decreased broadband access and increased cancer burden. [29][30][31] Internet connectivity has been increasingly important since the COVID-19 pandemic, as there was a dramatic increase in the adoption of telehealth patient-provider interactions via video-conferencing platforms. 32,33 The implementation of telephone-and internet-based connected health technologies has been associated with greater patient and provider satisfaction, as well as improved cancer symptom management outcomes, particularly in rural communities. 31  reported 18% lower odds of receiving surgery for patients with hepatocellular carcinoma who resided in highly socially vulnerable counties compared with individuals living in relative affluence. 11 In the current study, increasing sociodemographic DII was associated with a 9% lower likelihood of receiving surgery for liver cancer.
Increasing infrastructure-access DII was also associated with 8% lower odds of undergoing hepatic resection, even after adjusting for traditional, nondigital SDoH factors. Although the underlying reasons are likely multifactorial, these findings may in part be due to the increasing role that digital resources play in patient awareness and education, which may lead to earlier symptom recognition and presentation. To this point, increasing infrastructure-access to DII was associated with 8% higher odds of patients presenting with latter-stage disease. Papadakos et al. had previously reported that online websites were among the most preferred mediums for patients seeking information on GIC. 38 In fact, up to 74% of patients F I G U R E 1 Distribution of total DII ranked scores across the United States. Ranked-digital inequity scores in the total composite of infrastructure access-usage and sociodemographic categories were assigned per county. DII, Digital Inequity Index; NA, not available.
T A B L E 2 Logistic regression analysis to assess the impact of the Digital Inequity Index on stage at presentation and receipt of surgery. ask their healthcare providers about reliable websites that can be used to acquire more information about their diagnosis, as well as prognosis. 39 In fact, "cancer" is the second most common healthcarerelated search term across the internet, and specific search-terms such as "gastrointestinal cancer" and "bowel cancer" remain among the most frequent inputs. 39 41 As such, digital disparities may be a modifiable factor to target via local and national policies to close the "digital divide" and mitigate health disparities.
The findings in the current study should be interpreted in light of several limitations. Owing to the inherent limitations of the datasets, only patients diagnosed with GIC between 2013 and 2017 were included in the analytic cohort. As such, the impact of the COVID-19 pandemic relative to digital inequity could not be assessed. Future studies will need to assess the level of digital inequity after the implementation of the Medicare telehealth coverage waiver. Furthermore, as a large proportion of the present cohort was male and belonged to the White race/ethnicity, these findings may not be generalizable to female and minority patients. In addition, the SEER database does not provide several relevant sociodemographic characteristics that may have influenced or exacerbated the impact of digital disparities.
In conclusion, the DII metric is a comprehensive metric to quantify digital inequity. Variations in DII, which defined different cohorts of patients with varied access to digital resources, demonstrated that digital disparities were strongly associated with GIC care and outcomes-even after adjusting for traditional SDoH factors such as socioeconomic status, education, and disability. Patients with the highest DII were the most likely to be diagnosed at an advanced stage, while also being the least likely to receive surgery. Furthermore, increasing DII was associated with shorter surveillance time and worse long-term survival. DII may be used as a tool to identify communities with the least accessibility to digital resources, which may help inform policies to increase digital equity, and ultimately ensure equitable GIC care and outcomes.
F I G U R E 3 Relative decreases in mean months of surveillance with increasing DII scores. The percentage decreases from lowest to highest DII quintiles based on mean months surveyed for total-DII score and subcomponent DII-theme subscores. GIC patients were assigned DII scores and split into relative quintiles. DII, Digital Inequity Index; GIC, gastrointestinal cancer.
F I G U R E 4 Linear regression trends in months survival across increasing DII quintiles. Linear regressions across all the represented values (i.e., not the mean values) in each of the boxplot quintiles were performed to assess for continuous trend significance of the survival period for increasing total DII. Boxplots = median, IQR, 1.5 × IQR; mean months survived per quintile = maroon diamonds; outliers = black dots; p value for regression. DII, Digital Inequity Index; IQR, interquartile range.
F I G U R E 5 Relative decreases in mean months survival with increasing DII scores. The percentage decreases from lowest to highest DII quintiles based on mean months survived for total-DII score and subcomponent DII-theme subscores. GIC patients were assigned DII scores and split into relative quintiles. DII, Digital Inequity Index; GIC, gastrointestinal cancer.