Skip to main content

Table 2 Change scores and standardized response means for fatigue measures

From: Comparative responsiveness and minimally important difference of Fatigue Symptom Inventory (FSI) scales and the FSI-3 in trials with cancer survivors

Fatigue measure Global category MBSR trials (N = 106, Worse n = 11, Same n = 34, Better n = 59) BEAT Cancer trial (N = 222, Worse n = 37, Same n = 89, Better n = 89)
Change score (SD) p value SRM (95% CIs) Change score (SD) p value SRM (95% CIs)
FSI Total
 Worse − 0.70 (1.57) 0.631 − 0.45 (− 1.16, 0.31) 1.22 (1.37)  < 0.001 0.89 (0.53, 1.20)
 Same − 0.98 (1.70) − 0.58 (− 0.83, − 0.28) 0.05 (1.15) 0.05 (− 0.16, 0.26)
 Better − 1.96 (1.74) 0.010 − 1.13 (− 1.47, − 0.81) − 1.19 (1.32)  < 0.001 − 0.90 (− 1.10, − 0.70)
FSI Severity
 Worse − 0.55 (1.75) 0.759 − 0.31 (− 1.10, 0.37) 1.22 (1.29)  < 0.001 0.95 (0.63, 1.28)
 Same − 0.75 (1.95) − 0.38 (− 0.70, − 0.08) 0.12 (1.37) 0.09 (− 0.13, 0.29)
 Better − 1.65 (1.88) 0.031 − 0.88 (− 1.18, − 0.57) − 1.27 (1.62)  < 0.001 − 0.78 (− 0.99, − 0.57)
FSI Interference
 Worse − 0.87 (1.85) 0.769 − 0.47 (− 1.28, 0.28) 1.19 (1.74)  < 0.001 0.69 (0.37, 0.99)
 Same − 1.06 (1.89) − 0.56 (− 0.81, − 0.26) 0.03 (1.46) –−  0.02 (− 0.19, 0.23)
 Better − 2.16 (2.01) 0.011 − 1.08 (− 1.40, − 0.74) − 1.11 (1.42)  < 0.001 − 0.78 (− 0.99, − 0.57)
FSI-3
 Worse − 0.45 (1.92) 0.332 − 0.24 (− 0.97, 0.51) 1.41 (1.66)  < 0.001 0.86 (0.47, 1.20)
 Same − 1.20 (2.25) − 0.53 (− 0.82, − 0.21) 0.18 (1.34) 0.13 (− 0.08, 0.33)
 Better − 2.18 (2.05) 0.035 − 1.06 (− 1.44, − 0.64) − 1.51 (1.56)  < 0.001 − 0.96 (− 1.18, − 0.76)
  1. MBSR = Mindfulness-based stress reduction. BEAT Cancer trial = Better Exercise Adherence after Treatment for Cancer trial. SRM = Standardized Response Mean. FSI = Fatigue Symptom Inventory
  2. “Worse” and “Better” are defined as ≥ 1 SEM decrease/increase from baseline to follow-up in the Short Form-36 Vitality subscale. Follow-ups for all trials occurred post-intervention (i.e., 2 months post-baseline for MBSR trials and 3 months post-baseline for the BEAT Cancer trial). P-values are from pairwise t-tests of change scores in the “Worse” or “Better” group compared with the change score in the “Same” group for each measure. Each SRM was calculated as the change score/SD of the change score. The 95% CIs for each SRM were calculated via bias-corrected bootstrapping with 10,000 samples using MedCalc software