Background: Integrated care pathways (ICPs) map the predicted course of an episode of patient care. They detail the expected interventions during the episode and document departures from the expected pathway (variance). This study describes the use of an ICP to audit the rehabilitation of patients with nontraumatic spinal cord injury admitted between 1997 and 1999. Methods: The ICPs and outcomes of 85 patients with nontraumatic spinal cord injury admitted to the Neurorehabilitation Unit at the National Hospital for Neurology and Neurosurgery, Queen Square, London were analysed. Data extracted included diagnosis, level of the lesion and duration of stay. The numbers and categories of goals and the rates of goal achievement were extracted and the variance patterns analysed. Results: An average of 28 patients were admitted each year. The level of disability on admission and the duration of stay decreased over the three-year period, while the average patient age increased from 48 to 54 years. None of these changes were statistically significant. On average each patient had three new goals set each week. Ninety per cent of all goals were achieved; this was not dependent on the category of goal. Sixteen patients (19%) accounted for 58% of all nonachieved goals. These patients tended to have acute-onset disability. The number of variances fell from 15 to 7 over the three-year period. Conclusions: The pathway enables monitoring of the rehabilitation process. As the Unit becomes more experienced there is a trend to shorter, more focused admissions with fewer variances. Specific groups of patients with particular needs can be identified. Future patients benefit from closure of the ‘audit loop’ and the implementation of clinical change based on information obtained from the ICP.