Resilience Indirectly Affects Functional Capabilities Through Physical Activity Engagement in Individuals With Multiple Sclerosis Free

In 2020, it was estimated that 2077 people per month were diagnosed with multiple sclerosis (MS) in the United States.¹  Several mechanisms can contribute to MS progression, including inflammation, neurodegeneration, and demyelination.²  Individuals can experience several symptoms, including fatigue, mood changes, and impaired mobility, which may increase in severity with greater disease duration.³  Not only do individuals with MS report having more difficulty completing their daily activities, but they report that several of their symptoms contribute to their limitations.⁴  Many individuals also report that their diagnosis had a negative impact on their emotional well-being, self-esteem, coping ability, and future outlook.⁴ 

That said, individuals with MS who are resilient exhibit lower psychological distress,⁵  describing themselves as “thriv[ing]” and “liv[ing] the best life possible,”⁶  and they report being able to maintain stability in their well-being.⁷  Resilience in MS is highly driven by psychological factors (eg, self-efficacy and affect).⁷  Social support and physical well-being indirectly contribute to resilience through the psychological components, acting as barriers or facilitators.^6,7  As resilience has been associated with better gross motor functioning in MS,⁸  it has raised the question of whether resilience contributes to functional capabilities.⁹  However, it has also been suggested that lifestyle behaviors, not resilience itself, are the underlying contributors.¹⁰ 

The importance of healthy lifestyle behaviors has been well documented in MS. Physical activity, for example, has been associated with higher health-related quality of life (HRQOL)^11,12  and improved motor functioning.¹³  A healthy diet has also been positively associated with HRQOL.¹¹  Conversely, unhealthy behaviors (eg, alcohol misuse and smoking) can have negative effects on MS-related disability.¹⁴  Resilient individuals with MS tend to engage in more healthy lifestyle behaviors (ie, more physical activity and healthier diet choices).¹⁵  The role of lifestyle behaviors was also recognized as a component of MS-related resilience in the development of the MS Resiliency Scale (MSRS).¹⁶ 

It can be conceptualized that resilience indirectly influences functional capabilities, with that relationship explained by lifestyle behaviors. Theoretically, as resilient individuals have higher self-efficacy and more effective coping strategies,⁷  this would contribute to their healthy lifestyle behaviors¹⁵  and thus lead to better functional outcomes. The rationale for this pathway is supported by social cognitive theory–based interventions for improving physical activity in MS, which show connections between engagement, self-efficacy, and goal setting.¹⁷  Not only is self-efficacy a significant factor in resilience among individuals with MS,^7,18  but goal setting is a component of at least 1 MS-specific resilience-building program.¹⁹ 

While there is a theoretical basis for this pathway, it has yet to be tested empirically in a mediation model. Furthermore, it is unknown which specific lifestyle behaviors mediate the relationship between resilience and functional capabilities. Therefore, this study aimed to explore whether resilience might indirectly influence functioning in individuals with MS through specific lifestyle behaviors.

This study was a planned secondary analysis of a test-retest reliability investigation of the MSRS.¹⁶  Participants were individuals with MS who had agreed to be contacted about research at the Joyce D. and Andrew J. Mandell Center for Comprehensive Multiple Sclerosis Care and Neuroscience Research at Trinity Health Of New England. In addition to an MS diagnosis, eligible individuals were between the ages of 18 and 89 years, able to read and respond to questionnaires in English, and had access to both the internet and a valid email address for the electronic surveys. Data from the first time point (n = 64) were used for these analyses. While the sample size determination was based on the parent study’s primary aim,¹⁶  a post hoc power analysis conducted for the present analyses suggested that it could detect a large effect in a linear regression with 7 predictors with 92.5% power and a 5% error rate.²⁰ 

