Abstract

Significance to VA: Colorectal cancer (CRC) is the second leading cause of cancer mortality and the leading cause of cancer death among men under age 50. Screening for CRC can reduce the risk of CRC death by half. However, many Veterans do not complete guideline-recommended CRC screening. Screening rates vary across sociodemographic groups, especially race and ethnicity, and across VA facilities. This project aligns closely with several VA priorities, including improving precision oncology services; caring for Veterans with CRC, a presumptive condition for Gulf War era and post-9/11 Veterans under the PACT Act; and improving Veterans’ timely access to services like colonoscopy. Innovation and Impact: Our project links VA encounters, Community Care claims, and CMS claims to obtain a more complete view of CRC screening rates and disparities among Veterans than has been analyzed previously. Our analyses will contribute key missing information on referral rates for follow-up colonoscopies among Veterans with positive stool-based test results and isolate between- and within-facility contributions to overall disparities. In partnership with the National Gastroenterology and Hepatology Program, we will develop an equitable risk prediction algorithm to prioritize enhanced outreach to Veterans at highest risk for CRC. Specific Aims: This CDA will provide Dr. Nieser with mentoring and training to conduct the following aims: 1. Provide a detailed analysis of the quality and equity of screening across VA facilities. I will assess disparities in each step of CRC screening pathways and quantify how much overall disparities are driven by where patients receive care (between-facility differences) versus within-facility differences in care. 2. Qualitatively investigate patient-, provider, and system-level mechanisms explaining low screening rates and disparities in screening rates. I will interview primary care providers and patients at facilities with high disparities, to understand barriers to screening and group-specific drivers of low screening rates. 3. Develop an equitable risk prediction algorithm for colorectal neoplasia that VA can use to target screening outreach efforts to patients most likely to benefit. I will develop a risk prediction algorithm and assess and improve its algorithmic fairness (i.e., similar predictive performance across groups). Methodology: In Aim 1, using logistic regression models, I will describe racial/ethnic disparities in (a) provider initiation of screening, (b) receipt of screening through colonoscopy or return of stool test, and (c) referral and (d) receipt of follow-up colonoscopy among patients with positive stool test results. I will use the Kitagawa- Blinder-Oaxaca decomposition method to estimate how much disparities are explained by between-facility differences in quality. In Aim 2, I will interview primary care providers and patients at facilities with large disparities to gain a better understanding of the quantitative results from Aim 1. In Aim 3, I will fit a Cox proportional hazards regression model to predict risk of advanced colorectal neoplasia within 8 years, among average risk patients. I will address differences in model performance across racial and ethnic groups. Path to Translation/Implementation: Based on findings from Aims 1 and 2, I will develop and pilot a detailed quality monitoring report, enhancing the existing national CRC screening reminder reports. National and local stakeholders can use this report to track the quality and equity of each of the CRC screening pathway steps. Following Aim 3, I will work with operational partners (the National Gastroenterology and Hepatology Program office) to integrate the risk prediction algorithm we develop into existing national CRC screening reminder reports. We will conduct a hybrid type 1 effectiveness-implementation study of an equitable risk-stratified outreach intervention to improve CRC screening rates. More generally, this CDA will support my development into an independent VA researcher who uses innovative data science methods to inform and evaluate interventions that improve the health of all Veterans.

NIH Reporter Project Information: https://reporter.nih.gov/project-details/11343579


PUBLICATIONS:
None at this time.