Background: The most effective treatment for musculoskeletal shoulder pain is unknown. Physiotherapy is often the first point of referral. However, there is uncertainty as to which patients will benefit. Purpose: To identify which patient and clinical characteristics, commonly assessed at the first physiotherapy appointment, are associated with better or worse patient rated shoulder pain & function six weeks and six months later. Methods: This prospective multicentre cohort study recruited patients referred to physiotherapy for the management of musculoskeletal shoulder pain. It took place within 11 NHS trusts and social enterprises in the East of England, including primary and secondary care, between November 2011 and October 2013. Seventy one potential prognostic factors were collected prior to and during the patient's first physiotherapy appointment and included individual and lifestyle characteristics, psychosocial factors, shoulder symptoms, general health, clinical examination findings, activity limitations and participation restrictions. Physiotherapy treatment was unaffected. Outcome measures included two self-report postal questionnaires; the Shoulder Pain and Disability Index (SPADI) (MacDermid et al, 2006; Roach et al, 1991) and the Quick Disability of the Arm, Shoulder and Hand (QuickDASH) (Beaton 2005). Multivariable linear regression was used to analyse prognostic factors associated with outcome. The protocol has been published previously (Chester et al, 2013). Results: 1030 participants were recruited and provided baseline data, 82% (n = 840) provided outcome data at 6 weeks, 79% (n = 811) at 6 months. Ten prognostic factors were consistently associated with the SPADI and QuickDASH at one or both time points. Five factors were associated with better outcomes at both time points: lower baseline disability, patient expectation of “complete recovery” compared to “slight recovery” as “a result of physiotherapy treatment,” higher pain self-efficacy, lower pain severity at rest, and for patients who were not retired, being in employment or education. Only three clinical examination findings were associated with outcome and each at one time point only. For the SPADI in particular, a greater range of shoulder abduction was associated with a better outcome at six weeks follow up, and a smaller difference between active and passive abduction was associated with a better outcome at six months follow up. For both the SPADI and QuickDASH, at six months follow up only, a reduction in pain or increase in range of shoulder elevation with manual facilitation of the scapula during elevation of the arm, was associated with a better outcome. Conclusion: A wide range of biopsychosocial factors were associated with patient rated outcome. Psychological factors were consistently associated with outcome at both time points. Clinical examination findings associated with a specific structural diagnosis were not. Clinical examination findings associated with symptom modification during manual facilitation of the scapula during elevation of the arm was consistently associated with both outcomes at six months. Implications: When assessing people with musculoskeletal shoulder pain psychological in addition to medical information should be considered.