BACKGROUND Little is known about the epidemiology of pain in head and neck cancer, the effects of curative treatment on this pain, and the impact that pain experience may have on patients' quality of life (QL). METHODS The prevalence and severity of pain was studied in 93 patients who were first seen with a diagnosis of head and neck cancer, were treated, and remained disease free at two years. QL assessment utilised the life-satisfaction scale and the General Health Questionnaire as specific measures. Pain was assessed by a linear analogue scale anchored by words and numbers. RESULTS Forty-eight percent had head and neck pain when first seen, whereas only 25% and 26% had such pain at 12 and 24 months. Approximately 8% of patients rated the pain as "severe" when first seen, whereas 3% had severe pain at 12 months and 4% at 2 years. The prevalence of shoulder and arm pain increased from 14% at diagnosis to 37% at a year and 26% at 24 months, but the percentage of patients with severe pain at any stage postoperatively was only 5% and 2%, respectively. Any pain (pain in either in the head and neck or shoulder and arm or both) at 2 years was strongly predicted by earlier posttreatment pain (at 3 months or at 12 months.) Shoulder and arm pain at 2 years was strongly correlated with surgical treatment of the neck, although no difference in pain experience was noted between those who had radical neck dissections and those who had more conservative procedures. There was no correlation between radiotherapy to the neck and subsequent shoulder and arm pain. Pain had an adverse effect on the general well-being and psychological distress of head and neck cancer patients who were free of disease. CONCLUSIONS Pain is common among those presenting with curable head and neck cancer. Pain can be reduced by curative treatment but neck dissection may cause increased shoulder and arm pain. Ongoing pain is predictable and impacts adversely on patients QL.