PREDICTING PRESCRIBING COSTS IN GENERAL-PRACTICE USING PRACTICE DEMOGRAPHY

Prescription and dispensing costs form a large part (c. 56%) of primar y care expenditure in the NHS and concern has been expressed at its ev er increasing total. Previous predictive models have either failed to account for a high proportion of costs or else have not been able to e xplain adequately the role practice list demography plays upon costs. Using prescription data and the practice demography, our model account s for 91.4% of the variation in primary health care prescribing costs in Northern Ireland thus explaining them to a much greater extent than previous models and, in addition, explains a large part of the variat ion in total monthly consultations and numbers of prescriptions. In ad dition to comprehensiveness it has a high degree of parsimony, needing only three independent variables for each practice, namely, the numbe r of children aged 0-4 years, the number of persons aged 60 + years an d the number of partners in the practice, all of which are immediately comprehensible by GPs and their negotiators. Thus, it could form a va luable addition to the 'evaluation kit' of prescribing advisers and ot hers concerned with auditing and containing costs. Previous studies ha ve shown the importance of the age-sex structure of practice lists in relation to prescribing costs but none has been able to develop such a powerful, simple and comprehensible predictive model. (C) 1997 by Joh n Wiley & Sons, Ltd.