Walking index for spinal cord injury (WISCI): criterion validation

Retrospective examination.

To compare the Walking Index for Spinal Cord Injury (WISCI) and current scales for their sensitivity to walking changes in subjects with a spinal cord lesion (SCL) and further validate the WISCI for use in clinical trails.

A large rehabilitation hospital in the center of Italy.

Retrospective review was performed on 284 patient records with an SCL. Measurements included neurological evaluation with Lower Extremity Motor Scores (LEMS) according to the American Spinal Injury Association (ASIA) and walking status assessed by Barthel Index (BI (0-15)), Rivermead Mobility Index (RMI (three levels)), Functional Independence Measure (FIM (1-7)), Spinal Cord Independence Measure (SCIM (0-8)), and WISCI (0-20). The WISCI is a 21-level hierarchical scale which incorporates gradations of physical assistance and devices required for walking. Improvement in walking is based on the change of scores from admission to discharge. Statistical analysis included Spearman rank correlation and chi2 test; P<0.05.

There was a significant positive correlation between WISCI and other scales (WISCI and BI r=0.67, P<0.001; WISCI and RMI r=0.67, P<0.001; WISCI and SCIM r=0.97, P<0.001; WISCI and FIM r=0.7, P<0.001). The initial ASIA grade was predictive of mobility outcome on the WISCI: of the 78 ASIA A patients, only five achieved independent walking versus 4/17 ASIA B (P=0.02), 56/109 ASIA C (P<0.001) and 39/44 ASIA D (P<0.001). The correlation of LEMS to the WISCI was 0.58 (P<0.001). At discharge, patients were distributed into 12 WISCI levels versus four FIM, three BI, two RMI and five SCIM levels. The most frequent WISCI levels at discharge were 13 (walker, no braces or assistance), 16 (two crutches, no braces or assistance) and 20 (no devices or assistance).

Similar correlation between the WISCI and the other scales indicates that all these measures address the same concept, mobility, which is a measure of concurrent validity. The correlation is not 100% because of conceptual differences (the WISCI incorporates gradations of physical assistance and devices required for walking while most of the other scales focus on burden of care or mobility in the environment). The WISCI is more detailed and appears more sensitive to walking recovery than the other scales, as demonstrated by our patients' score distribution at discharge. Within each of the most frequent WISCI levels (13, 16, 20) LEMS and other walking features varied; therefore the scale would benefit from further refinement based on speed, distance and energy cost.