Predictors of Rates of Resistant Gram-Negative Bacteria

Antibacterial drug use

Systemic antibacterial drug use in adult inpatients discharged between January 1, 2006 and December 31, 2010 was obtained from patient-level billing records. These data were aggregated and reported for each hospital as days of therapy per 1000 patient days (DOT/1000PD) as previously described [10]. Any dose of an antibiotic received by a patient during a 24 hour period is counted as one DOT. For example, administration of imipenem/cilastatin 1000 mg every 8 hours, or administration of 500mg every 6 hours, is counted as one DOT. We have recently reported advantages of measuring antibiotic use by DOTs versus the metric usually recommended, the Defined Daily Dose (DDD) [12].

Antimicrobial susceptibility

In year 2009 we requested the annual cumulative hospital antibiograms from the 50 hospitals for the years in which we had carbapenem drug use. The contact at each hospital was usually the antimicrobial stewardship pharmacist, but may also have included the infectious diseases physician(s), clinical microbiologist or infection control practitioner. We utilized antibiograms with a full calendar year of susceptibility reported for all clinical isolates (at least 30 isolates), the total number of isolates and the proportion of resistant isolates.

All hospitals received an on-line survey requesting additional information regarding susceptibility testing methods and antibiogram construction. We used the secure survey instruments provided by REDCap (Research electronic data capture; www.project-redcap.com) to send and compile survey responses. Specifically we inquired about the inclusion/exclusion of duplicate clinical isolates in the antibiogram, method(s) of susceptibility testing, policy regarding surveillance cultures and we attempted to verify that Clinical Laboratory Standards Institute (CLSI) interpretative breakpoints were used for all years. As recommended by Schwaber we recorded both proportions and rates of carbapenem resistant P. aeruginosa [13]. The resistant proportion was the number of resistant isolates divided by total number of isolates, and the resistant incidence rate was the number of resistant isolates per 1000 adult patient days (PD).