Abstracts

Addressing Fecal Immunochemical Test Underutilization for Colorectal Cancer Screening in Asian Americans: Findings from a Patient-Centered Needs Assessment

Jeemin Kwon¹, Selena Zhou^1-2, Jessica J. Tuan⁴, Jennifer K. Bocek⁵, Jessica D. Calderon⁵, Sarah D. Meshkat⁵, Folasade P. May^1-4

¹UCLA Department of Internal Medicine, Los Angeles, CA, USA; ²Vatche and Tamar Manoukian Division of Digestive Diseases; UCLA David Geffen School of Medicine; ³Greater Los Angeles Veterans Affairs Health System, Los Angeles, CA, USA; ⁴UCLA Kaiser Permanente Center for Health Equity, Jonsson Comprehensive Cancer Center; ⁵UCLA Office of Population Health and Accountable Care

• Submission category: Health Services and Public Health Research; Health Care Delivery/Access/Quality OR Colorectal Neoplasia Screening; Non-Endoscopic
• Character limit: 2,900/2,900

Background: Colorectal cancer (CRC) screening rates are low among Asian Americans. In our medical center, a Mailed Fecal Immunochemical Test (FIT) Outreach Program was established in 2015 to increase screening participation, with only 16% uptake among Asian American patients. In this study, we aimed to identify the primary reasons why Asian American patients did not complete screening to inform future interventions.

Methods: The study setting is a large, urban academic health center with a defined primary care population and robust referral-based care. The Mailed FIT Outreach Program sends education, a FIT kit, and instructions in English to average-risk primary care patients overdue for CRC screening. Between 8/2024 and 1/2025, there were 334 Chinese, Japanese, or Korean patients who did not return a FIT. We contacted 201 (60.2%) via an electronic patient portal message (Epic MyChart), mailed letter, and/or phone call to complete a 12-item survey. Portal and mailed letters were in the patient’s documented preferred language; phone calls were conducted in English. We calculated frequencies and proportions for patient characteristics and survey item responses.

Results: The survey completion rate was 14.9% (30/201) (Table 1). The mean age of respondents was 57.5 years (sd=9.2), and 19 (63.3%) were female. Ten (33.3%) respondents were Chinese, 10 (33.3%) were Japanese, and 10 (33.3%) were Korean; 16 (53.3%) were born in the US. Twenty-one (70%) responded that CRC screening is “very important;” 21 (70%) reported knowing the age to start screening, of which 10 (47.6%) were correct; 27 (90%) reported knowing what a colon polyp is; and 20 (66.7%) reported prior screening (Table 2). Of those previously screened (20), colonoscopy was the most frequent method (16, 80%). Twenty-four (80%) recalled receiving the mailed FIT kit, of which 9 (37.5%) did not open the kit, 5 (20.8%) meant to complete the kit but did not, 4 (16.7%) did not want to complete the kit, and 3 (12.5%) misplaced the kit (Table 2). The most common reasons for not completing the kit were insufficient time (4,16.7%), felt up-to-date with screening (4,16.7%), and discomfort collecting stool (3,12.5%). The top three suggested ways to encourage FIT completion were an easier FIT kit (6, 25%), additional reminders (7, 29.2%), and more discussion with a medical provider (7, 29.2%). Respondents also provided free text responses indicating preference for colonoscopy and concerns about FIT accuracy (Table 2).

Conclusions: In this survey study of Asian American patients who did not complete a FIT, there was a mismatch between perceived importance of CRC screening and screening participation. Additional patient education on screening indications and FIT sensitivity, reminders, FIT kit redesign focused on user experience, or offering colonoscopy may increase participation in this patient demographic.

