Statistics on local drug sales: A tool to identify problem areas and to follow effects of education on drug use

Sammanfattning: Aims: 1) To assess if drug sales data indicate differences in morbidity, mortality, socio-economic conditions and deviating prescribing habits among physicians; 2) to identify areas for educational interventions and to analyse changes in prescribing after educational activities. In addition, an effort was made to estimate differences between pharmacy sales and purchases by the population, and the amount and value of unused drugs returned to pharmacies. Method: Information on drug sales was gathered from three Swedish and one Nordic drug registers, namely, Sales Statistics on Drugs, Nordic Statistics on Medicines, The National Prescription Survey and The Diagnosis-Prescription Survey, as well as from local prescription studies, copies of prescriptions processed at pharmacies, copies of issued prescriptions and computerised patient records at a health care centre. Results: There was a significant correlation between sales of tranquillizers and hypnotics/sedatives, on the one hand, and mortality, suicides and UnderPrivileged Area score, on the other, in the 33 municipalities of Skåne County. Benzodiazepine sales in the city of Helsingborg were the highest in the country and higher than the national average to all age groups and both genders. A minority of the physicians, <5%, had issued about 10 times as many benzodiazepine prescriptions as the average physician in Helsingborg. High compliance rates among district physicians after information by pharmacists, to brands recommended by the Drugs and Therapeutics Committee, attributed to cost savings several times the costs of the information campaign. Repeated verbal producer-independent education produced significant changes in attitudes towards drugs, drug treatment and drug information. Both this programme and a local educational programme at a health care centre resulted in an overall and sustained reduction in antibiotic prescribing. There was a substantial difference between sales at the local pharmacy and the purchases on prescription by the local population. Less than 4% of the purchased drugs were returned unused to pharmacies for destruction. Conclusion: There are large differences in sales of drugs between municipalities within the same county. The results show that local drug sales covariates with, and hence may indicate, socio-economic conditions, morbidity and mortality. This information may be used to identify areas where there is a need of information and education activities to promote rational prescribing. Local sales data are reasonably well suited to follow effects of such information and educational activities. However, local sales of drugs do not fully correspond to purchases and drug use by the local population, neither to prescribing by the local physicians. In order to identify target groups for information and education, there is a need for additional information, as deviant prescribing habits among a minority of doctors may be influential. Nevertheless, successively improved quality and accessibility of local sales data may provide a useful and cost-effective means to analyse and improve prescribing and use of drugs.