This study assesses the treatment attitudes and practices of general practitioners in the management of pain caused by cancer. The study consisted of two parts: (1) a questionnaire survey of 328 GPs (including trainees) in Tayside, Scotland and (2) an analysis using duplicate prescriptions of all opioid analgesic prescribing over a five-year period among the doctors in a primary health care centre (population 11500). In the questionnaire survey, 63% (206) of the GPs replied. Of these respondents, 80% were prepared to manage malignant pain on their own and 17% would refer 'frequently' to the local hospice. Only 19% of respondents work in collaboration with the pain clinic and, if so, the pain clinic undertakes most of the decisions made on pain management. Co-proxamol (paracetamol plus dextropropoxyphene hydrochloride preparation) was the preferred analgesic for mild to moderate pain due to cancer and controlled-release morphine sulphate was preferred for severe pain. Duplicate prescription monitoring showed that 24% (2722) of patients received one or more prescriptions for opioid analgesics over the study period. Forty-five per cent of the patients were prescribed co-proxamol. When morphine sulphate or diamorphine was prescribed (3% and 1 % of patients respectively), laxatives were frequently coprescribed (64% and 58% respectively), as were anti-emetics (57% and 32% respectively). This study therefore shows that most GPs manage pain caused by cancer without help from hospital specialists. This study also shows a degree of awareness concerning good prescribing practice in cancer pain. Cette étude évalue les conceptions et pratiques des médecins généralistes en ce qui concerne le traitement dans la gestion de la douleur d'origine cancereuse. L'étude se divise en deux parties: (1) un questionnaire envoyé & a g r a v e ; 328 généralistes (y compris stagiaires) dans le Tayside en Ecosse et (2) une analyse & a g r a v e ; I'aide des duplicata des ordonnances de toutes les prescriptions d'analgésiques opioïdes, sur une période de cinq ans, des médecins appartenant à un centre de soins généraux (population 11 500). 63% (soit 206) des généralistes ont répondu. Parmi ceux-ci, 80% étaient disposés & a g r a v e ; gérer la douleur d'origine cancereuse par leurs propres moyens et 17% préféreraient se référer "fréquemment" à I'hospice local. Seuls 19% travaillent en collaboration avec le service clinique de la douleur et, dans ce cas, la clinique se charge de la plupart des décisions en ce qui concerne la gestion de la douleur. Le Co-proxamol (une préparation de paracétamol additionné de chlorhydrate de dextropropoxyphène) était I'analgésique p r & e a c u t e ; f & e a c u t e ; r & e a c u t e ; pour le traitement d'une douleur d'origine cancéreuse faible & a g r a v e ; moyenn...