Claims-based surveillance of surgical site infections following vascular surgery.

Brigham And Women'S Hospital, Boston MA

Investigators

Linked publications & trials

Abstract

DESCRIPTION (provided by applicant): This project will develop and validate a standardized, claims-based surveillance method to rank hospitals nationally based on diagnosis and procedure codes suggestive of surgical site infection (SSI) following vascular surgery. This work will be part of an ongoing collaboration between the Harvard-based CDC Prevention Epicenter, the CDC's Division of Healthcare Quality Promotion, and the Centers for Medicare and Medicaid Services (CMS). The first aim is to develop a set of diagnosis and procedure codes that suggest SSI following vascular surgery. We will perform a retrospective cohort study of Medicare patients who underwent one of the vascular procedures targeted by the national Surgical Care Improvement Project (SCIP) in 2005. Approximately 100 codes that have been pre-selected as SSI indicators will be pilot tested in five CDC Prevention Epicenter hospitals using full medical record review. We will use the pilot test to refine our code list to ensure a sensitivity of at least 90% for all confirmed SSIs while attempting to maximize specificity. The second aim is to apply our refined claims-based surveillance method to the 2005-2008 Medicare national claims database. We will rank all hospitals that performed at least 200 targeted vascular procedures in Medicare patients between 2005 and 2008 (average of 50 procedures per year). Ranking will be based on the fraction of patients at each individual hospital with>1 SSI indicator according to the claims-based surveillance method. Rankings will adjust for age, gender, and Romano comorbidity score via a generalized linear mixed model that controls for clustering within institutions. Hospitals will be grouped into deciles based on their adjusted SSI rates. Finally, the third aim is to validate the ranking of hospitals in the best and worst deciles based upon nationwide review of up to 1,500 full-text records. We will perform logistic regression to assess the relationship between having a vascular surgery performed in a best versus worst deciles hospital while controlling for case mix.