Poor correlation with Jarman index may be due to changes in deprivation since 1981

To validate the index the authors correlate it with the level of prescribing of certain drugs, hormone replacement therapy, and tamoxifen, which they regard as indicators of the social class composition of the practice population.This procedure involves unsupported assumptions, particularly in relation to hormone replacement therapy.An underlying problem with the authors' use of PACT data is the implicit assumption that prescribing is a standardised procedure in which general practitioners diagnose and treat disease systematically.The literature does not support this assumption; indeed, it shows great variation among doctors, much of which can be explained only by personal idiosyncrasy.2It seems dangerous to base an objective measure of deprivation on such subjective behaviour.Although the correlation between exempted precriptions and deprivation at health authority level is high, the measure proposed by Lloyd and colleagues is intended for practice based research rather than epidemiological research on larger populations.For example, the authors suggest that it would be useful in researching topics such as prescribing.Given that PACT data are Britain's main source of prescribing data and also provide the data for the authors' deprivation index, however, this suggestion seems tautologous.Further efforts are needed to develop a reliable and valid measure of deprivation at the practice level.