Author: Amira Mandour
Biostatistician | Clinical Trials & Statistical Modeling
Expert
Introduction
Fatigue is a common and disabling symptom in people with multiple sclerosis (MS). This study compares two exercise interventions—Cognition-Targeted Exercise (CTE) and Symptom-Targeted Exercise (STE)—to evaluate their effectiveness in reducing fatigue over time.
The data used in this analysis are simulated to preserve participant confidentiality. The simulation is based on summary statistics from the real dataset.
Primary Objective
The aim of this study was to examine the effect of time and intervention type on MS-related fatigue and to explore the interaction between time and intervention.
The primary outcome is the Modified Fatigue Impact score measured at three time points: Baseline, Post-intervention, and Follow-up.
Secondary Objectives
To explore whether the longitudinal patterns of anxiety, perceived stress, and functional impairment differ between the CTE and STE groups over time.
Statistical Analysis
Generalized Estimating Equations (GEE) model was applied to evaluate how fatigue changed over time in the CTE and STE groups and whether these changes differed by intervention. Additional GEE models were used to assess group differences in anxiety, perceived stress, and functional impairment across the study period.
Generalized Estimating Equations (GEE) Model:
Outcome: Modified Fatigue Impact:
| Table 1. GEE Model Results | |||
| Outcome: Modified Fatigue Impact Scale | |||
| Characteristic1 | Beta1 | 95% CI2,1 | p-value1 |
|---|---|---|---|
| Group | |||
| STE | — | — | |
| CTE | -4.7 | -7.4, -2.1 | <0.001 |
| Time | |||
| Baseline | — | — | |
| Post | -4.7 | -7.6, -1.9 | 0.001 |
| Followup | -5.2 | -7.4, -3.0 | <0.001 |
| Group × Time Interaction | |||
| CTE * Post | -0.44 | -4.9, 4.0 | 0.8 |
| CTE * Followup | -4.3 | -8.6, -0.01 | 0.050 |
| 1 STE group and Baseline time point are reference categories. | |||
| 2 CI = Confidence Interval | |||
A Generalized Estimating Equation (GEE) model was used to evaluate the effects of exercise group (CTE vs. STE), time, and their interaction on fatigue scores. Both groups exhibited significant reductions in fatigue over time. At baseline, the CTE group reported significantly lower fatigue than the STE group. Notably, while no significant group difference was found immediately post-intervention, the CTE group demonstrated a significantly greater reduction in fatigue at follow-up compared to the STE group. Pairwise comparisons supported these findings, indicating that the long-term benefit of cognition-targeted exercise became statistically significant only at the follow-up assessment. These results suggest a delayed yet meaningful added effect of CTE on fatigue reduction. The visual trajectory further illustrates this pattern, with the CTE group maintaining a more sustained decrease in fatigue over time.
This line plot illustrates the estimated marginal means of fatigue scores across three time points for both the CTE (Cognition-Targeted Exercise) and STE (Symptom-Targeted Exercise) groups. Both groups demonstrated improvement over time; however, the CTE group exhibited a slightly greater and more sustained reduction in fatigue, particularly evident at the follow-up assessment. The error bars, representing 95% confidence intervals, provide a visual indication of the precision of the estimates and the variability in group differences at each time point. Overall, the figure emphasizes the potential long-term advantage of cognition-targeted exercise in reducing fatigue among individuals with MS.
Outcome: Work and Social Adjustment Scale:
| Table 2. GEE Model Results | |||
| Outcome: Work and Social Adjustment Scale | |||
| Characteristic1 | Beta1 | 95% CI2,1 | p-value1 |
|---|---|---|---|
| Group | |||
| STE | — | — | |
| CTE | -7.4 | -10, -4.7 | <0.001 |
| Time | |||
| Baseline | — | — | |
| Post | -4.9 | -7.3, -2.5 | <0.001 |
| Followup | -4.6 | -6.8, -2.3 | <0.001 |
| Group × Time Interaction | |||
| CTE * Post | 4.1 | 0.57, 7.6 | 0.023 |
| CTE * Followup | 2.4 | -1.7, 6.6 | 0.2 |
| 1 STE group and Baseline time point are reference categories. | |||
| 2 CI = Confidence Interval | |||
A GEE model was used to examine the effects of intervention group (CTE vs. STE), time (Baseline, Post-intervention, Follow-up), and their interaction on work and social adjustment, as measured by the WSAS. At Baseline, participants in the CTE group had significantly better work and social adjustment scores compared to the STE group (β = -7.38, p < 0.001). Both groups showed significant improvements in WSAS scores over time: At Post-intervention, scores decreased by 4.89 points (p < 0.001). At Follow-up, scores decreased by 4.59 points (p < 0.001), relative to Baseline. A significant group-by-time interaction was observed at Post-intervention (β = 4.06, p = 0.023), suggesting that the STE group showed greater short-term improvement than the CTE group. However, the interaction at Follow-up was not statistically significant (β = 2.43, p = 0.25), indicating that differences in improvement between groups were not maintained over time.