Table 1: Survey participant characteristics, n=30

Table 2: Survey items and responses

Area Deprivation Index is a Predictor of Colonoscopic Follow-up After Abnormal FIT in a Large United States Health System

Yousif Arif, MD^1-2; Alexandra Thomson, MD,MPH^1-2; Jessica Tuan, MPH⁴, Jayraan Badiee, MPH^1-2; Folasade P. May, MD, PhD, MPhil<sup>1-4

1</sup>Department of Medicine, David Geffen School of Medicine, UCLA Ronald Reagan Medical Center, University of California Los Angeles; ²Vatche and Tamar Manoukian Division of Digestive Diseases, Department of Medicine, David Geffen School of Medicine, University of California Los Angeles; ³Greater Los Angeles Veterans Affairs Healthcare System, Los Angeles, CA, USA; ⁴UCLA Kaiser Permanente Center for Health Equity, Jonsson Comprehensive Cancer Center

• Submission category: Colon; Clinical Science; Health Disparities
• Character count: 2876/2900 (including spaces)

Background: Patients with abnormal fecal immunochemical test (FIT) results require colonoscopic follow-up to complete colorectal cancer screening. Known predictors of follow-up include age, comorbidities, and health system practices (e.g., reminders, navigation). Less is known about how neighborhood-level deprivation may contribute. We aimed to evaluate the relationship between Area Deprivation Index (ADI), a validated composite measure for neighborhood-level disadvantage, and colonoscopic follow-up after abnormal FIT.

Methods: We conducted a retrospective cohort study in a large, multi-site academic health system in an urban setting. We included average-risk adults aged 45-75 with an abnormal FIT result between 7/1/2022 and 11/7/2024 and at least 12 months follow-up. The primary study outcome was colonoscopy completion within 6 months of abnormal FIT. Follow-up at 12 months was the secondary outcome. The primary independent variable was state-level ADI, a composite score derived from income, education, employment, and housing quality to provide a measure of a neighborhood's social and economic conditions. State ADI is categorized into deciles with the first decile (approx. ADI score 1-15) representing the least disadvantaged and the tenth decile (approx. ADI score 51-72) the most disadvantaged. Control variables included age, sex, race/ethnicity, and Elixhauser Comorbidity Index. We used descriptive statistics to summarize patient characteristics and performed multivariable logistic regression and Cox proportional hazards models to evaluate predictors of colonoscopic follow-up, with particular attention to the role of ADI.

Results: We identified 1,083 patients with an abnormal FIT result during the study period. Mean age was 63 years (IQR 54-74); 53% were male, and 31.5% were non-Hispanic White (Table). The median ADI decile was 3. Overall, 547 patients (50.5%) completed colonoscopy by 6 months, and 625 (53.8%) did by 12 months. In adjusted logistic regression, higher ADI decile (aOR 0.89, 95% CI 0.82–0.97) was significantly associated with lower odds of colonoscopy completion at 6 months. This association remained significant at 12 months (aOR 0.91, 95% CI0.84-0.98). In the adjusted Cox proportional hazards model, higher state ADI was a significant predictor of lower likelihood of colonoscopy completion over time (aHR 0.94, 95% CI 0.89–0.99) (Figure).

Conclusion: In our large, academic, urban health center, high neighborhood deprivation was associated with lower odds of colonoscopic follow-up at 6 months, 12 months, and overall, for patients with an abnormal FIT result. ADI may serve as a practical risk stratification tool to identify patients at highest risk of being lost to follow up after an abnormal FIT result and can guide targeted, equity-focused interventions to address structural barriers to colonoscopic follow-up.

Table: Baseline characteristics of the study cohort by colonoscopy completion status.

Figure: Adjusted probability of overall colonoscopy completion by state-level area deprivation index (ADI) deciles.

Colonoscopy Complications in Older Adults: A Systematic Review and Meta-Analysis

Nicole P. Mirabadi, BA; Divya P. Prajapati, MD; Yi Le, MS; Alec Zadikian, BA; Lin Liu, PhD; Karen Heskett, MSI; Folasade P. May, MD, PhD, MPhil; Samir Gupta, MD, MSCS

• Proposed category: Colon; Clinical Science; Colorectal Neoplasia Surveillance; Endoscopic

Introduction: Risks of colonoscopy in older adults may inform decision-making surrounding colorectal cancer (CRC) screening and surveillance. We performed a systematic review and meta-analysis to characterize colonoscopy complications in older adults.

