Table 3.
Age- and Sex-Adjusted Associations of Individual Physical, Cognitive, and Mental Health Status Variables With Fatigue Symptoms at 6- and 12-Month Follow-up (N = 732)
| Variable (Scaled by 0.5 SD) | Mean Difference (95% CI) in Fatiguea [Positive Value = Less Fatigue] | P |
|---|---|---|
| Physical Component summary (SF-36v2) [∼6 points]b | 5.0 (4.6 to 5.4) | < .001 |
| Physical Functioning (FPI-SF) [∼0.5 point] | 4.7 (4.3 to 5.1) | < .001 |
| Cognition (MMSE) [∼1 point] | 0.9 (0.6 to 1.3) | < .001 |
| Mental Component summary (SF-36v2) [∼7 points] | 5.1 (4.7 to 5.5) | < .001 |
| PTSD symptoms (IES-R) [∼0.5 points] | –4.3 (–4.7 to –3.9) | < .001 |
| Anxiety symptoms (HADS-Anxiety subscale) [∼2.5 points] | –4.9 (–5.2 to –4.5) | < .001 |
| Depression symptoms (HADS-Depression subscale) [2.5 points] | –5.9 (–6.2 to –5.6) | < .001 |
For SF-36, FPI-SF, and MMSE, higher scores = better function; for IES-R and HADS, higher scores = greater symptoms. FPI-SF = Functional Performance Inventory-Short Form; HADS = Hospital Anxiety and Depression Scale; IES-R = Impact of Events Scale-Revised; MMSE = Mini-Mental State Examination; PTSD = posttraumatic stress disorder; SF-36v2 = Short Form-36 Version 2.
Each row reports the results of a separate regression model that evaluates the age- and sex-adjusted association of fatigue symptoms with the variable named in that row. All analyses evaluate fatigue symptoms and the variable named in the row at the same follow-up time point. Analyses were conducted by using a longitudinal time-averaged random effects regression model. Fatigue symptoms were measured by using the transformed score from the validated Functional Assessment of Chronic Illness Therapy-Fatigue Scale (range, 0-100), with higher scores representing less fatigue. Values presented represent the estimated mean difference in fatigue score for a 0.5 SD difference in the variable named in the row, over both 6- and 12-month follow-up time points.
Interpretation of this first row is as follows: “If Patient A had a Physical Component Summary score that was 0.5 standard deviations higher than Patient B, then Patient A’s expected fatigue score would be 5.0 points higher than Patient B.”