Antimicrobial Expenditures and Usage at Four University Hospitals

Background: The increasing use of antimicrobial drugs is resulting in enormous hospital expenditures. Careful assessment of inappropriate prescribing and a search for more cost-effective treatment strategies are urgently required. Comparisons between hospitals should help identify areas of inappropriate prescribing as well as effective drug use programs, but such analyses may be severely biased if the impact of different case-mixes is not recognized. Patients and Methods: We studied antimicrobial usage and expenditures at four state university hospitals in southwestern Germany and assessed the significance of differences between hospital services after adjustment for patient variables. A prevalence survey was done with review of 2,254 charts of patients admitted to the surgical, medical and pediatric services to obtain information on antimicrobial drug prescription and expenditures in the week preceding the survey. Results: According to pharmacy data for the year 1994, maximal diferences between these hospitals in the antibiotic costs per patient-day were 1.9-fold (surgical services), 1.5-fold (medical services), and 1.6-fold (pediatric services). In a multivariate analysis, adjusted antibiotic prescription prevalence rates did not differ for medical and pediatric service patients, but did differ for surgical service patients (p = 0.03). Similarly, adjusted expenditures per patient-week differed significantly between hospitals for surgical service patients (p = 0.001), but only marginally for medical (p = 0.14) and pediatric (p = 0.05) service patients. The adjusted difference in expenditures between surgical departments was as large as 2.8-fold (95% CI 1.8 to 4.3) and was primarily related to preferential use of expensive iv antimicrobial drugs. Conclusion: In two hospitals, lowest expenditures in either surgery or medicine were assiciated with active antimicrobial drug use programs suggesting an impact of these programs on drug use and expenditures limited to these services. The identification of such large patient-mix unrelated differences in antimicrobial usage and expenditures offers opportunities for quality improvements and cost reduction.