Measures of Adult Antibacterial Drug Use Risk Adjusted by Patient Mix in 74 US Academic Medical Centers: Comparison of Days of Therapy (DOT) to Days of Antibiotic Therapy (DOAT)

Background: SHEA/IDSA guidelines for antimicrobial stewardship programs (ASP) recommend measurement of “risk adjusted” antimicrobial drug use for interhospital comparison (“benchmarking”). The recommended drug use metric is Defined Daily Dose (DDD) intended to estimate the Days of Therapy (DOT). Direct measurement of DOT avoids some limitations of DDD measures. In both methods, each antimicrobial drug is counted separately, whether given in combination or sequentially.  We propose an alternative to DOT, the days of antimicrobial therapy (DOAT), which counts both mono- and combination therapy given on the same day as one DOAT. DOAT may be a useful alternative to DOT for benchmarking since they measure different patterns of use.

Objective: To characterize and compare antibacterial drug use, after risk adjustment by patient mix, by DOT and DOAT in a sample of US academic medical centers (AMCs).

Methods: The data source included administrative claims data for 74 AMCs that participate in the University HealthSystem Consortium (UHC) Clinical Resource Manager (CRM) program. Adult inpatients were grouped into one of 35 clinical service lines (CSLs) based on the Medicare Severity Diagnosis Related Group (MS DRG) to which they were assigned. Antibacterial drug use by CSL in adult patients discharged in CY 2009 was assessed using DOT and DOAT.

Results: Of 1.93 million patients, 1.16 million received at least one dose of an antibacterial drug (mean = 60%; range by CSL = 13% [psychiatry] to 94% [ventilator support]). The mean DOT and DOAT were markedly different when grouped by CSL (Figure). DOAT were always less than DOT, but the ratio of DOT to DOAT depended on the CSL, e.g., patients on ventilator support received a mean of ~40 DOT (interhospital range = 6.6 to 81 days), but the DOAT was half that (mean = 20.3 days, range = 1.0 – 35 days). In contrast, the mean DOT and DOAT for psychiatric patients was short and of similar duration (<1 day).

Conclusions: Antibacterial therapy in adult inpatients can be characterized and risk adjusted by patient mix across multiple hospitals using CSLs. Hospitals with more acutely ill patients contribute to high drug usage because a high proportion receive long courses of combination therapy. However, hospitals with similar DOAT values within a CSL can have very different DOT values because of greater or less use of combination therapy. Both measures may be useful for benchmarking antibacterial drug use.