We used multivariate analyses to mathematically simplify a set of 10 factors to two predictors of shoulder pain. The multivariate model had a good level of accuracy, and explained 63% of the variance in the dataset. Additional factors, such as age and altered tone, did not enhance the model, which suggests that the fit of the model was good. Nevertheless, given that any model is highly dependent upon its derived dataset (Tabachnick and Fiddell 2001), the findings should be replicated in other samples before being recommended AZD8055 mouse for wider use. Our findings support that shoulder pain post-stroke is heterogeneous in nature (Price 2002). Level of risk and underlying mechanisms

are likely to vary according to the type and severity of impairments, and personal (eg, age and premorbid shoulder problems) and environmental factors (eg,

trauma) (Ratnasabapathy et al 2003). It therefore seems important to develop clearer diagnostic classifications in order to direct clinical management. Our findings indicate that the Motor Assessment Scale Upper Arm item selleck inhibitor score may be helpful for this issue. For instance, a score of < 4 indicates a high risk of developing shoulder pain, as proposed in the Management Tool for Acute Hemiplegic Shoulder (Nicks et al 2007). For this group of patients, who are also more likely to have shoulder subluxation, clinical management including use of arm support, electrical stimulation, education, and active motor training to promote shoulder girdle control, as outlined by Nicks and colleagues, seems highly appropriate. However, despite the lower odds, patients admitted with a score of 4 or 5 in our study also had shoulder pain. Physiotherapists would need to employ other approaches to manage these people as different mechanisms for pain, such as shoulder

impingement, are likely (Bender and McKenna 2001, Blennerhassett et al 2009). Despite the observed association with pain, reduced passive range and motor control at the shoulder cannot be considered the cause of post-stroke 3-mercaptopyruvate sulfurtransferase shoulder pain. Nevertheless, the findings suggest that clinical attention could be directed to improving pain free shoulder joint range, or promoting active shoulder girdle control to align the glenohumeral joint and enable arm elevation. Training should be carefully structured and monitored, given the importance of highly co-ordinated muscular control within the shoulder girdle (Dontalelli 2004), and the potential for impingement, wear and tear, inflammation, and subsequent pain at the shoulder – particularly when the muscles are weak or fatigued, or while performing overhead activities (Ludewig and Reynolds 2009). Education and training of staff, carers, and patients in how to care for the arm are also warranted (Nicks et al 2007, Turner-Stokes and Jackson 2002), given the vulnerability of a weak shoulder and the events described that may have contributed to the development of shoulder pain.