The Trinity Health Of New England Institutional Review Board approved all study procedures and data were collected through the University of Connecticut Health Center-hosted REDCap electronic data capture tools.^21,22  Age, sex, race, MS type, and disease duration were extracted from the parent study. The impacts of fatigue, depression, and anxiety, which can influence the relationship between resilience and functioning,¹⁰  were evaluated using their respective items on the SymptoMScreen, a 12-item self-reported assessment.²³  Participants rated how much their MS has affected each symptom, using a 7-point Likert scale (not affected at all to total limitation/I’m unable to do most daily activities). The individual items have been validated against the Performance Scales (fatigue and depression),^23,24  Patient Health Questionnaire-8 (depression),²⁴  and Generalized Anxiety Disorder-7 (anxiety).²⁴  A composite score was created, which had good internal consistency (α = 0.85); higher scores were indicative of greater symptom impact.

Resilience was measured with the Multiple Sclerosis Resiliency Scale (MSRS),²⁵  an MS-specific multidimensional questionnaire that includes psychological, physical, and social risk and protective factors of resilience. For each item, participants rated their level of agreement (strongly agree to strongly disagree) using a 4-point Likert scale. To reduce the influence of the Physical Activity and Diet subscale, which taps into some lifestyle behaviors, those 3 items were excluded. The modified 22-item version had good internal consistency (α = 0.88), with higher scores indicating greater resilience.

Physical activity engagement, dietary choices, and tobacco use were assessed with the Simple Lifestyle Indicator Questionnaire (SLIQ)²⁶ ; alcohol consumption was evaluated using the Alcohol Use Disorders Identification Test (AUDIT-C).²⁷  Both measures have been previously used as self-reported lifestyle behavior measures in MS.^28,29  For smoking, participants were categorized as nonsmokers, current smokers, and previous smokers. Higher scores on the SLIQ Physical Activity and Diet subscales indicated more physical activity engagement and healthier dietary choices, respectively, while higher AUDIT-C scores reflected more negative alcohol-related behaviors.

The Patient-Reported Outcomes Measurement Information System Physical Function-Short Form 20a (PROMIS-PF)³⁰  was used to measure functional abilities. Participants rated how much difficulty they experienced completing different activities of daily living and their health-related physical limitations. Raw scores (range, 20-100) were converted to t scores (range, 9.2-62.7) using a conversion table, with higher scores indicating better functioning. A score of 50 (SD = 10) is considered average for the general population.³¹  In the current sample, the PROMIS-PF had strong internal consistency (α = 0.97) and was highly correlated with the Patient-Determined Disease Steps (PDDS; ρ = –0.83, P < .001) scale.

Analyses were conducted with IBM SPSS version 26 and there were no missing values. Bivariate analyses were conducted by examining the relationships between the PROMIS-PF, MSRS, and individual lifestyle behaviors. Spearman correlations (ρ) were used for nonnormally distributed variables (SLIQ Physical Activity and AUDIT-C), Pearson correlations (r) for normally distributed variables (SLIQ Diet and MSRS), and analysis of variance (ANOVA) for categorical variables (SLIQ Smoking). If a significant relationship was observed (P < .05), mediation analyses were run using the Hayes PROCESS modeling tool,³²  with the MSRS as the independent variable, individual lifestyle behaviors as the mediator, and PROMIS-PF as the dependent variable. The modified SymptoMScreen, age, sex, MS type, and disease duration were included as covariates based on the literature.^10,15  The mediation analyses generated a total effect for the MSRS, which is the summation of the direct and indirect effects,³²  with the parameter estimates in the total effect model equivalent to a linear regression without the mediators. An indirect effect was considered significant if its confidence interval did not cross 0.³³  All assumptions for a linear regression were met.

