Closing a Gap in Colorectal Cancer Screening
Jennifer K. Maratt,David A. Leiman,Thomas F. Imperiale
DOI: https://doi.org/10.1001/jamanetworkopen.2024.2652
2024-03-26
JAMA Network Open
Abstract:Colorectal cancer (CRC) is one of the most prevalent, but highly preventable, cancers in the US. Much progress has been made in reducing CRC incidence and mortality; however, gaps in CRC screening remain, arguably the most critical of which is timely follow-up after positive results of stool-based tests (SBT), such as the fecal immunochemical test (FIT) and FIT-DNA test. The yield of a positive SBT result for advanced neoplasia (the combination of CRC and advanced, precancerous polyps) is in the 20% to 35% range, 1 representing one of the highest yields for advanced neoplasia of any indication for colonoscopy, 2 yet rates for colonoscopy completion in this context are variable, reported as low as 18% in some centers. 3 Delaying or foregoing this second step of screening is associated with more advanced-stage disease and doubled 10-year CRC-related mortality. 4 In this study, Ciemins and colleagues 5 describe a novel CRC screening completion measure to quantify this gap in care. Ciemins and colleagues 5 used deidentified electronic health record (EHR) data from OptumLabs Data Warehouse, which comprise demographic, clinical, and utilization and visit data for individuals from more than 50 health care organizations (HCOs). The data warehouse also contains administrative claims data for approximately 10% of individuals with EHR data. Ciemins and colleagues 5 defined their measure as the proportion of individuals who received a colonoscopy within 180 days of a positive SBT result (numerator) out of the total number of adults aged 50 to 75 years with an abnormal CRC screening SBT result within the measurement year (denominator). Patients with a history of CRC, total colectomy, and those receiving palliative or hospice care were excluded, as were those who had underwent SBT for diagnostic purposes or in the inpatient or emergency care setting. Feasibility testing was performed by using the National Quality Forum's Feasibility Scorecard to evaluate the capability of 3 HCOs, each using a different EHR vendor, to capture data elements required to calculate this measure. Reliability testing was performed across 38 HCOs in 5 measurement years (2016-2020). While the process of measuring face validity was not explicitly described, it involved content expert advisors representing 2 HCOs and 3 national societies. 5 Among 20 581 individuals across 38 HCOs who met inclusion criteria and had a positive SBT result in 2018, a median (IQR) of 47.9% (37.4%-53.2%) had a colonoscopy within 6 months, with a median (IQR) follow-up of 53 (28-115) days. Performance rates for this measure ranged from 13% to 70% across HCOs. 5 Patients with Medicare or Medicaid insurance had lower colonoscopy completion rates within 6 months after a positive SBT result, at 49.2% of Medicare patients and 38.6% of Medicaid patients, compared with 50.9% of commercially insured patients. Black (37.1%; 95% CI, 34.6%-39.5%) and Hispanic (38.4%; 95% CI, 34.6%-42.1%) individuals had similar colonoscopy completion rates, which were lower than those of White individuals (49.0%; 95% CI, 48.2%-49.7%). 5 Based on a subset of 2164 individuals for which both EHR and claims data were available, 2 sensitivity analyses were performed. The first of these found that 59.9% (95% CI, 57.9%-62%) of individuals with claims and EHR data had a follow-up colonoscopy, compared with 51.3% (95% CI 49.2%-53.4%) of individuals with EHR data alone, resulting in missed capture of 14% of patients who had colonoscopy within 180 days. 5 The second sensitivity analysis included measuring performance rates at 90 days and showed that colonoscopy was completed by a median (IQR) of 39.7% (28.7%-44.0%) of individuals with positive SBT results. 5 Findings from this large, geographically diverse study by Ciemins and colleagues 5 highlight the suboptimal rates of timely follow-up colonoscopy after positive SBT results and draw attention to the significant variation in care across HCOs. Also notable is the disparity in CRC screening completion related to race, ethnicity, and insurance status. As acknowledged by Ciemins and colleagues, 5 there are limitations to this novel measure that affect feasibility (all initially specified exclusion criteria were unable to be captured in field testing), validity (the 14% underestimate requires further development and testing), and generalizability beyond HCOs included in the data warehouse. Evaluating measure performance among individuals at average risk for CRC, for whom SBTs are intended, and for individuals aged 45 to 49 years would provide additional useful context. We commend Ciemins and colleagues 5 for focusing attention on a meaningful approach to measuring high-quality CRC screening and providing guidance for standardized measurement. Several questions arise -Abstract Truncated-
medicine, general & internal