3097 — Overuse and Underuse of Colonoscopy for Screening and Surveillance in the VA Healthcare System
Fisher DA, Durham VAMC; Grubber J, Durham VAMC; Murphy C, UNC Chapel Hill; Johnson MR, Durham VAMC; Sandler RS, UNC Chapel Hill; Provenzale D, Durham VAMC;
Overuse of colonoscopy is expensive and reduces capacity. Underuse is a missed opportunity for colorectal cancer (CRC) detection and prevention. Our objectives were to 1) examine physician non-adherence to colonoscopy interval guidelines 2) examine patient receipt of follow-up colonoscopy and potential correlates.
Data were abstracted from the VA electronic medical record for 2,443 patients across 25 VA facilities. Facilities were randomly selected among 85 qualifying facilities stratified by academic affiliation, geographic region, and resource level. Patients were randomly selected among those aged 50-64 who underwent colonoscopy at each VA in fiscal year (FY) 2008 and had no colonoscopy in the prior 10. Patients with incomplete colonoscopies, inadequate bowel preparations (prep), or ambiguous CRC risk were excluded (n = 988). Physician non-adherence was a recommendation that did not match the guideline follow-up interval. We examined four CRC risk groups, defined by index colonoscopy result: normal, hyperplastic polyp, low risk adenoma (1-2 adenomas < 1cm), and high risk adenoma (3-10 adenomas or any adenoma > = 1 cm or high grade dysplasia). Physician cluster-adjusted logistic regression models were created to estimate the association between non-adherence and potential factors. Then patients were followed in VA administrative claims data through FY 2014 to determine if a colonoscopy was performed. Guideline adherent follow-up was calculated for three groups: due in 10 years (no adenoma = normal or hyperplastic polyp), due in 5-10 years (low risk adenoma), and due in 3 years (high risk adenoma). Adjusted logistic regression models were created to estimate the association between colonoscopy use and potential predictors.
Physician recommendation non-adherence rates were 37% (532/1455) overall, 26% (236/893) for normal colonoscopy, 61% (123/203) for hyperplastic polyps, 46% (107/231) for low risk adenomas, and 52% (66/128) for high risk adenomas. The non-adherent recommendation was too soon for 100% of the normal, hyperplastic and low risk adenoma patients and for 86% of the high risk adenoma patients. Adjusted models indicated CRC risk group, prep quality and region were associated with non-adherence. In cumulative follow-up, 19% of patients with no adenomas, 32% with low risk adenomas, and 27% with high risk adenomas underwent colonoscopy too early. The majority of high risk patients (55%) underwent colonoscopy too late or had not undergone colonoscopy. In the no adenoma and low risk groups, overuse was associated with non-adherent physician recommendation at the time of index colonoscopy and treatment in a non-academic facility. None of the examined patient, physician or facility factors were associated with colonoscopy underuse in the high risk group.
Physician recommended intervals for polyp follow-up were frequently shorter than guidelines and varied by region and prep quality. The physician recommendation was associated with overuse of colonoscopy for low risk patients in the 6 years of follow-up. Underuse of colonoscopy for high risk patients was high with no significant associations among the factors examined.
VA colonoscopy capacity may be increased by reducing overuse for polyp surveillance. The complex picture of overuse and underuse supports the need for better tracking of patients by clinical risk.