Shoulder complaints are common and have an unfavourable outcome in many patients. Only 50% of all new episodes of shoulder disorders end in complete recovery within 6 months. There is no consensus about prognostic indicators that can identify patients at high and low risk of chronicity. By a systematic search of the literature we identified 16 studies focusing on the prognosis of shoulder disorders. The methodological quality of these 16 studies was assessed. Six of these were considered to be of relatively 'high quality'. There was a wide variety among the studies in length of follow-up, study population, evaluated prognostic factors, type of outcome measure and method of analysis. Due to this large heterogeneity, we refrained from statistical pooling. Instead, we used a best-evidence synthesis. There is strong evidence that high pain intensity predicts a poorer outcome in primary care populations and that middle age (45-54) is associated with poor outcome in occupational populations. There is moderate evidence that a long duration of complaints, and high disability score at baseline predict a poorer outcome in primary care. These results need to be interpreted with caution because of the small number of studies on which these conclusions are based, and the large heterogeneity among studies regarding follow-up, outcome measures, and analysis.