Physiological and pathological covariates of persistent concussion-related fatigue: results from two regression methodologies.

Background: Fatigue severity in persons with mild traumatic brain injury (mTBI) has received little research attention, despite its typically positively skewed nature. Investigation of covariates across a range of fatigue severity may provide insight into important contributors. Objective: To assess the relative significance of a priori-hypothesized covariates of physiological and pathological (mental and physical) fatigue in persons with mTBI/concussion, applying ordinary least squares (OLS) and quantile regression (QR) approaches. Methods: We conducted a cross-sectional investigation in 80 participants with mTBI/concussion (mean age 45.4 +/- 10.1 years, 59% male). The fatigue severity scale (FSS) was used as an outcome measure. Predictors of this outcome, grouped into physiological and pathological models of fatigue were assessed using OLS and QR. Results: The mean total FSS score was 46.13 +/- 14.59, and the median was 49 (interquartile range 37-57), demonstrating positive skewness. Fatigue severity was associated with variables within the mental, psychological and psychiatric domains at different levels of the fatigue score distribution. Conclusion: Results highlighted that some covariates had a significant impact on the FSS total score at non-central parts of its distribution, while others exhibited significant impact across the entire distribution. Addressing covariates of fatigue across the severity continuum can enhance research and clinical management.