Distribution-based estimates of clinically significant changes in the International Standards for Neurological Classification of Spinal Cord Injury motor and sensory scores

Abstract

Background: Although the psychometric properties and statistical significance of the International Standards for Neurological Classification of Spinal Cord Injury Patients (ISNCSCI) have been widely examined, the clinical significance of motor and sensory scores (i.e., the improvement in score that has a meaningful impact on patients) is unknown.

Aim: To calculate the clinical significance of the International Standards for Neurological Classification of Spinal Cord Injury Patients (ISNCSCI).

Design: Analysis of prospectively collected data.

Setting: Spinal Cord Unit of a rehabilitation hospital in the centre of Italy.

Patients and methods: Analysis of the data of 600 patients with registration of the ISNCSCI Motor scores (total score and separately upper and lower extremity scores) and ISNCSCI Sensory scores. Clinical significance was calculated per several distribution-based approaches: minimal important differences, effect size-based estimates for small and substantial changes, standard error of measurement, and minimal detectable change. The calculated clinical significance was compared with improvements by the patients to determine the percentage of patients who achieved significant improvement. Furthermore, the functional status (as evaluated by the Spinal Cord Independence measure [SCIM III]) was studied in patients who achieved significant improvement compared to those who did not achieve them.

Results and discussion: The results of the study showed that motor scores were more amenable to change than sensory scores. A 5-point change in motor score resulted in a clinically significant improvement of 0.2 standard deviation units, and an 11-point change in motor score was associated with an improvement of 0.5 standard deviation units. The percentages of patients with a significant improvement varied from 8 to 80% according to the level and severity of the lesion. In some AIS grade/level of lesion groups, patients who achieved clinical significant scores also showed a better functional status with significantly higher SCIM III scores than those who did not achieve clinical significant scores: the functional relevance of the ISNCSCI change is maximal for AIS C cervical and thoracic lesions, whereas it appears to be limited for AIS A cervical and thoracic lesions, and null for lumbar lesions of any kind.

Conclusion: The results of the study provide information on the amount of motor and sensory scores that can be considered as true and clinical significant, as well as on the percentage of patients that achieve clinical significant improvements. Worthy to note, most groups of patients showed a statistically significant improvement of MS and SS between admission and discharge, but in some of the groups (for example thoracic AIS A patients), this improvement was limited to a very low percentage of patients.

Clinical rehabilitation impact: Our results provide useful benchmarks for clinicians and researchers with which changes in patient ISNCSCI motor and sensory scores can be interpreted as true and clinically meaningful, thus allowing a clinical judgment on interventions based on patients' progress. The proportion of patients with clinically significant improvements may be a useful benchmark in clinical trials: an intervention should be considered to be effective not only if it produces a greater statistically significant improvement in neurological status than another intervention or the natural course of the lesion, but also if it effects an increase in the percentage of subjects who achieve a clinically significant improvement.