• prescribing costs
• general practice

Rising prescribing costs are a major concern in modern healthcare systems worldwide. Wherever costs are analysed and combined with any kind of quality measures, large variations between practices can be seen which cannot be explained fully by the underlying morbidity of the patients treated in these practices. For this reason, the knowledge, attitudes, and beliefs of doctors themselves has become the focus of research.

Since the early work of Parish¹ we have known that prescribing has an irrational as well as a rational basis. Balint et al² showed that writing a prescription is an easy way of cutting the consultation time for overworked doctors. Howie³ in his study of clinical judgement and antibiotic use in general practice demonstrated the influence of situational factors on prescribing decisions. The personal views of the doctor can also play a similar role⁴; often GPs think that patients expect to get a prescription at the end of the consultation. In European countries such as Germany, Belgium, and Switzerland which have no fixed list system, this may lead to a fear of losing patients, although large patient surveys in these countries have shown that more patients change their GP because of overprescribing than because they think their doctor tries to keep down prescribing costs.⁵

The paper by Watkins et al⁶ in this issue of QSHC identifies more factors and summarises some well known findings on the basis of a cross sectional study. The influence of the pharmaceutical industry again becomes apparent. We know that it is pervasive but almost invisible to individuals—but what else could we expect? Why should an industry invest billions of dollars per year in changing behaviour their way if they did not see this is as an effective intervention?⁷

A question which often arises in this context is whether we need “sticks or carrots” to change prescribing patterns. The answer is unclear. What we now know from research on implementing change is that a multimodal approach, tailored to the individual needs of the doctor, is likely to be more effective than a single “one shot” intervention.⁸ We have to remember that not all doctors are the same, and that they are part of a complex health system, on the one hand, and part of a complex interaction with the patient on the other. Prescribing decisions, like all medical decisions, are the result of an interaction between the GP and patient, so their “rational” basis will always be affected by a combination of attitudes, beliefs, and knowledge of the two parties involved. This is why we have to understand the personal needs, beliefs, and attitudes of GPs and patients in order to give them the kind of support which might be most effective for them. In future this could mean that interventions on the doctor’s side will have to be tailored, not only on a regional basis, but also to the individual practice and the individual doctor. An intervention would mean a flexible framework which gives information and feedback in an environment with trust and support combined with financial and other incentives. “Sticks” will then be necessary only for those few (if any) who are unwilling to do anything. As Bradley⁹ has recently pointed out, we need greater insight from qualitative research and must adopt the findings such as those of Watkins et al to improve our interventions. I am sure we can.