Validation of a clinical prediction rule for ambulation outcome after non-traumatic spinal cord injury.

Abstract

OBJECTIVES: To validate a Clinical Prediction Rule (CPR) for ambulation in a non-traumatic spinal cord injury population (NTSCI).

SETTING: Tertiary spinal rehabilitation inpatient service, Melbourne, Australia.

METHODS: Adults with confirmed NTSCI were recruited between April 2013 and July 2017. Data based on the original van Middendorp CPR (age and four neurological variables) were collected from participant's medical records and by interview. The Spinal Cord Independence Measure item 12 was used to quantify the ability to walk at 6 and 12 months. A receiver operator curve (ROC) was utilised to determine the performance of the CPR. Ambulatory outcomes were compared for AIS A, B, C and D and aetiology groups.

RESULTS: The area under the ROC curve (AUC) was 0.94 with 95% confidence interval (CI) 0.86-1.0 (n = 52). Overall accuracy was 75% at 6 months and 82% at 12 months. For the whole cohort the sensitivity at 12 months was 95% and specificity 73%. However, specificity for AIS C and D was only 50%.

CONCLUSION: The CPR correctly predicted those who did not walk at 6 and 12 months following NTSCI, but was less accurate in predicting those who would walk particularly those with an AIS C or D classification. This CPR may be useful to inform planning for future care in individuals with NTSCI, particularly for those who are not expected to walk. Further research with larger sample sizes is required to determine if the trends identified in this study are generalisable.