Minimal Clinically Important Difference for the Rasch Neuropsychiatric Inventory Irritability and Aggression Scale for Traumatic Brain Injury

Abstract

Objective To determine the minimal clinically important difference (MCID) for a Rasch measure derived from the Irritability/Lability and Agitation/Aggression subscales of the Neuropsychiatric Inventory (NPI)—the Rasch NPI Irritability and Aggression Scale for Traumatic Brain Injury (NPI-TBI-IA). Design Distribution-based statistical methods were applied to retrospective data to determine candidates for the MCID. These candidates were evaluated by anchoring the NPI-TBI-IA to Global Impression of Change (GIC) ratings by participants, significant others, and a supervising physician. Setting Postacute rehabilitation outpatient clinic. Participants 274 cases with observer ratings; 232 cases with self-ratings by participants with moderate-severe TBI at least 6 months postinjury. Interventions Not applicable. Main Outcome Measure NPI-TBI-IA. Results For observer ratings on the NPI-TBI-IA, anchored comparisons found an improvement of 0.5 SD was associated with at least minimal general improvement on GIC by a significant majority (69%–80%); 0.5 SD improvement on participant NPI-TBI-IA self-ratings was also associated with at least minimal improvement on the GIC by a substantial majority (77%–83%). The percentage indicating significant global improvement did not increase markedly on most ratings at higher levels of improvement on the NPI-TBI-IA. Conclusions A 0.5 SD improvement on the NPI-TBI-IA indicates the MCID for both observer and participant ratings on this measure.