Implications of a defined daily dose fixed database for drug utilization research studies: The case of statins in Portugal

Aims Given the discrepancies between PDDs (prescribed daily doses) and DDDs (defined daily doses), we aimed to assess the extent of error in the results of an 18-year population-level study on statin utilization in Portugal. Methods The Portuguese regulatory agency provided data for the period 2000-2018 on statin dispensing (C10AA). The DDDs were gathered from the ATC/DDD database. DDDs were calculated by the DDD year-by-year approach (DDDYEAR) and by the DDD last-year approach (DDDLAST). PDDs were calculated according to the year-by-year approach (PDDYEAR). Statin annual utilization rates per 1000 inhabitants per day were also calculated. Percent errors were calculated for PDDYEAR and DDDYEAR units. Results The DDDYEAR approach revealed decreases in the consumption of atorvastatin, fluvastatin, lovastatin, pravastatin and simvastatin in 2009, when their DDD was modified. Conversely, the results from both DDDLAST and PDDYEAR approaches indicated gradual changes in the actual consumption of all statins in Portugal. Before 2009, atorvastatin, pravastatin and simvastatin utilization was greatly overestimated by DDDYEAR/1000 inhabitants/day. The average dose of lovastatin prescribed in the past 18 years (20 mg) was below the assigned DDDs during the study period, varying from 30 mg to 45 mg. Conversely, the PDD for fluvastatin was above the DDD values (ranging from 40 mg in 2000 to 70 mg in 2016). For atorvastatin, pravastatin and simvastatin, national PDDs were above the assigned DDD until the DDD modification in 2009. Conclusions A more dynamic system, based on national and annually updated DDDs, should be able to reduce discrepancies between DDDs and PDDs and the bias in utilization studies.