Improving Colonoscopy Quality for Colorectal Cancer Screening in the National VA Healthcare System

Overview

High quality screening colonoscopy is critical for colorectal cancer (CRC) prevention in Veterans. There is significant variability in colonoscopy quality in VA that is directly linked to differences in CRC incidence and death. The investigators developed the VA Endoscopy Quality Improvement Program (VA-EQuIP) that the National GI program office will implement using centralized quality measurement and reporting for adenoma detection rates (ADR), bi-annual audit and feedback with provider benchmarking to local and national performance, and collaborative learning to support colonoscopy quality improvement. Using a cluster randomized controlled trial, the investigators will study the implementation of VA-EQuIP and determine the efficacy of its intervention on adenoma detection rates, which are directly linked to CRC incidence and death.

Colorectal cancer (CRC) prevention is a top VA priority. Veterans are highly impacted by CRC, the third most common cancer diagnosed with a 3-year mortality rate of 35%. Colonoscopy screening can save lives, and is the cornerstone of effective prevention. The National Polyp Study showed that removal of adenomas during screening colonoscopy decreases the subsequent development of CRC by up to 90% and death by up to 50% in comparison to historical controls. In VA, colonoscopy is the primary CRC screening modality with over 200,000 colonoscopies performed each year. Eighty-two percent of Veterans are up-to-date on CRC screening, with 89% having a colonoscopy. However, the quality of those VA colonoscopies is unknown. High quality colonoscopy is critical for CRC prevention. Observational studies found significant rates of CRC even after normal colonoscopy, likely due to colonoscopists' varied performance in detecting and removing polyps. Among Medicare beneficiaries, approximately 7% of all CRC occurred in individuals who had colonoscopy that apparently missed the CRC diagnosis. A pooled analysis from eight surveillance studies suggested 52% of incident CRCs after colonoscopy were due to missed lesions. These interval cancers highlight the need to focus on colonoscopy quality indicators. Adenoma detection rate (ADR) significantly varies by provider and has been strongly linked to both CRC incidence and mortality. In essence, polyps that are not found cannot be removed for CRC prevention. ADR, defined as the proportion of screening colonoscopies performed by a physician that has one or more histologically-confirmed adenomatous polyps or CRC, is the primary benchmark for colonoscopy inspection quality. Increasing ADR correlates with a lower risk of interval and fatal CRC. A landmark US study of 314,872 colonoscopy exams showed marked variability in providers' ability to detect polyps, with the ADR ranging from 7.4 to 52.5%. In patients of providers with the highest ADRs, as compared with those in the lowest, the adjusted hazard ratios (HR) for fatal interval CRC was 0.38 (95% Confidence Interval \[CI\], 0.22 to 0.65). Each 1% increase in ADR was associated with a decrease risk of 3% for interval CRC and 5% for CRC death. Using the methods the team developed, the investigators have powerful pilot results showing colonoscopy quality variability in the VA and direct association with CRC death. From 1999-2011, there were 634,331 Veterans with baseline colonoscopy performed by one of 3101 colonoscopists with ADRs ranging from 13-79%. In patients who had a colonoscopy without polyps found, the higher the provider ADR, the more protected the patient was from developing future CRC: adjusted HR (adjHR) 0.57 for incident (95%CI: 0.42-0.79; ptrend\< 0.001) and 0.73 for fatal (95%CI: 0.50-1.06; p=0.047) CRC for the highest vs. lowest ADR quintile. Among individuals who had a colonoscopy with polypectomy, increasing ADR was also associated with reduced risk: adjHRs 0.46 for incident (95%CI: 0.34-0.62; ptrend\< 0.001) and 0.53 for fatal (95%CI: 0.34-0.84; ptrend=0.006) CRC for the highest vs. lowest quintile. A 5% absolute increase in ADR was associated with relative reductions in fatal CRC risk of 4% after normal colonoscopy (adjHR 0.96, 95%CI: 0.95-0.97) and 6% after polypectomy (adjHR 0.94; 95%CI: 0.89-0.98) Other important quality metrics, including bowel preparation quality and cecal intubation rate also impact patient outcomes with poor bowel preparation and incomplete examinations associated with missed lesions and need for earlier repeat procedures. Audit and feedback of colonoscopy quality improves performance and outcomes. The VA Surgical Quality Improvement Program (VASQIP) is an example of a VA nationwide reporting system that formed to focus on quality and has led to a decrease in mortality of 47% in patients 30 days after surgery. VASQIP provides information to providers for self-assessment and quality improvement purposes. There is similarly strong evidence that providing colonoscopy performance