Background/Purpose: Fatigue is a prevalent, debilitating symptom across chronic diseases that is difficult to treat and negatively impacts patients’ quality of life. Patients with rheumatologic diseases, such as systemic sclerosis (SSc), report high levels of fatigue. The 29-item Fatigue Assessment Instrument (FAI) is an expanded version of the 9-item Fatigue Severity Scale (FSS). The FAI is multidimensional and yields 4 subscales measuring different components of fatigue (global severity, situation specific, psychological consequences, responds to rest/sleep), while the FSS is unidimensional and has been more widely used to assess fatigue in chronic disease populations. However, there have been few psychometric studies of the FAI and FSS in SSc, and structural validity has yet to be assessed. The present study examined the previously established 1-factor structure of the FSS and the 4-factor structure of the FAI in an SSc sample.

Methods: SSc patients (N = 390; 84.1% female; 59.0% diffuse SSc; 52.1% White, 26.2% Hispanic, 17.7% Black) were sampled from the Genetics versus ENvironment In Scleroderma Outcome Study (GENISOS). GENISOS is a multicenter, prospective cohort study of early SSc patients based in Texas. All patients fulfilled the ACR/EULAR classification criteria for SSc and had disease duration < 5 years at enrollment. Confirmatory factor analysis was used to assess the 1-factor structure of the FSS and the 4-factor structure of the FAI. Cronbach's coefficient alpha and McDonald's omega were used to evaluate internal consistency reliability.

Results: The 1-factor model of fatigue demonstrated good model fit for the FSS based on 2 indices (Comparative Fit Index [CFI] = .96, Standardized Root Mean Residual [SRMR] = .04). The Root Mean Square Error of Approximation (RMSEA) indicated slightly less than acceptable model fit (RMSEA = .09). All standardized factor loadings for items were statistically significant (all ps < .001), with values ranging from .338 to .842. The 4-factor model of fatigue demonstrated acceptable model fit for the FAI based on 2 indices (SRMR = .07, RMSEA = .08). However, the Comparative Fit Index (CFI) indicated less than acceptable model fit (CFI = .86). All standardized factor loadings for items were statistically significant (all ps < .001), with values ranging from .23 to .82. The FSS (α = .89, ω = .90) and the global fatigue severity subscale of the FAI had adequate internal consistency reliability (α = .92, ω = .92). The other 3 FAI subscales had Cronbach's coefficient alpha and McDonald's omega values ranging from .62 to .69.

Conclusion: Findings provided support for the expected factor structure of both the FSS and the FAI. However, given the superior model fit of the FSS, the lower internal consistency reliability of 3 subscales of the FAI, and that only 22 of the 29 items on the FAI are scored, the FSS is recommended for use as a general, more efficient measure of fatigue in SSc patients.

« Back to ACR Convergence 2021

ACR Meeting Abstracts - https://acrabstracts.org/abstract/confirmatory-factor-analysis-of-the-fatigue-assessment-instrument-and-fatigue-severity-scale-in-systemic-sclerosis/