PRESCRIBER PROFILE AND POSTMARKETING SURVEILLANCE

The response rate to requests to general practitioners (GPs) to supply post-marketing data on new drugs has been falling within the prescrip tion-event monitoring (PEM) system organised by the Drug Safety Resear ch Unit in Southampton, UK. To find out why, we looked at the characte ristics of prescribers and the pattern of their prescribing for twenty seven PEM drugs and 543 788 treatments dispensed in England between S eptember, 1984, and June, 1991. 28 402 GPs identified during PEM studi es were divided into six groups according to the largest number of pre scriptions for one or more of the drugs, ranging from group 1 (none of the drugs prescribed) to group 6 (one or more drugs for over 60 patie nts). From group 1 to group 6 the proportion of the GPs who were women decreased from 46% to 9%, and the proportion of overseas-qualified do ctors increased from 13% to 47%. 10% of doctors who had prescribed mos t heavily accounted for 42% of total prescribing. 19 doctors had each prescribed a drug for more than 120 patients during the early post-mar keting period. There was a consistent inverse relation between the num ber of prescriptions and the resonse to requests for post-marketing in formation. The overall response was 53% but the heaviest 10% of prescr ibers returned only 44% and the heaviest 1% returned only 34% of quest ionnaires. No differences in medical need can account for such variati ons in prescribing practice. Heavy prescribing by a minority of doctor s during the period immediately following licensing for marketing may be placing patients at unnecessary risk. These doctors also affect the success of attempts to monitor the safety of new drugs.