2027. Does VA Colorectal Cancer Screening Performance Measure Reflect Quality?
Louise C Walter, MD, San Francisco VAMC, NP Davidowitz, San Francisco VAMC, PA Heineken, San Francisco VAMC, KE Covinsky, San Francisco VAMC

Objectives: To determine whether compliance with the national VA colorectal cancer screening (CRC) performance measure reflects quality medical care.

Methods: A study of the 229 veterans whose medical records were reviewed by external auditors contracted by the national VA Office of Quality and Performance to compute the FY 2002 CRC performance measure for the San Francisco VAMC. We abstracted data about compliance with the CRC performance measure (defined by FY 2002 VHA Technical Manual) as well as comorbid conditions, patient preferences and adherence, and screening recommendations.

Results: The mean age of audited patients was 69 years (31% > 75 years). 69% had a Charlson comorbidity-age combined risk score >=4 (predictive of > 50% 10-year mortality). Compliance with the CRC performance measure was 65%. However, of the veterans credited with compliance, 39% had tests performed for diagnostic rather than screening purposes. Of the 81 veterans not counted as compliant, 43% had declined screening and 12% failed to keep appointments for scheduled tests.

Conclusions: The VA CRC performance measure over-samples patients with poor prognoses. As a result, in many veterans selected for audit, CRC screening is not indicated and in some cases would represent poor quality care. Patient preference was a major reason for not screening and should factor into measures of quality. Furthermore, the measured “screening” rate is heavily determined by needs for diagnostic tests rather than screening recommendations.

Impact: The VA CRC performance measure should be changed since it is a poor measure of the quality of medical care.