Methods: We searched PubMed, Web of Science, CINAHAL, and Embase through 11/2025 with keywords relevant to colonoscopy complications in older adults to identify studies with data reported for adults age >65; studies were excluded if data for patients with known CRC or disease states predisposing to CRC could not be disaggregated from patients without elevated CRC risk. Complication rates for adults aged ≥65 and ≥80 years were pooled for meta-analysis using random-effects models of proportions. Freeman–Tukey double arcsine transformation was applied to stabilize variance due to low event counts, and pooled estimates were expressed as rates per 10,000 colonoscopies with 95% confidence intervals (CI). A random effect model was also used for comparing complications for individuals age < 65 vs. > 65 years, and relative risk for complications was estimated.

Results: Our search yielded 12,808 articles; 45 studies met inclusion criteria, and the number of studies included in meta-analysis of each outcome varied based on whether data specific to each outcome were reported (Figure 1). Mean reported age in studies of age ≥65 was 70.4 years with 47% female patients. Among adults age ≥65, the overall GI-related adverse event rate was 26.4/10,000 colonoscopies (95% CI: 10.2 – 48.8), with a perforation rate of 3.7/10,000 (95% CI: 0 – 11.7), and a GI bleeding rate of 26.3/10,000 (95% CI: 6.6 – 57.2; Figure 1). The overall non-GI adverse event rate was 133/10,000 (95% CI: 52.8 – 245). Mean reported age in studies of age ≥80 was 83.9 years with 61.3% female patients. Adults age ≥80 had an overall GI-related adverse event rate of 85.5/10,000 colonoscopies (95% CI: 30.3 – 162), perforation rate of 14/10,000 (95% CI: 0 – 48.8), GI bleeding rate of 71.7/10,000 (95% CI: 7.6 – 179), and overall non-GI adverse rate of 218/10,000 (95% CI: 113 – 351; Figure 2). Adults age ≥65 compared to <65 years had increased risk for GI-related adverse events (RR 2.0; 95% CI: 1.5 – 2.6), perforation (RR 2.6; 95% CI: 1.7 – 4.3) and GI bleeding (RR 1.8; 95% CI: 1.4 – 2.2); risk for non-GI adverse events approached significance (RR 1.55; 95% CI: 1.0 – 2.5).

Conclusion: Older adults are at increased risk of colonoscopy-associated complications compared to younger adults, with the highest rates observed among individuals age 80 and older. As the most comprehensive systematic review to-date of colonoscopy-associated complications in older adults, this analysis can inform patient-provider discussions as clinicians weigh benefits of CRC prevention against procedural risks in older adults.

Figure 1: Colonoscopy complications among patients age ≥65 years. 

Figure 2: Colonoscopy complications among patients age ≥80 years.

Colorectal Cancer Outcomes Among Disaggregated Asian Subgroups: A 2006-2020 SEER Analysis

Matthew Y. Zhao, MD¹ ([email protected]); Jayraan Badiee, MPH² ([email protected]) Folasade P. May, MD, PhD, MPhil^2-4 ([email protected])

¹Department of Medicine, Icahn School of Medicine at Mount Sinai, New York, New York; ²Vatche and Tamar Manoukian Division of Digestive Diseases, Department of Medicine, David Geffen School of Medicine, UCLA, Los Angeles, CA, USA; ³Greater Los Angeles Veterans Affairs Healthcare System, Los Angeles, CA, USA; ⁴UCLA Kaiser Permanente Center for Health Equity, Jonsson Comprehensive Cancer Center, UCLA, Los Angeles, CA, USA

• Disclosures: None to report
• Character count: 2867/2900
• Abstract due: December 4, 2025; 9PM EST 
• Abstract submission category: Colon; Clinical Science; Colorectal Cancer; Epidemiology and Diagnosis

Introduction: Despite varying colorectal cancer (CRC) outcomes across distinct Asian American subgroups, population-level analyses often assess Asian patients as a single aggregated category. We aimed to assess recent CRC outcomes among disaggregated Asian populations to better characterize subgroup-level heterogeneity.

Methods: Using National Cancer Institute (NCI) Surveillance, Epidemiology, and End Results (SEER) registry data from 2006 to 2020, we identified patients age ≥18 with a primary diagnosis of CRC. To assess outcomes among disaggregated Asian subgroups, we stratified patients into Chinese, Filipino, Japanese, Vietnamese, Korean, and Asian Indian/Pakistani. We also included a referent non-Hispanic (NH) White group. We identified predictors for regional or distant metastases at CRC diagnosis by multivariable logistic regression, as well as predictors for CRC-related mortality by multivariable Cox proportional hazards.

