Randomised controlled trials (RCTs) of manual physical therapy (from a physical therapist) given only to the glenohumeral joint for shoulder pain in adults (18 to 80 years) were eligible. Studies had to report at least one of the outcomes active or passive range of motion, a shoulder-specific functional outcome, quality of life and pain. Participants with cervical or thoracic symptoms, arm pain other than the shoulder, evidence of joint instability, a history of traumatic dislocation or radicular symptoms were excluded. People who had undergone surgery were excluded. Studies that incorporated general anaesthesia were excluded.
The conditions of included participants varied, but were commonly either idiopathic adhesive capsulitis or shoulder impingement syndrome. Many studies compared different types of manual therapy and four types were used: mobilisation with movement, a Cyriax approach and static mobilisation at end- or mid-ranges of motion. Some studies had sham mobilisation or heat treatment as comparators. Between three and 24 treatment sessions were administered per week for between two and 12 weeks. All studies assessed range of motion. Most studies assessed pain (using a visual analogue scale).
One reviewer selected abstracts and four reviewers independently screened full papers and compared their results to decide on exclusions.