Adapting to disruption? Changes in colorectal cancer screening modalities and disparities during the COVID-19 pandemic
Principal Investigator
Cameron Kaplan
Associate Professor, Medicine, Director of the Gehr Family Center for Health Systems Science and Innovation, Keck School of Medicine, University of Southern California
Co-Investigators
- Sarah Axeen, Director of Data and Analytics, Schaeffer Center; Assistant Professor, Division of Emergency Medicine Research, University of Southern California
- Chun Nok Lam, Research Scientist, Emergency Medicine, University of Southern California
- Sonali Saluja, Assistant Professor, Gehr Family Center for Health Systems Science and Innovation, University of Southern California
- Todd Schneberk, Assistant Professor, Clinical Emergency Medicine, University of Southern California
- Jennifer Tsui, Associate Professor, Population and Public Health Sciences, University of Southern California
Funded By
National Institute on Minority Health and Health Disparities (NIMHD)
The problem:
The COVID-19 pandemic profoundly altered access to routine services, including routine in-person cancer screenings. For some screenings, such as for colon cancer, there were alternative methods that were safer and introduced less risk than in-person screenings.
Non-invasive Fecal Immunochemical Tests (FIT) were commonly used by Hispanic and Asian adults before the pandemic. However, colonoscopies were more common among non-Hispanic Whites.
Kaplan and colleagues aimed to understand how use of colorectal cancer screening modalities changed during the pandemic and if there were any disparities in usage type.
The approach:
The researchers employed two sources of data:
• OPTUM Clinformatics Data Mart, a nationwide commercial claims dataset, from 2018-2021, and
• Los Angeles County Department of Health Services encounter data for four hospitals and 24 clinics from 2018-2022.
From these data sets, the team reviewed colorectal cancer screenings by month for people ages 40-75. Using linear probability models and difference-indifference regressions, the team predicted monthly group-level probabilities of screening use, controlling for demographic variables and comorbidities.
The findings:
Colonoscopies were more commonly used in the prepandemic period (January 2019-February 2020) and the post-closure pandemic period (July 2020-December 2021). However, during the acute pandemic period (March-June 2020), both FIT tests and colonoscopies decreased, but FIT tests were more commonly used.
Individuals with one or more comorbidities were consistently less likely to undergo colonoscopies throughout the study period. Notably, during the acute pandemic phase, there was a statistically significant decrease in colonoscopies among these individuals, and this decrease was not offset by an increase in FIT tests.
Looking at differences in screening modalities by race and ethnicity, researchers found that Hispanic adults experienced a relative decline in both FIT tests and colonoscopies compared to non-Hispanic Whites. For colonoscopies, disparities between Hispanics and non-Hispanic Whites as well as between Asians and Whites widened, while screening rates for Blacks increased relative to Whites.
While the pandemic may have encouraged non-invasive screenings, persistent gaps remained without sufficient catch-up, highlighting a critical need for enhanced outreach to communities with lower screening rates, particularly the Hispanic population.
A line graph titled “Adjusted FIT test monthly screening rate pre/post pandemic closure by race/ethnicity.” The graph illustrates the pre- and post-COVID monthly FIT test screening rates by race and ethnicity for four groups.
Non-Hispanic White: 0.64% pre-COVID (0.641-0.646 CI), 0.57% post-COVID (0.568-0.571 CI).
Non-Hispanic Black: 0.56% pre-COVID (0.559-0.57 CI), 0.53% post-COVID (0.527-0.536 CI). Coefficient: 0.04%, p<0.01.
Hispanic: 0.94% pre-COVID (0.93-0.944 CI), 0.79% post-COVID (0.782-0.793 CI). Coefficient: 0.08%, p<0.01.
Non-Hispanic Asian: 0.88% pre-COVID (0.869-0.89 CI), 0.81% post-COVID (0.803-0.82 CI). Coefficient: 0.01%, p=0.406.