Results: Among 25,379 Asian patients, there were 6,585 (25.9%) Chinese, 6,399 (25.2%) Filipino, 4,309 (17.0%) Japanese, 3,047 (12.0%) Vietnamese, 2,964 (11.7%) Korean, and 2,075 (8.2%) Asian Indian/Pakistani patients included for analysis. We also included 234,948 NH White patients. Among all patients, median age was 66.0 (IQR 56.0-76.0), 53.0% were male, 59.2% had metastases at time of CRC diagnosis, and 27.9% suffered CRC-related mortality. Among disaggregated Asian subgroups, Vietnamese patients had the highest rates of metastasis at diagnosis (64.6%) and CRC-related mortality (28.2%) (Figure). In the multivariable logistic regression, Vietnamese (aOR 1.23, 95% CI 1.14,1.32), Korean (aOR 1.20, 95% CI 1.11,1.30), and Asian Indian/Pakistani (aOR 1.10, 95% CI 1.01,1.21) subgroups had increased odds of metastasis at time of diagnosis compared to NH White patients. Moreover, low socioeconomic status (SES) by Yost index quintile (YQ) was also associated with increased odds of metastatic disease (Table). In the multivariable Cox proportional hazards model, the Chinese (aHR 0.92, 95% CI 0.88,0.97) and Asian Indian/Pakistani (aHR 0.88, 95% CI 0.80,0.97) subgroups had decreased likelihood of mortality compared to NH White patients, whereas low SES and persistent census tract poverty predicted increased mortality risk (Table).

Conclusion: In this national cancer registry analysis, disaggregated Asian subgroups demonstrated heterogeneity in CRC outcomes. Vietnamese, Korean, and Asian Indian/Pakistani individuals had higher odds of presenting with metastatic disease, while Chinese and Asian Indian/Pakistani patients had lower mortality risk compared to White individuals with CRC. These findings highlight the importance of disaggregating Asian populations to accurately survey national CRC trends, identify groups at highest risk, and develop culturally relevant approaches to CRC prevention and management.

Figure: Survival curve by disaggregated Asian subgroup

Table: Predictors of regional or distant stage at colorectal diagnosis by multivariable logistic regression and colorectal cancer related death by multivariable cox proportional hazards

Increasing Incidence of Early-Onset Gastric Carcinoma in the United States: A Population-Based SEER Analysis by Histologic Subtype

Sigrid S. Young, MD¹ ([email protected]); Jayraan Badiee, MPH¹ ([email protected]); Folasade P. May, MD, PhD, MPhil^1-3 ([email protected])

¹Vatche and Tamar Manoukian Division of Digestive Diseases, Department of Medicine, David Geffen School of Medicine, UCLA, Los Angeles, CA, USA; ²Greater Los Angeles Veterans Affairs Healthcare System, Los Angeles, CA, USA; ³UCLA Kaiser Permanente Center for Health Equity, Jonsson Comprehensive Cancer Center, UCLA, Los Angeles, CA, USA

• Limit: 2,880/2900 characters including spaces
• Submission deadline: December 4, 2026;  6PM PST, 9PM EST
• Presentation type: Oral or poster
• Abstract track: Stomach & duodenum
• Category: Gastric Pre-Neoplastic and Neoplastic Conditions; Epidemiology and Diagnosis

Background: Gastric cancer (GC) is the fifth most common cancer and the fourth leading cause of cancer-related deaths worldwide. In the United States, routine endoscopic screening is not recommended for average-risk individuals, and most patients present with advanced disease. Prognosis varies substantially by histology, with 5-year survival of 36% for gastric adenocarcinoma (GA) versus 81% for gastric neuroendocrine tumors (GNET). Given rising gastrointestinal cancer rates in young adults, we aimed to evaluate temporal trends in GC incidence by histologic subtype and age.

Methods: We analyzed 2004–2022 data from the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) registry. We included patients (1) aged ≥15 years, (2) diagnosed with a primary gastric malignancy, (3) with histologically confirmed GA, GNET, signet ring cell carcinoma (SRCC), or other invasive carcinoma, and (4) no concurrent or prior cancer diagnoses. Age-standardized incidence rates (ASIR) and annual percentage change (APC) were calculated and stratified by age group (15–49 vs. ≥50 years) and GC subtype. Joinpoint regression was used to assess trends and statistical significance.

