Assessment of ability realization using the 4th version of the Spinal Cord Independence Measure

Context: Change in ability realization reflects the main contribution of rehabilitation to improvement in the performance of daily activities after spinal cord lesions (SCL). Objective: To adapt a Spinal Cord Ability Realization Measurement Index (SCI-ARMI) formula to the new Spinal Cord Independence Measure version 4 (SCIM4). Methods: Using data from 156 individuals for whom American Spinal Injury Association Motor Score (AMS) and SCIM4 scores were collected, we obtained an estimate for the highest possible SCIM4 given the patient’s AMS value, using the 95th percentile of SCIM4 values at discharge from rehabilitation (SCIM95) for patients with any given AMS at discharge. We used the statistical software environment R to implement the quantile regression method for linear and quadratic formulas. We also compared the computed model with the SCIM95 model obtained using data from the present study group, positioned in the SCIM95 formula developed for SCIM3. Results: The coefficients of the computed SCIM95 formula based on SCIM4 scores were statistically non-significant, which hypothetically reflects the small sample relative to the goal of estimating SCIM4 95th percentile. Predicting the ability using SCIM4 scores positioned in the SCIM95 formula used for SCIM3, however, yielded SCIM95 values, which are very close to those of the new SCIM95 formula (Mean difference 2.16, 95% CI = 1.45, 4.90). Conclusion: The SCI-ARMI formula, which is based on the SCIM95 formula developed for SCIM3, is appropriate for estimating SCI-ARMI at present, when SCIM4 scores are available. When sufficient additional data accumulates, it will be appropriate to introduce a modified SCI-ARMI formula.