Unsuspected patterns of drug utilization revealed by interrogation of a regional general practitioner prescribing database

The classification of drugs according to the specificity of their use has been shown to reveal unsuspected patterns of GP prescribing (McGavock, 1988). The present aim was to define the frequency distribution of rate of issue and cost of prescriptions of (a) drugs used to treat proven pathology, (b) drugs used mainly to relieve symptoms, and (c) drugs like antibiotics, whose use should be specific, but which are often used presumptively. Prescribing data were retrieved from the GP prescription pricing database, specific for individual practices and at the level of the therapeutic group (e.g. hypnotics). A new microcomputer program sorted these data into categories (a), (b) and (c), and produced frequency distributions for all practices in Northern Ireland. The distributions in all three categories were unimodal, positively skewed. A seven–tenfold range in the rate of prescribing was seen in all three categories (cost showed a threefold range), between lowest and highest practices. For each practice, prescribing rate and cost in each category were displayed, with the Northern Ireland average for comparison. This revealed the emphasis placed by each practice on the use of symptomatic treatment and the doctors' tendency to use category (c) drugs. Above all, it revealed the frequency of prescribing of drugs for proven pathology (a) which, prima facie , should be directly related to the diagnoses of diseases such as diabetes, asthma and hypertension. The interpretation of these patterns for practice audit is discussed, together with problems and caveats.