Predictors of Hospital-Acquired Urinary Tract-Related Bloodstream Infection in VA

Veterans Health Administration, Decatur PA

Investigators

Linked publications & trials

Abstract

DESCRIPTION (provided by applicant): Preventive efforts to reduce urinary tract-related bloodstream infection in veterans are hampered by the scarcity of analytic studies that identify etiologic factors, particularly those specific to the time-dependent elements that may impact prognosis during hospitalization. Therefore, the ability of clinicians to intercede in a meaningful way has been restricted. The objective of this VA-based clinical study is to delineate both host- and time- dependent factors that contribute to urinary tract-related bloodstream infection in hospitalized veterans. This study is unique in that it will ascertain potential causal mechanisms of bloodstream infection taking into account the time course of a veteran's hospital stay and the multiple procedures and treatments that occur during this stay. Specific Aim #1 addresses the relation between obesity, diabetes, and bloodstream infection. Specific Aim #2 is directed to time-related treatment factors that impact the incidence of infection. Of particular interest are the use of antimicrobials, immunosuppressant therapies and statin use. Specific Aim #3 assesses the relationship between transfusion-related risk factors and bloodstream infection. A case-control study is planned that will include adult patients from three VA medical centers in Ann Arbor, Detroit, and Indianapolis. Incidence density sampling will be utilized for random selection of controls. Information from cases and controls will be extracted from electronic databases (electronic medical, pharmacy, and laboratory records) over an eleven-year period, 2000 to 2010. Analyses will identify factors related to the development of urinary tract-related bloodstream infection in hospitalized veterans. Conditional logistic regression with clustering by hospital location will be utilized for the analyses, as well as sliding time windows for log-linear models and cubic B-splines with time-specific weight functions. We anticipate that the information obtained from the proposed study - which evaluates both novel risk factors and is relevant to the clinical care and safety of veterans - will enhance tools for clinical decision-making and provide a framework for preventive trials in VA patients.