Results: There were 137,641 people with GC (62% male; median age 66.0 [IQR 56.0–76.0]; 86.5% ≥50 years). Most had GA (71.7%), followed by SRCC (17.7%), GNET (7.3%), and other invasive carcinoma (3.4%). Overall, 49.8% were non-Hispanic White, 23.6% Hispanic, 13.4% non-Hispanic Black, 12.5% non-Hispanic Asian/Pacific Islander, and 0.7% non-Hispanic American Indian/Alaska Native. For all ages combined, incidence declined for GC, GA, SRCC, and other gastric carcinoma (APC −1.30, p<0.001; APC −1.30, p<0.001; APC −2.0, p=0.03; APC −15.8, p=0.004), while it increased for GNET (APC 3.90, p<0.001). GA incidence decreased in people age ≥50 years (APC −1.56, p<0.0001) but increased for people age <50 (APC +1.01, p=0.03). Overall GC incidence declined in patients age ≥50 (APC −1.67, p<0.0001) but rose modestly in those age <50 (APC +0.60, p=0.057). In both age cohorts, SRCC incidence declined, while GNET incidence increased over time (Table, Figure).

Conclusion: Despite overall declines in GC incidence and sustained decreases among people age ≥50 years, incidence is rising in people under age 50, driven primarily by increasing GA2 incidence in this younger cohort. GNET incidence has also increased across all ages since 2004. These age- and histology-specific shifts demonstrate that the emerging burden of early-onset gastrointestinal cancers includes GC and is not uniform across subtypes. The growing incidence of GA in younger adults, in contrast to declines in older populations, suggests an evolving epidemiologic pattern. Further research is needed to identify the drivers underlying these selective increases in order to address the growing burden of early-onset disease.

Table: Annual percentage change (APC) in gastric cancer incidence by gastric cancer subtype and age cohort

Figure: Age-standardized incidence rate over time by gastric cancer subtype for a) all patients (n=137,641); b) people age 15-49 (n=18,527); and c) people age 50 and older (n=119,114)

A Decade of Massive Patient Population Growth but Minimal Improvements in Colorectal Cancer Screening Participation in United States Federally Qualified Health Centers from 2014 to 2024

Megan R.M. Aaronson, MD, MS¹ ([email protected]); Matthew Y. Zhao, MD² ([email protected]); Jayraan Badiee, MPH³ ([email protected]) Folasade P. May, MD, PhD, MPhil^3-5 ([email protected]) 

¹Department of Medicine, David Geffen School of Medicine, UCLA, Los Angeles, CA, USA; ²Department of Medicine, Icahn School of Medicine at Mount Sinai, New York, New York; ³Vatche and Tamar Manoukian Division of Digestive Diseases, Department of Medicine, David Geffen School of Medicine, UCLA, Los Angeles, CA, USA; ⁴Greater Los Angeles Veterans Affairs Healthcare System, Los Angeles, CA, USA; ⁵UCLA Kaiser Permanente Center for Health Equity, Jonsson Comprehensive Cancer Center, UCLA, Los Angeles, CA, USA

• Disclosures: None to report
• Character count: 2873/2900
• Abstract due: December 4, 2025; 9PM EST
• Submission category: Colon; Clinical Science; Health Disparities

Introduction: United States (US) Federally Qualified Health Centers (FQHCs) provide primary care services, including colorectal cancer (CRC) screening, for millions of Americans regardless of their ability to pay. Populations served in these settings have suboptimal CRC screening rates and outcomes. We aimed to evaluate 10-year trends in FQHC characteristics and CRC screening participation.

