Rasch Analysis of the 7-Item Fatigue Severity Scale (FSS–7) in Traumatic Brain Injury

OBJECTIVE: To evaluate validity, reliability, and targeting of the Norwegian 7-item Fatigue Severity Scale (FSS-7) in adults with moderate-to-severe traumatic brain injury (TBI). SETTING: Inpatient and outpatient acute care and postacute rehabilitation facilities. PARTICIPANTS: Ninety-four patients with intracranial injury, 20% women, with a mean age of 45.8 years (SD = 13.6), provided 185 complete FSS-7 responses at 6 months (n = 94) and 12 months (n = 91) post-TBI. DESIGN: Secondary analysis of data from a longitudinal observational study using the Rasch model and graphical log-linear Rasch models (GLLRM). MAIN MEASURES: The FSS-7 is a self-report fatigue measure derived from the original FSS by omitting items 1 and 2. Differential item functioning (DIF) was examined for sex, age, education, and sleep disturbances (Epworth sleepiness scale and insomnia severity index). RESULTS: The FSS-7 did not fit the Rasch model due to poor fit of item 4 and local dependence between items 5 and 6. A 6-item version (FSS-6), excluding item 4, provided good fit to a GLLRM accounting for the local dependence between items 5 and 6. No DIF or item parameter drift over time was detected. The FSS-6 demonstrated excellent reliability (r = 0.936), while targeting was less than optimal, as on average, only 57% of the maximum obtainable test information was reached. CONCLUSIONS: The FSS-6, excluding item 4, provides a valid measure of fatigue with excellent reliability for adults at 6 and 12 months post-TBI. The total raw score is invariant across sex, age, educational level, and sleep disturbances with no item parameter drift from 6 to 12 months postinjury. The FSS-6 may be useful for screening and monitoring fatigue in TBI rehabilitation and research, although a reassessment of targeting is recommended for future research on a larger and broader sample of patients.