Society: AGA

Background: Although colorectal cancer (CRC) screening has been recommended for decades, about one-third of eligible individuals are not up-to-date with screening in the US. The COVID-19 pandemic has worsened this situation. Specific recommendations from health care providers have been found to have a critical role in promoting screening. We aimed to explore the impact of COVID-19 pandemic on CRC screening recommendations and its distinct effects on different populations.
Method: We analyzed the 2019 and 2021 datasets from the National Health Interview Survey (NHIS) in a retrospective cross-sectional study. Using a multistage probability design, the NHIS represents a broad geographical representation of the country. US adults aged 50-75 with no prior CRC and not up-to-date with CRC screening who had available information for CRC screening recommendations were included (N= 4270 [representing 31,665,798] and 3634 [representing 28,496,028] in 2019 and 2021, respectively). The 50-75 age group was considered to have better consistency between 2019 and 2021. A multivariable logistic model was used to identify factors associated with the likelihood of receiving screening recommendations from a provider in the past 12 months. In addition to analysis of the pooled dataset, we conducted separate analyses for 2019 and 2021 to determine if the COVID pandemic disproportionally impacted screening recommendations for specific populations. Weighted frequencies, odds ratio (OR) and confidence interval were estimated.
Results: 1,015 (22.22%) eligible individuals received recommendations for CRC screening in 2019, whereas 660 (17.9%) received them in 2021 (adjusted OR = 0.81, p<.001). Factors associated with less likelihood of receiving a recommendation both in 2019 and 2021 included low income, lack of insurance, racial and ethnic minorities, and no access to a usual care facility. Disparities related to race and insurance coverage became more pronounced in 2021 (Table 1). Also, the higher likelihood of referral in patients with higher education in 2019 was not seen in 2021. Opportunistic referrals were generated from urgent office visits in 2019 vs. emergent hospital visits in 2021.
Conclusion: A lower proportion of eligible individuals received recommendations for CRC screening from a provider during the COVID-19 pandemic in 2021 compared to 2019. While there was a disparity gap in 2019 in individuals with low socio-economic status, ethnic minorities, and those with no access to usual care facility, the gap widened further among ethnic minorities and the uninsured during the COVID-19 pandemic. Targeted efforts are even more needed as a result of the pandemic in order to decrease the disparities gap in CRC screening.

Introduction: Medicaid expansion is a provision in the Affordable Care Act that increases access to health insurance and preventive services for low-income individuals, however not all U.S. states participate. Individuals with Medicaid and in medically underserved areas often receive preventive services at Federally Qualified Health Centers (FQHCs). We aimed to study the impact of Medicaid expansion on colorectal cancer (CRC) screening rates in FQHCs by comparing CRC screening rates and predictors of screening in FQHCs in Medicaid expansion and non-expansion states.

Methods: We used national FQHC quality data from the 2021 Uniform Data System (UDS) to perform a cross-sectional analysis of all U.S FQHCs. We obtained 2021 CRC screening rates for each FQHC and for each state (FQHCs only) for patients age 50-74. (CRC screening data for patients age 45-75 are not available in UDS.) We then used Wilcoxon rank sum and chi-sq tests to compare FQHC patient-mix characteristics based on Medicaid expansion status. Finally, we performed mixed-effects linear regression models to determine FQHC patient-mix characteristics that predict high and low CRC screening participation in FQHCs in Medicaid expansion states and in FQHCs in non-expansion states. Medicaid expansion status was determined from Kaiser Family Foundation data in 11/2022.

Results: Overall, there were 6,940,879 patients eligible for CRC screening in 1,284 U.S. FQHCs in 2021. FQHC patient characteristics differed based on Medicaid expansion status (Table 1). The median CRC screening rate among all FQHCs in 2021 was 40.8%. CRC screening rates were significantly higher in FQHCs in Medicaid expansion states than in non-expansion states (42.1% v. 36.5%, p≤0.0001) (Table 1). In the adjusted model for states without Medicaid expansion, FQHCs in rural settings (urban coef 4.21, 95%CI 0.03, 8.39) or with a high proportion of uninsured patients (coef -10.27, 95%CI -14.00, -6.54) had significantly higher odds of lower CRC screening rates (Table 2). FQHCs located in Medicaid expansion states, however, experienced significantly lower CRC screening rates if they had large proportions of male, Black, Hispanic, low income, unhoused, or uninsured individuals (Table 2).

Discussion: CRC screening rates in U.S. FQHCs are significantly higher in states that implemented Medicaid expansion than in non-expansion states. The impact of being uninsured on participation in CRC screening remains profound in non-expansion states, while race/ethnicity, homelessness, and poverty also predict screening utilization in Medicaid expansion states. Our results suggest that Medicaid expansion states have minimized CRC screening disparities due to uninsured status and that targeted interventions to improve CRC screening participation should differ in FQHCs in Medicaid expansion and non-expansion states.