Both CTE and STE interventions were associated with improved work and social functioning over time among individuals with MS-related fatigue. Although the STE group showed greater short-term gains post-intervention, these differences were not sustained at follow-up. These findings suggest that both interventions may be beneficial, but the long-term trajectories of improvement are similar across groups.
At baseline, participants in the CTE group reported better work and social adjustment compared to those in the STE group. Over time, both groups showed significant improvements in WSAS scores. The STE group demonstrated a greater reduction in WSAS scores immediately post-intervention, as indicated by a significant group-by-time interaction. However, by follow-up, this difference was no longer statistically significant, suggesting that both interventions led to similar improvements in work and social functioning in the long term.
Outcome: Hospital Anxiety:
| Table 3. GEE Model Results | |||
| Outcome: Hospital Anxiety | |||
| Characteristic1 | Beta1 | 95% CI2,1 | p-value1 |
|---|---|---|---|
| Group | |||
| STE | — | — | |
| CTE | -5.4 | -7.5, -3.4 | <0.001 |
| Time | |||
| Baseline | — | — | |
| Post | -1.9 | -3.9, 0.10 | 0.062 |
| Followup | -2.3 | -4.0, -0.62 | 0.007 |
| Group × Time Interaction | |||
| CTE * Post | 1.7 | -1.1, 4.5 | 0.2 |
| CTE * Followup | 1.9 | -1.4, 5.2 | 0.3 |
| 1 STE group and Baseline time point are reference categories. | |||
| 2 CI = Confidence Interval | |||
A Generalized Estimating Equations (GEE) model was used to evaluate the effects of intervention group (CTE vs. STE), time (Baseline, Post-intervention, Follow-up), and their interaction on Hospital Anxiety scores. The CTE group had significantly lower anxiety scores at baseline compared to the STE group (β = -5.45, p < 0.001), indicating better initial emotional well-being. Over time, both groups showed some reduction in anxiety: At Post-intervention, scores decreased by 1.92 points, although this change was not statistically significant (p = 0.062). At Follow-up, scores decreased by a statistically significant 2.29 points relative to Baseline (p = 0.007), indicating improvement in anxiety symptoms over the longer term. The group-by-time interactions (CTE × Post and CTE × Follow-up) were not statistically significant, suggesting that the pattern of change over time was similar in both groups. Specifically, the CTE group did not differ significantly from the STE group in their anxiety score changes at either Post-intervention or Follow-up.
Both intervention groups (CTE and STE) demonstrated reductions in Hospital Anxiety over time, with a significant improvement observed by Follow-up. While the CTE group started with lower anxiety levels at baseline, the lack of significant interaction effects suggests that both interventions led to similar patterns of change in anxiety symptoms over time. These findings imply that both CTE and STE may be effective in reducing anxiety among individuals with MS-related fatigue, but neither showed a superior trajectory of improvement.
Participants in the CTE group had lower anxiety levels at baseline compared to the STE group. Both groups showed reductions in Hospital Anxiety scores over time, with significant improvement observed by follow-up. However, the pattern of change did not differ significantly between groups, suggesting similar effects of both interventions on anxiety over time.
Outcome: Perceived Stress:
| Table 4. GEE Model Results | |||
| Outcome: Perceived Stress | |||
| Characteristic1 | Beta1 | 95% CI2,1 | p-value1 |
|---|---|---|---|
| Group | |||
| STE | — | — | |
| CTE | -2.3 | -4.4, -0.23 | 0.030 |
| Time | |||
| Baseline | — | — | |
| Post | -4.8 | -7.3, -2.4 | <0.001 |
| Followup | -3.8 | -5.6, -2.1 | <0.001 |
| Group × Time Interaction | |||
| CTE * Post | 1.0 | -2.6, 4.6 | 0.6 |
| CTE * Followup | 1.3 | -2.6, 5.1 | 0.5 |
| 1 STE group and Baseline time point are reference categories. | |||
| 2 CI = Confidence Interval | |||
Participants in the CTE group had significantly lower stress levels at baseline compared to those in the STE group (β = -2.32, p = 0.030), indicating a baseline advantage in stress levels for the CTE group.
Over time, both groups showed statistically significant reductions in perceived stress: At Post-intervention, stress scores were 4.84 points lower than baseline (p < 0.001). At Follow-up, scores remained significantly lower, with a 3.84-point reduction from baseline (p < 0.001).
However, the interaction terms (group × time) were not statistically significant: CTE vs. STE at Post-intervention: β = 1.03, p = 0.575 CTE vs. STE at Follow-up: β = 1.26, p = 0.521
Conclusion:
Both cognition-targeted exercise (CTE) and symptom-targeted exercise (STE) interventions were effective in improving key outcomes related to MS-related fatigue, including fatigue severity, work and social functioning, hospital anxiety, and perceived stress. While CTE showed a greater long-term reduction in fatigue, particularly evident at follow-up, the trajectories of improvement in other outcomes—such as anxiety, stress, and functional impairment—were largely similar between the two groups. These results suggest that both interventions are beneficial, with CTE potentially offering an added advantage for sustaining fatigue relief over time.