feedback improves quality and patient outcomes. Kahi and colleagues showed in a single-center VA setting that a quarterly report card improved colonoscopy quality. Most recently, in a large prospective European cohort study evaluating annual feedback and quality benchmark indicators on screening colonoscopy performance, the majority of the endoscopists (74.5%) increased their annual ADR. Moreover, individuals examined by endoscopists in the highest ADR quintile (\> 24.6%) had significantly lower risk of interval CRC and death. When compared with no increase in ADR, reaching or maintaining the highest quintile ADR decreased adjusted hazard ratios for interval and fatal CRC to 0.27 (95% CI, 0.12-0.63; p = .03), and 0.18 (95% CI, 0.06-0.56; p = .03). Quality Gap: Quality reporting improves patient outcomes. The national VA healthcare system currently lacks a reliable, efficient program to routinely measure and report colonoscopy quality. Accurate measurement of quality metrics is challenging, because validated quality metrics are not available in structured VA data from VA Corporate Data Warehouse (CDW). The challenge is multifactorial. Colonoscopy procedure documentation resides in text notes in Vista/ Computerized Patient Record System (CPRS) or endoscopic reporting software. The commonly used VA endoscopy note writer software programs (i.e. Endopro®, Provation®, etc) do not facilitate tracking pathology data and quality measurement. None of the current VA endoscopy reporting programs link to VA pathology data (to determine ADR); and production-level pathology data is not in the CDW. Thus, within VA, there has not been a reliable, efficient way to track critical procedure and pathology results to measure colonoscopy quality and ensure optimal protection from CRC incidence and death for Veterans. Proposed Solution and Rationale for Study: The foundation of VA-EQuIP is a novel natural language processing (NLP) algorithm and informatics reporting infrastructure that the investigators developed to measure the quality of colonoscopies performed by individual VA colonoscopists. The team will evaluate a national quality improvement program to better understand how and why such a program is effective or ineffective in different contexts. This study is critically needed to allow us to perform a rigorous evaluation to answer implementation science questions and to advance the clinical and epidemiologic knowledge on the impact of colonoscopy quality on CRC mortality. Leveraging the informatics investment made by VA Health Services Research and Development Service (HSR\&D) as well as the National Gastroenterology Program office partnership in quality improvement initiatives, the roll out of VA-EQuIP is a rare opportunity to study, in real time, the deployment and impact of a large-scale learning health system initiative. Most existing performance measures are calculated at the facility level. The concept of calculating metrics from individual provider procedures and comparing these metrics not only with local peers but with national providers represents a profound culture shift. This is where implementation science becomes essential, especially with respect to stakeholder engagement, adaptation and tailoring. A primary justification for this study is that VA-EQuIP has a high probability of improving a quality metric outcome (ADR) directly associated with CRC death, one of the most common cancers in Veterans. A randomized controlled trial of this magnitude is an opportunity to show, for the first time, that even small improvements in clinical performance from audit and feedback can save Veterans lives, since even a 1% increase in ADR translates to a 3% reduction in interval cancer death. The proposal for a randomized program evaluation is a tremendous opportunity to determine the large scale effect of the audit and feedback strategy on changes in the colonoscopy quality metrics of individual endoscopists over time. The evaluation of VA-EQuIP implementation will identify factors associated with effective implementation and colonoscopy quality improvement (impact) at VA sites. Prior and planned work builds toward long-term goal to reduce mortality in Veterans by increasing early detection of CRC and inform national quality improvement initiatives such as remediation training for continual low performing endoscopists. Finally, with upcoming expansion to community care, VA will be making policy decisions about what services should be furnished by VA versus non-VA providers. It is more crucial than ever that health services researchers apply rigorous methods to identify meaningful quality metrics that are linked by evidence to patient outcomes so that these measures of quality will be used to justify such crucial "make versus buy" determinations. It is hopeful that the investigators study will not only improve colonoscopy quality but also provide a framework for assessing the quality of other procedures such as bone marrow biopsy or bronchoscopy.