Appropriateness of imaging in lung cancer in a national cohort.

Abstract

7546 Background: Rising healthcare costs prompt emphasis on quality, appropriate care, and cost containment. Imaging is a source of high healthcare expenditures in cancer. We sought to examine variability of adherence to national guidelines for staging, and appropriateness criteria for imaging for patients with locally advanced lung cancer. Methods: Stage IIB, IIIA, or IIIB lung cancer patients were identified from the national VA Central Cancer Registry from 2004-2008 with linkage to VA data and Medicare claims. Imaging was assessed 180 days pre and post diagnosis per National Comprehensive Cancer Network guidelines and American College of Radiology Appropriateness Criteria for Imaging. Multivariate logistic regression with robust variance estimates (adjusting for within-cluster correlation by facility) was used to control for covariates and results reported as adjusted risk differences (95% confidence intervals [CI]). Results: Recommended imaging was performed in 69.5% of patients for brain imaging and 51.1% of patients for positron emission tomography (PET). Overutilization, with combined bone scintigraphy and PET (BS/PET), occurred in 19.7% of patients and did not vary over time (15.8% to 19.5% p=0.168). Facilities affiliated with a medical school had a 14.8 (CI -25.2, -4.4) percentage point lower utilization of PET. Facility volume and tumor board availability were not associated with variability. Appropriateness of imaging varied significantly by region. New England had the highest rates of imaging. Relative to New England, brain imaging was lowest in the Great Basin Region with 26.9 (-40.0, -14.0) percentage point decrease. Both recommended PET use and overutilization of BS/PET was lowest in the Mississippi Region with 25.0 (CI -49.3, -0.6) and 11.7 (CI – 23.3, -0.1) percentage point decrease respectively compared to New England. Conclusions: A significant proportion of patients do not receive recommended imaging and many undergo excessive imaging for lung cancer. Furthermore, there is substantial regional variation in imaging utilization. These observations are hallmarks of poor quality. A disease-based quality improvement plan aimed at modification via policy and reimbursement initiatives may mitigate poor quality care.