Methods: We used the Uniform Data System (UDS), the only source for national FQHC quality data, to perform cross-sectional analyses of CRC screening rates for all US FQHCs. We abstracted CRC screening rates for each FQHC from 2014 to 2024 and FQHC characteristics in the years 2014 and 2024, including the percentage of patients who were: age 45-74, uninsured, non-English speaking, >200% of the federal poverty level (FPL), and experiencing homelessness. We grouped FQHCs by majority race/ethnicity served (non-Hispanic [NH] White, NH Black, Hispanic/Latino, NH Asian, NH Other), urban/rural status, and US region (West, South, Midwest, Northeast). We used paired Wilcoxon signed-rank tests to compare FQHC characteristics and screening rates for all FQHCs in 2014 and 2024. We also calculated percent change in FQHC characteristics from 2014 to 2024 and performed linear regression to evaluate the relationship between FQHC characteristics and change in CRC screening participation over time.

Results: There were 7,430,162 patients eligible for CRC screening at 1,271 US FQHCs in 2024, compared to 3,936,450 patients in 1,192 FQHCs in 2014 (Table). FQHC patient-mix characteristics changed significantly from 2014 to 2024 (Table). The median CRC screening rate for all FQHCs was 30.0% in 2014 and 40.8% in 2024 (Figure 1A). Screening participation increased from 2014 to 2024 regardless of the majority race/ethnicity served, US region, and urban/rural status (Figure 1B, 1C & 1D). In our adjusted model, FQHCs in states that adopted Medicaid expansion (coef 4.05, 95% 1.02 - 7.08) and FQHCs with a greater percentage of patients with income >200% FPL (coef 0.16, 95% CI 0.02 - 0.30) experienced a significant increase in CRC screening participation from 2014 to 2024. Conversely, FQHCs with a high percentage of uninsured patients (coef -0.22, 95% CI -0.37 - -0.08) saw a significant decrease in screening participation over time (data not shown).

Discussion: The number of age-eligible patients served by FQHCs nearly doubled from 2014 to 2024, and FQHCs became more diverse. Despite this population growth, gains in CRC screening participation have slowed and remain 2 below pre–COVID-19 levels. Our findings suggest that structural factors, including Medicaid expansion status and socioeconomic composition, impact a FQHC’s capacity to screen patients. Sustained policy support for FQHCs is essential to advance progress and ensure equitable CRC screening on a national scale.

Table: Characteristics of United States federally qualified health centers in 2014 and 2024

Figure 1: Trends in median colorectal cancer screening rates from 2014-2024 for (A) All U.S. FQHCs; (B) FAHCs by race/ethnicity of majority of patients served; (C) FQHCs by U.S. region; and (D) FQHCs by urban/rural status; (n=1,271)

Population Health Outreach with a Shared Decision Aid Increases Uptake of Colorectal Cancer Screening for Average-Risk Adults

Thomson, Alexandra E., MD, MPH^1-2, Tuan, Jessica J., MPH³, Fujimoto, Jeffrey MD, MBA¹, Smith, Laurie N., MPH⁴, Badiee, Jayraan, MPH², May, Folasade P., MD, PhD, MPhil^1,2,3,5

¹Department of Medicine, David Geffen School of Medicine, UCLA, Los Angeles, CA, USA; ²Vatche and Tamar Manoukian Division of Digestive Diseases, Department of Medicine, David Geffen School of Medicine, UCLA, Los Angeles, CA, USA; ³UCLA Kaiser Permanente Center for Health Equity, Jonsson Comprehensive Cancer Center, UCLA, Los Angeles, CA, USA; ⁴WiserCare, Los Angeles, CA, USA; ⁵Greater Los Angeles Veterans Affairs Healthcare System, Los Angeles, CA, USA

• Disclosures: None to report
• Funding: Institutional, Melvin and Bren Simon Digestive Diseases Center
• Submission category: Practice management: clinical science: Performance Metrics, Process Improvement, Quality Improvement and Implementation
• Character count: 2895/2900

Introduction: Population health outreach can increase colorectal cancer (CRC) screening participation within health systems. The WiserCare (WC) CRC screening module is an electronic shared decision aid that facilitates patient-provider discussions about CRC screening test options to promote screening participation. We evaluated the impact of the module on screening uptake in a large United States health system.

