Objective: To determine the natural history of shoulder pain in the population, and predictors of outcome on the basis of clinical and individual factors. In addition, to determine whether outcome is influenced by the definition of shoulder pain used.
Methods: A prospective cohort study, over a 3 year period, of subjects recruited from a cross sectional population screening study of shoulder pain, conducted in the Greater Manchester area of the UK.
Results: Of 92 subjects classified as having shoulder pain in the cross sectional study, 50 (54%) reported shoulder pain at followup about 3 years later. In 90% of cases this was accompanied by some disability specifically relating to the symptoms. Baseline factors that predicted symptoms at followup were: pain (indicated on a manikin) within a more narrowly defined shoulder region, shoulder pain related disability, pain on the day of examination, symptoms lasting more than one year, and a high score on the General Health Questionnaire, a measure of psychological distress.
Conclusion: Shoulder pain in the population is a longterm disabling symptom, although many subjects do not seek early medical consultation. Disability (independent of whether there was restriction of movement on examination) is a strong predictor of continuing symptoms. The outcome observed in epidemiological studies of shoulder pain will be influenced by the initial definition of symptoms.