Drug utilization 90% : using aggregate drug statistics for the quality assessment of prescribing

Abstract: Introduction: There is room for improvement in prescribing. Although a range of indicators have been developed to assess the quality of prescribing, none of them focus on all drugs and there is a gap between what is credible for health care professionals and what can be achieved with the readily available data. It is therefore important to further develop methods for comparisons of drug utilization to provide a better basis for decision-making and better feedback to prescribers. This thesis focuses on the principle of how to present and implement aggregate drug utilization statistics as a basis for improving drug prescribing. Aims: To (1) assess the usefulness of prescribing feedback based on pharmacy dispensing data at the level of individual doctors, primary health care (PHC) centres and clinics; (2) develop a method for the quality assessment of drug prescribing based on routinely available aggregate drug statistics; (3) analyse the strengths and limitations of the proposed method when it is applied in various therapeutic areas and settings; and (4) assess the applicability of the method for international comparisons of drug utilization. Methods: Four studies were observational while three studies described interventions with feedback on prescribing patterns performed in different settings and with varying designs. Data on dispensed drugs were collected from local prescription studies, the Swedish National Prescription Register and databases on dispensed drugs in Funen (Denmark), Bologna (Italy) and Rijeka (Croatia). Drug utilization was expressed in Defined Daily Doses (DDDs), prescription items and costs. The prescribing patterns were analysed using the DU90% (Drug Utilization 90%) method, i.e., focusing on the number of drugs accounting for 90% of the volume and adherence to guidelines within this segment. The feedback provided consisted of DU90% prescribing profiles and corresponding profiles based on costs (Drug Cost 90% DC90%). The prescribers' attitudes were evaluated using questionnaires. Results: The DU90% method was easy and inexpensive and could be adapted to provide comparative data between individual doctors, PHC centres, hospitals and regions.There was a substantial variation in the range of drugs prescribed and adherence to guidelines between different PHC centres and hospital clinics. Feedback with prescribing profiles had a positive effect on the adherence to guidelines issued by a regional Drug and Therapeutics committee. The profiles acted as a catalyst for discussion around deviations from prescribing drugs with better evidence-based documentation or a lower cost, and the prescribers considered them useful in assessing the overall quality of prescribing. The international comparisons revealed important regional differences in prescribing patterns of NSAIDs, irrespective of their relative GI toxicity. This information could be used as a basis for preparing prescribing guidelines. Conclusions: A proper evaluation of the quality of prescribing includes both monitoring the overall costs and volumes of drugs and more sophisticated evaluations using disease- or patientspecific quality indicators. The DU90% method offers an opportunity to add qualitative aspects to routinely available aggregate data, which will increase its credibility and usefulness for prescribers. The method can be considered as a further development of the ATC - DDD methodology and it has proven to be useful both for international comparisons of drug utilization and feedback to doctors on prescribing patterns.