Intro: Rates of colorectal cancer screening (CRC) in medically underserved populations (MUPs) are lower compared to the general population, but optimal interventions to improve screening rates in MUPs have not been well-defined. The objectives of this study were to review published interventions aimed at improved CRC screening in MUPs and to identify characteristics associated with successful programs.

Methods: A systematic review of the PubMed, Google, Embase, and Cochrane databases was performed using standard PICO format to construct keyword searches. Studies were included if they described interventions to improve CRC screening in MUPs. Both colonoscopy and stool-based CRC screening methods were included. Studies outside of the United States or that did not include data for baseline, or post-intervention screening rates were excluded. The primary outcome was the success of the intervention, evaluated by calculating the percentage increase (PI) in CRC screening rate ((post-intervention – pre-intervention)/pre-intervention x 100). Interventions that failed to achieve a PI of 125% were not included in the final analysis. Two reviewers assessed each study's eligibility, and a third reviewer served as a tiebreaker. Study quality was evaluated using standard NIH and CASP quality assessment checklist tools. Interventions were grouped by type, and interventions with >2 elements were classified as "multicomponent."

Results: The initial search yielded a total of 572 studies, from which 22 studies including 33 interventions met the criteria for final analysis. Dates ranged from 2003-2018. The mean sample size was 805 ± 1203 (SD) individuals (range: 119-5970). CRC screening method was variable, with 20 interventions to improve rates of stool-based testing and 15 focused on colonoscopy (12 included both). The most common types of interventions included those that were multicomponent (N=19), those that centered around the use of a patient navigator (N=7), and those that primarily used mailings to patients for screening stool-based kits and provider recommendation letters (N=3). The Median PI for the collection of studies was 233% (IQR: 185%, 552%). The most successful interventions were mailer-based kits including a provider letter (median PI 1187%, [IQR:684-1193]), multicomponent interventions (median PI 233% [IQR:187,633]), and interventions using patient navigators (median of 224% [IQR:168,295]).

Conclusion: Randomized studies focused on improving colorectal cancer screening rates in MUPs vary in their definitions of MUP and in intervention type. The most effective published interventions include mailed stool-based kits plus provider letters, those with multiple intervention components, and those built around patient navigation. Clinicians serving MUPs should consider these specific strategies when designing programs to improve CRC screening.

Background: <20% of individuals suspicious for Lynch syndrome and other inherited cancer syndromes undergo genetic testing in the US. Undiagnosed patients will not benefit from enhanced cancer screening and prophylactic interventions. It is imperative we find mechanisms to dramatically improve the identification of candidates for LS genetic testing.
Method: We created a patient registry by converting National Comprehensive Cancer Network (NCCN)/American College of Medical Genetics based genetic testing criteria for all cancer syndromes into distinct rule-based conditional logic statements in the HER. The registry leverages structured data from the EHR including personal and familial history. Here we report patients identified because they met NCCN genetic testing criteria for Lynch syndrome. Patient evaluation status, as well as data on demographic, socio-economic, and clinical factors were collected. Next, logistic regression modeling was exploited to identify independent determinants of likelihood for evaluation. Association of determinants with evaluation status was adjusted for personal and family history of cancer in the model.

Results: Among 4,090 individuals meeting testing criteria for Lynch syndrome, only 1,118 (27.33%) had been evaluated. The mean age of the patients included was 49.88±15.45 and 3,173 (77.6%) were female (Table 1). Hispanic and non-Hispanic African Americans (OR=0.55 and OR=0.56, both P<.001 compared to non-Hispanic White) and individuals with Medicaid or Medicare insurance (OR=0.55, P<.001 and OR=0.56, P<.001 compared to commercial insurance) had been less likely evaluated (Table 2). In contrast, female patients (Odds ratio (OR) = 1.25, P=.01 vs males), and those with higher number of office visits in the prior three years (OR = 2.07, P<.001 in those with >38 visits) had more likely been evaluated.

Conclusion: The novel EHR-based algorithm identified many individuals that by current standard of care criteria should undergo genetic testing to rule out Lynch syndrome, yet they had not been identified by providers. Among those, less than a third had undergone testing, with the lowest numbers among the underserved and publicly insured populations. In spite that eligible Medicaid patients have genetic testing covered in Connecticut, these patients were still less likely to receive genetic testing to rule out Lynch syndrome than other patients cared for by the same health system. The use of tools as the one we developed can have an important impact in the identification of LS patients and can decisively contribute to decrease disparities.

Xavier Llor

Yale University School of Medicine

Appearances