OBJECTIVE: To investigate the effectiveness of a contracture preventive positioning procedure for the hemiplegic arm in subacute stroke patients in addition to conventional physio- and occupational therapy. DESIGN: A single-blind pilot randomized controlled trial. SETTING: Inpatient neurological units from three rehabilitation centres in the Netherlands. SUBJECTS: Nineteen subacute stroke patients (minus two drop-outs) with a severe motor deficit of the arm. INTERVENTIONS: All subjects underwent conventional rehabilitation care. Nine subjects additionally received a positioning procedure for two 30-min sessions a day, five days a week, for five weeks. MAIN MEASURES: Passive range of motion of five arm movements using a hydrogoniometer and resistance to passive movement at the elbow using the Ashworth Scale. Secondary outcome measures were pain at the end range of passive motions, the arm section of the Fugl-Meyer Assessment and Barthel Index scores for ADL-independence. Outcome measures were taken after five weeks and additional measurements after 10 weeks by two assessors blinded to group allocation. RESULTS: Comparison of the experimental (n = 9) with the control subjects (n = 8) after five weeks showed that additional positioning significantly slowed down development of shoulder abduction contracture (P = 0.042, -5.3 degrees versus -23 degrees). No other differences were found between the groups. CONCLUSIONS: Applying a contracture preventive positioning procedure for the hemiplegic arm slowed down the development of shoulder abduction contracture. Positioning did not show significant additional value on other outcome measures. Since the sample size was small, results of this study need future verification.