Methods: The study setting is an urban, tertiary care health system with over 300 primary care providers, 67 clinics, and 5 endoscopy sites. In 2/2023, the health system launched My Action Plan (MAP), an EHR-based population health outreach program that sends electronic education and shared decision aids to patients overdue for preventive care measures. For primary care patients overdue for CRC screening, MAP sends the WC CRC screening module, which helps patients select a screening modality that meets their risk profile and personal preferences (fecal immunochemical test [FIT], colonoscopy, CT colonography, sigmoidoscopy, or stool DNA). Patients who remain unscreened are mailed a FIT. For this evaluation, we included primary care patients age 45 to 75 at average-risk for CRC and overdue for screening between 2/1/2023 and 7/31/2024. The primary outcome was CRC screening participation at 9 months (any modality), We compared screening rates for patients who (1) completed the WC module (group 1), (2) did not access the WC module (group 2), and (3) accessed but did not complete the WC module (group 3). We used Kruskal Wallis, chi-square tests, and pairwise comparisons to summarize patient characteristics and compare screening completion in the 3 groups.

Results: Among the 36,856 patients sent the WC module during the study period, median age was 54.1 years (IQR 49.1-62.1), 54% were female, and 49% were non-Hispanic White (Table). Overall, 9.5% accessed the module. Of patients who completed the module (n=1,958), 56% were female, 54% were non-Hispanic White, and 79% were privately insured (Table). Screening participation in group 1 was 28.7%, compared to 18.5% in group 2 and 22.2% in group 3 (all pairwise p<0.001). For group 1 and group 3, colonoscopy was most common (49.4% and 52.5%, respectively), followed by FIT (45.5% and 30.3%, respectively). For group 2 patients, FIT (60.3%) was followed by colonoscopy (30.5%) (Figure).

Conclusions: In this population health outreach program, engagement with the shared decision aid was low, but screening participation was significantly higher among patients who at least accessed the module. Patients who accessed the module were more likely to complete colonoscopy than FIT. The findings support that population outreach through shared decision aids can increase CRC screening participation independent of clinical visits, offloading primary care providers and increasing health system quality performance metrics.

Table: Patient characteristics and study outcomes by CRC screening module completion status; N=36,856

Figure: Screening participation and modalities chosen for each study group and overall at 9 months; n=7,078

Provider Perspectives on Colorectal Cancer Surveillance Among Veterans Age 75 and Older: The SurvOlderAdults (Surveillance Colonoscopy in Older Adults) Study

Neetu Chawla, PhD; Tamra Burns Loeb, PhD; Samir Gupta, MD; Lin Liu, PhD; Joshua B Demb, PhD; Jiyue Qin, MS; Ashley Earles, MPH; Mark Lamm; Benjamin J Seligman, MD, PhD; Folasade P May, MD, PhD, MPhil

• Characters: 2858 (limit 2900) 

Background: Older adults (age75) and their providers face challenging decisions regarding whether to continue or forgo post-polypectomy colonoscopic surveillance, given the need to balance cancer detection and prevention with procedure risks and competing healthcare priorities. We conducted qualitative interviews with providers in the Veterans Health Administration (VA) to assess perspectives on surveillance decision-making for older adults.  

Methods: We conducted 24 semi-structured interviews with primary care providers (PCPs), gastroenterologists (GIs), and geriatricians from two VA sites after reaching out to 90 total providers (26.7% response). Interviews were structured using domains from a multi-level framework modelled after the Socio-Ecological framework, including factors at the patient-, provider-, system-, and policy-levels. We used a rapid analysis approach to identify themes according to the multi-level framework.

Results: Our sample included 12 PCPs, 8 GIs, and 4 Geriatricians. Providers discussed the following patient-level factors when considering surveillance for older adults: age, life expectancy, health status/comorbidities, patient preferences, and identifying the highest priority for patients’ overall health needs. Key provider-level themes were that: many follow GI provider recommendations; some factor in whether patients would pursue treatment if tumors were found through surveillance; some heavily discuss patient preferences and make decisions based upon individual circumstances; and differences existed in attitudes towards whether to continue surveillance based on provider type, with geriatricians being more likely to note that some 75 year olds “can run a marathon” while others are “bed bound.” Providers also brought up factors related to the healthcare system, such as the use of reminders, with some finding them “helpful” and others noting they are “good in principle but have their limitations.” Providers had differing impressions about continuing surveillance colonoscopies for older patients within VA, with some noting “poor resource utilization” and needing “a way to screen patients to see if they are appropriate or not,” while others said that continuation was “the status quo” but it would “be nice to define who does not need it based on individual risk.”

