4008 — Overuse of Repeat Upper Endoscopy in the Veterans Health Administration

Lead/Presenter: Joel Rubenstein, COIN - Ann Arbor
All Authors: Rubenstein JH (Ann Arbor VA) Pohl H (White River Junction VA) Adams MA (Ann Arbor VA) Kerr E (Ann Arbor VA) Holleman R (Ann Arbor VA) Vijan S (Ann Arbor VA) Dominitz JA (Puget Sound VA) Inadomi JM (University of Washington) Provenzale D (Durham VA) Francis J (Veterans Health Administration, DC)

Objectives:
Americans undergo approximately 7 million esophagogastroduodenoscopies (EGDs) annually, and among Medicare beneficiaries, one third undergo a repeat EGD within 3 years. As many as 43% of those repeat EGDs are inappropriate. We aimed to determine the rate of repeat inappropriate EGD within the Veterans Health Administration (VHA), and identify factors associated with repeat EGD.

Methods:
We conducted retrospective analyses of Veterans undergoing an index EGD at 159 VHA facilities between 1/1/2003 and 6/30/2007. We excluded Veterans without regular use of VHA for healthcare and 5 years follow-up. Appropriateness of repeat EGDs was classified based on diagnostic and procedure codes into 3 categories: Likely Appropriate, Possible Overuse, and Probable Overuse. The electronic algorithm was validated against manual chart review of 150 randomly selected cases of index EGD at the Ann Arbor VA. The proportion of repeat EGDs in each category was tabulated. Multi-level logistic regression was performed to estimate the impact of patient-level and site-level factors on the odds of repeat EGD.

Results:
Compared to review of electronic medical records, the electronic algorithm demonstrated 99% specificity for EGD, 92% specificity for Probable Overuse, and 97% specificity for either Possible or Probable Overuse. Of 235,855 included Veterans, 85,690 (36.3%) underwent a repeat EGD within 5 years. 42,412 (49.5%) of the repeat EGDs were Likely Appropriate; 35,503 (41.4%) represented Possible Overuse, and 7,756 (9.1%) represented Probable Overuse. Patients with more frequent encounters with primary care providers, and access to facilities performing EGD and with greater complexity of services were more likely to receive repeat EGD, regardless of whether the repeat EGD was appropriate or overuse. Women were slightly more likely to undergo repeat EGD in Probable Overuse situations.

Implications:
Up to 50% of repeat EGDs in VHA may represent overuse. Repeat EGD was associated with more frequent encounters with primary care providers, and access to facilities with greater complexity of services and to facilities performing EGD.

Impacts:
Efforts are needed to better understand the motivations for overuse and barriers to appropriate use, and to promote appropriate use of repeat EGD.