The purpose of this study was to evaluate the effect of a community-oriented intervention in one part of the Free Town of Bremen, northern Germany (population 541,000) on the prescription prevalence of World Health Organization (WHO) class III opioids for cancer patients in their final year of life. A community-oriented, multimodal intervention included information, teaching, and training modules tailored to physicians, pharmacists, nursing staff, and patients and their relatives, and the public. Prescription prevalences were calculated for the intervention region (Bremen-Nord) and a control region (Bremen-Mitte) before and after the intervention. Specifically, a population-based, controlled, quantitative assessment of opioid prescriptions for patients with cancer during their final year of life was undertaken for two time periods, prior to 1992-1993 and after 1995-1996, respectively. Prescription ascertainment was based on duplicates kept in the pharmacies. Patients comprised two anonymized complete 4-month samples who died in 1993 and 1996, respectively, with cancer as the primary or a contributory cause of death on their death certificates. A total of 1282 prescriptions were abstracted from duplicates in 109 of 119 pharmacies in Bremen-Mitte and all 31 pharmacies in Bremen-Nord (overall pharmacy participation proportion 93%) and individually matched to 856 patients with cancer in their final year of life. In 1993, 16.3% of all terminal cancer patients in Bremen-Mitte and 19.1% in Bremen-Nord had received at least one prescription for a WHO class III opioid. Corresponding numbers after the intervention were 20% and 21%, respectively. The total amount of class III opioids, however, increased 20% in Bremen-Mitte and 210% in Bremen-Nord after the intervention. In 1996, the spectrum of prescribed opioids had changed markedly toward the WHO recommendations. The proportion of prescribing physicians remained constant. These data suggest that a community-oriented intervention in one part of Bremen had a limited impact on cancer pain therapy on the population level. A measurable change of prescription practice seemed to be restricted to the minority of physicians, who had prior experience with prescribing WHO class III opioids.