Measurement units for antibiotic consumption in outpatients.

Abstract

Sir, Recently, Coenen et al. reported that number of packages per 1000 inhabitants per day (PID) is a more appropriate measure than number of defined daily doses (DDDs) per 1000 inhabitants per day (DID), when assessing outpatient antibiotic use over time and to measure the effect of interventions in countries dispensing ‘complete packages’. This is true if a significant change in DDDs per package happens over time. In Slovenia, since 2000 to the present day, we express antibiotic consumption in outpatients both in measurement units and in the number of prescriptions per 1000 inhabitants per year. During the period 2000–12, we reduced the use of antibiotics for systemic use by 28.5% expressed in DID (from 19.94 to 14.26), by 27.4% expressed in number of prescriptions per 1000 inhabitants per year (from 724.5 to 526) and by 32% expressed in PID (from 2.88 to 1.96). On average, in 2000 1.46 and in 2012 1.36 packages per prescription were prescribed and on average one package had in 2000 6.9 DDDs and in 2012 7.3 DDDs. These small changes can explain different percentages in reductions. Amoxicillin and co-amoxiclav represented in 2000 38.9% and in 2012 52.5% of total use. The contents of a package of amoxicillin did not change over time and had between 2.0 and 20 DDDs per package. The mean contents of a prescribed package had 6.8 DDDs in 2000 and 8.4 DDDs in 2012. For co-amoxiclav there has been a change in the amount of DDDs per package, with the registration of a new compound (amoxicillin/clavulanic acid 16 : 1), which increased the highest amount of DDDs per package from 12.25 to 28 DDDs, but the consumption of the latter was very low (only 95 packages). The mean contents of a package of co-amoxiclav increased in the investigated period from 7.2 to 8.3 DDDs. In contrast, in Belgium, increased number of DDDs per package for amoxicillin and co-amoxiclav changed significantly. DDDs per package increased by 50% for amoxicillin and by 70% for co-amoxiclav and these two antibiotics represent 54% of outpatient antibiotic consumption expressed in DDDs. The increased number of DDDs per package explains why Belgium has not decreased the consumption expressed in DID while the number of PID decreased. When the number of DDDs per package does not change over time or if the consumption of packages with increased amount of DDDs per package is low, we can express national data in DID. A good correlation with DID is also found with the number of prescriptions per 1000 inhabitants per year. Why do we measure antibiotic consumption? Reasons for national antimicrobial consumption surveillance are to identify trends in prescribing that require further investigation, measure the effect of stewardship activities including cost savings, provide feedback to prescribers, to study correlation between antibiotic use and resistance and regional, national and international benchmarking. For international benchmarking and to study correlation between antibiotic use and resistance, it is advisable to use DID. – 5 To identify trends in national prescribing and measure the effect of national stewardship activities, all three measurement units should be used especially if substantial changes of packages have happened over time. In 2011, we found almost the same consumption in outpatients in Slovenia and Sweden expressed in DID (14.4 versus 14.3), but large (42% and 29%) differences when we expressed the consumption in PID (2.02 versus 1.18) or number of prescriptions per 1000 inhabitants per year (540.5 versus 385.3). It appears that fewer, but larger, packages were prescribed in Sweden and higher doses per treatment contribute to the low resistance rates registered in Sweden; however, their guidelines do not support this explanation. It seems that for international benchmarking of antibiotic use in outpatients, DID is a better measurement unit than PID. If substantial changes in the amount of DDDs per package happen over time, then additional measurement units should be used, such as PID and number of prescriptions per 1000 inhabitants per year, to identify trends in national prescribing.