Table 4.

Predictors of alterations in levels of fatigue during 12-month follow-up selected with Random Forest algorithm (Figure 2) in 144 followed (44 worsened, 35 improved, 65 unchanged regarding fatigue) patients with systemic sclerosis using binary logistic regression analyses.

		Worsening (n = 44)			Improvement (n = 35)
		Odds Ratio	CI	p	Odds Ratio	CI	p
	Predictors of Worsening/Improvement Fatigue among Baseline Parameters.
Based on baseline data	rEUSTAR AI	0.825	0.616–1.087	0.183	1.374	1.034–1.850	0.030
	Cochin Hand Function Scale	0.982	0.949–1.014	0.289	1.037	1.003–1.073	0.031
	Disease duration	1.027	0.969–1.086	0.356	0.916	0.846–0.979	0.018
	Modified Rodnan Skin Score	0.901	0.806–0.995	0.052	0.924	0.837–1.011	0.101
	Interstitial Lung Disease	3.197	1.247–9.158	0.021	0.602	0.238–1.541	0.284
	Factors, which changes predict the worsening/improvement in fatigue levels.
1-year changing values	FACIT-EWB	0.868	0.776–0.962	0.009	1.338	1.157–1.593	<0.001
	SF36-PCS	0.951	0.916–0.982	0.004	1.045	1.003–1.094	0.042
	Cochin Hand Function Scale	1.020	0.946–1.100	0.590	1.006	0.923–1.082	0.876
	Hemoglobin	0.971	0.923–1.015	0.215	1.032	0.973–1.095	0.267
	Albumin	0.937	0.809–1.081	0.380	1.187	0.988–1.451	0.075
	UCLA SCTC GIT 2.0	1.953	0.383–10.478	0.422	0.077	0.007–0.616	0.020
	Pinch strength	1.004	0.967–1.041	0.822	1.043	0.999–1.091	0.055
	rEUSTAR AI	0.937	0.720–1.213	0.622	1.120	0.821–1.558	0.482

Statistically significant results are indicated in bold. CI, Confidence Interval; rEUSTAR AI, revised European Scleroderma Trials and Research Group Activity Index; FACIT-EWB, The Functional Assessment of Chronic Illness Therapy Emotional Well-being; SF36-PCS, Short Form 36 Physical Component Summary; GIT 2.0, University of California, Los Angeles Scleroderma Clinical Trial Consortium Gastrointestinal Tract Instrument 2.0.