OBJECTIVES: To compare the clinical effectiveness of an intensive three-month training programme with a less intensive programme on females suffering from chronic neck/shoulder pain. STUDY DESIGN: A prospective observer-blinded clinical trial including 12-month pretreatment follow-up. SETTING: Patients were referred to the Departments of Rheumatology and Physical Medicine at Hvidovre Hospital by their general practitioners. Training was undertaken at a satellite clinic for physiotherapy of Hvidovre Hospital. SUBJECTS: Female patients aged 18-65 years suffering from chronic neck/shoulder pain for a minimum of six months. INTERVENTION: Patients were examined by a physician in order to exclude serious diseases. They were then randomized to either an intensive neck/shoulder training programme or a programme of lesser intensity but of similar duration. MAIN OUTCOME MEASURES: Scales measuring pain and activities of daily living (ADL) were used, and strength and endurance measurements of the cervical and shoulder muscles were carried out at baseline and completion of the study. Follow-up measurements were carried out by postal questionnaire at 6 and 12 months after inclusion, and included pain, ADL and treatment satisfaction measurements. RESULTS: Seventy-seven patients were included in the trial, of whom 27 (69%) completed the intensive programme and 25 (61%) the lighter programme. Forty-one (>80%) completed the follow-up questionnaires. The patients in the two groups did not differ with regard to age, pain, ADL scores and physical measurements prior to training. Patients in both groups that completed the trial demonstrated statistically significant improvements in nearly all of the outcome measurements at completion. ADL scores maintained statistical significance at 12 months in both groups, but pain scores were only significantly improved in the intensive group at 12 months follow-up. There was no statistically significant difference between groups regarding pain or ADL, but overall 50% of all patients showed improvement. CONCLUSIONS: The type of low-tech dynamic training used in either of our two programmes resulted in both subjective and objective improvements in patients suffering from chronic neck/shoulder pain, but there were no statistically significant differences in outcome between the two approaches. The subjective improvements were maintained throughout the follow-up period.