Conclusions: VA providers identified several factors at the patient-, provider-, and health system-levels that influence surveillance colonoscopy decision-making for older adults in the VA. We identified notable differences in perspectives about the importance of patient age, reliance on GI recommendations, and resource utilization for continuing surveillance. Future research should assess patient-level attitudes towards CRC surveillance in older adults to understand the alignment between provider and patient perspectives.

Figure 1: Adapted version of Steve Taplin's multi-level framework for CRC surveillance among older Veterans

Table 1: Exemplar quotes from VA providers by interview domain

Root-Cause Analysis of Post-Endoscopy Esophageal Neoplasia at the Veterans Health Administration

Zainab Aziz; Rahul Karna; Mohammad Bilal; Mei Leng; Anders Westanmo ([email protected]); Jason A. Dominitz; Samir Gupta; Folasade May; David Katzka; Nicholas J. Shaheen; Sachin Wani; Jennifer M. Kolb

Background: Surveillance endoscopy is recommended in Barrett’s esophagus (BE) for early neoplasia detection; however current practices have not diminished death from esophageal adenocarcinoma (EAC). Post-endoscopy esophageal neoplasia (PEEN), defined as high-grade dysplasia (HGD) or EAC detected after an apparently negative exam and before the next appropriate surveillance interval, may indicate suboptimal surveillance or missed lesions. The specific contributors to PEEN are not well described.

Methods: We conducted a root-cause analysis to identify drivers of PEEN in a national cohort of BE patients from the Veterans Health Administration between 2010-2024. Veterans with an index endoscopy of non-dysplastic BE or indefinite for dysplasia and a subsequent diagnosis of HGD or EAC between 6 months to 3 years later were included. A panel of three BE experts reviewed primary source records and independently rated examination quality and follow-up recommendations (Figure). Index endoscopy was deemed adequate according to the following criteria 1) appropriate landmark identification, 2) use of virtual chromoendoscopy, and 3) use of Seattle biopsy protocol. A more lenient definition of adequacy excluded chromoendoscopy.

Results: 18 PEEN cases were identified (10 HGD, 8 EAC) with a mean time to diagnosis of 1.6 years (Table). 5 of the cancers were advanced stage T2/T3 and required chemotherapy and radiation. On index endoscopy, landmarks were appropriately identified in 59% of cases, chromoendoscopy use was documented infrequently (17%) and Seattle protocol biopsy was used less than half the time (44%). In 3 of the 18 cases a visible lesion was identified- 1 had inadequate sampling, and 2 had adequate sampling with possible causes for PEEN as inappropriate follow-up or possible additional missed lesion. For the remaining 15 cases without a visible lesion, the majority had inadequate endoscopy (14, 78%) with a possible explanation being a missed lesion. There was one case with panel disagreement over adequacy of exam; inappropriate follow-up was the likely cause of PEEN. Using less stringent criteria (landmarks + Seattle), 10 cases (56%) were determined to have inadequate examination, 1 adequate (6%), and there was disagreement over adequacy for the other 4 cases (22%).

Conclusion: Among a national cohort of Veterans with BE, root-cause analysis with expert panel adjudication of PEEN cases demonstrated suboptimal quality of endoscopic examination. Inadequate description of BE landmarks, with no demarcation of length or Prague classification and hiatal hernia, and limited use of the Seattle protocol and chromoendoscopy may have resulted in missed lesions. These results can inform intervention strategies, including use of standardized reporting systems, tissue sampling protocols, and optimized technology for neoplasia detection.

Figure 1: Root-cause analysis workflow for expert panel. While there may be other causes of PEEN following an adequate endoscopy (such as possible missed lesion or inadequate pathology read), in our cohort there was only one case, with inappropriate follow up recommended. PEEN=post-endoscopy esophageal neoplasia; HDG=high-grade dysplasia; EAC=esophageal adenocarcinoma.

Table 1: Characteristics of index endoscopies and cancer and dysplasia outcomes. BE=Barrett’s esophagus; GI= gastrointestinal; MAC=monitored anesthesia care; HGD= high-grade dysplasia; EAC=esophageal adenocarcinoma; AJCC=American Joint Committee on Cancer