BACKGROUND: Intravenous steroids are routinely used to treat disabling relapses in multiple sclerosis, and can be administered in an outpatient or home setting. We developed a rating scale that allowed us to compare the two strategies formally in a trial setting. METHODS: Patients who had a clinically significant multiple-sclerosis relapse within 4 weeks of onset were randomly assigned administration of a 3-day regimen of intravenous methylprednisolone either in an outpatient clinic (n=69) or at home (n=69). The MS relapse management scale (MSRMS) was developed to measure patients' experiences of relapse management as the primary outcome. Efficacy of the two treatment modalities was compared in terms of traditional measures and economic cost. A cost-minimisation analysis was also done. Analysis was by intention to treat. FINDINGS: Of 149 eligible patients, 138 consented to participate in the trial and were randomly assigned to a treatment group. Coordination of care was significantly better in the home-treatment group (median score 4.5 [IQR 3.0-11.4]) than in the hospital-treatment group (12.1 [3.0-18.6]; p=0.024). The other dimensions of the MSRMS did not differ between groups (p>0.10). Administration of steroids was equally safe and effective in either location, and cost was either the same or cheaper when delivered at home than when delivered in hospital. INTERPRETATION: Treatment of relapses in multiple sclerosis with intravenous steroids can be effectively and safely administered at home, from both patient and economic perspectives. Moreover, the trial indicates the importance of explicit and valid outcome measures of all aspects of service delivery when making decisions about health policy. This finding has implications for complex service delivery care models for long-term diseases.