Comparison of Defined versus Recommended versus Prescribed Daily Doses for Measuring Hospital Antibiotic Consumption

Background: Hospital antibiotic use is usually measured by calculating the volume as the number of daily doses defined by the World Health Organization/Anatomical Therapeutic Chemical (WHO/ATC) index (DDDs) divided by a denominator indicating clinical activity, such as the number of occupied bed days. Previous studies have found discrepancies between DDDs, daily doses as recommended in practice guidelines (recommended daily doses, RDDs), and truly prescribed daily doses (PDDs). Very few studies have quantified these discrepancies. Methods: A point prevalence survey was carried out in defined acute care service areas of a large German state university hospital. Of the 941 adult inpatients present on the day of the survey for whom adequate information was available, 267 (28%) had 471 prescriptions for antimicrobial drugs on that specific day. A check for any additional antimicrobial drug prescriptions for these 267 patients during the six days immediately preceding the survey day yielded a total of 2,304 PDDs, of which 1,754 were antibacterial agents. The Latter antibacterial drug PDDs constituted the basis for analysis. The proportion of PDDs different from RDDs and from DDDs was evaluated, and the deviations were calculated. RDD dose definitions corresponded to the local practice guideline recommendations of antibiotic therapy for adult hospitalized patients with normal renal function. Results: Of the 1,754 PDDs, 625 were matching DDD dose definitions (36%), and 1,024 (58%) were matching RDD dose definitions (p < 0.01). Corresponding values for patients with impaired renal function (creatinine clearance < 50 ml/min) were 36% and 42%, respectively. Large DDD discrepancies (< 50% of prescribed doses matching DDD definitions) were noted for betalactams and macrolides, while large RDD discrepancies were observed for aminoglycosides. Compared with PDDs, the use of DDDs as the measurement of hospital antibiotic use overestimated antibiotic use volumes by 32%, while the use of RDD led to a slight underestimation (-9%). Conclusions: The use of DDDs as currently defined by WHO/ATC for measuring hospital antibiotic consumption may be associated with a substantial overestimation of antibiotic use density, while using practice guideline-derived RDDs may yield more valid antibiotic exposure estimates that would be helpful in cross-sectional and longitudinal analyses of antibiotic consumption.