PURPOSE: The primary aim of this study was to assess the effectiveness of possible interventions for hemiplegic shoulder pain. The secondary aim was to investigate whether reduction of subluxation or spasticity can decrease shoulder pain and whether a change in shoulder pain is related to change in passive shoulder external rotation. METHOD: MEDLINE, EMBASE, CINAHL and the Cochrane Central Register of Controlled Trials databases were searched to obtain the randomised, controlled trials. Two authors independently extracted data and assessed the methodological quality of studies. RESULTS: Eight randomised trials were found in electronic databases. Aromatherapy plus acupressure, slow-stroke back massage and intramuscular neuromuscular electric stimulation were more effective than the controls at the end of treatment sessions. Intramuscular botulinum neurotoxin A injection and intraarticular triamcinolone acetonide injection were not helpful at one or three months after the end of treatment. Only intramuscular electric stimulation was effective at three months. These analyses found that shoulder pain improved independently of spasticity and subluxation. It was confirmed that the change in shoulder pain was associated with change in passive shoulder external rotation. CONCLUSIONS: Although five interventions were used for managing hemiplegic shoulder pain, their effects were limited in the context of trials.