Participants (TABLE) were, on average, 50.91 years old (SD = 11.12) and had a disease duration of 13 years (IQR = 15.00). The majority were White (81.3%), women (78.1%), and had relapsing-remitting MS (84.4%). Functional capabilities were positively associated with physical activity (ρ = 0.43, P < .001) and resilience (r = 0.37, P = .003). There was not a significant relationship between functioning and alcohol misuse (ρ = 0.15, P = .239), smoking (F[2,61] = 0.83, P = .440), or diet (r = 0.02, P = .848). In the mediation model, the total effect of resilience on functioning was nonsignificant (b = 0.13; 95% CI, –0.12 to 0.39; P = .295). Instead, symptom impact and progressive MS were associated with lower capabilities (FIGURE). However, resilience had a significant indirect effect on functional capabilities through physical activity (b = 0.16; 95% CI, 0.04-0.32), whereas its corresponding direct effect on functioning was nonsignificant.

Of the lifestyle behaviors examined, only physical activity was associated with functional capabilities. Consistent with previous work,¹⁰  mood and fatigue, rather than resilience, had a significant direct effect on functioning. This is also true for a diagnosis of progressive MS, which has also been associated with greater functional limitations.³⁴  However, resilience did have a positive, significant indirect effect, with physical activity explaining its relationship with functioning. This is consistent with the hypothesis that higher resilience would be associated with greater physical activity engagement,¹⁵  which in turn would be connected to better functional capabilities.¹³ 

As part of the comprehensive care model for MS, both pharmacological and nonpharmacological treatments are necessary.³⁵  While some individuals may be predisposed, resilience in response to chronic health conditions can be learned over time via intervention.³⁶  As such, fostering resilience may be an approach to promote physical activity engagement and indirectly improve functional capabilities. Despite the benefits of physical activity,^11–13,37  many individuals with MS do not meet the recommended 150 minutes per week.³⁸  Several physical activity barriers (eg, limited social support, low confidence, and depression)^38,39  could be addressed through a resilience-building intervention focused on increasing coping strategies and support systems.^19,40–44  Some programs specifically address physical activity engagement.⁴⁵  However, while improvements in physical HRQOL have been found,^40,42  to date, no study has included physical activity as an outcome. Future studies should investigate whether resilience-building interventions can be used to improve physical activity engagement and, subsequently, functional outcomes in individuals with MS.

Given their effects on MS-related disability,¹⁴  it was unexpected that alcohol misuse and smoking were not related to functioning; however, this may be due to less than 5% of participants being either current smokers (n = 3) or having unhealthy alcohol use (n = 2), defined as a score of 5 or above on the AUDIT-C.⁴⁶  In addition, participants may have made less healthy dietary choices, reducing the strength of the relationship between diet and functional outcomes. Although there was a positive association between diet and resilience (r = 0.25, P = .049), the average diet score was 7.02 (SD = 2.75). Comparatively, resilient older adults with MS had an average score of 9.86 (SD = 3.40) on the same measure, while those with low resilience had an average score of 8.98 (SD = 3.13).¹⁵  As such, the relationship between resilience, lifestyle, and functioning should be reexamined in a larger sample with a greater representation of certain behaviors.

The small sample size should also be taken into consideration when interpreting the results of the mediation, as it was only powered to detect large effects with the number of included variables. As such, the finding that resilience does not have a direct effect on functional capabilities should be interpreted with caution and verified in a larger sample. While the model was theoretically based, the data were cross-sectional; thus, causal conclusions cannot be definitively made. Furthermore, it is likely that some of the relationships examined are bidirectional (eg, lower functional capabilities contributing to decreased physical activity engagement). Functional capabilities were also measured via self-report. Although self-report and performance-based functional measures are correlated,⁴⁷  there is not a perfect agreement. While the depression and anxiety items of the SymptoMScreen have been validated against traditionally used scales,²⁴  the fatigue item has not been validated against measures such as the Modified Fatigue Impact Scale. In addition, most participants were White and had a relapsing form of MS, limiting the generalizability of the results. Finally, as the SLIQ Physical Activity score was used, a closer examination of the type and level of physical activity could not be done. Future research may examine the specific types of physical activity, confirmed objectively with an accelerometer, and their relationships with resilience and functioning.

Evidence suggests that resilience may indirectly contribute to functional capabilities through physical activity engagement. These findings highlight the important role of resilience in promoting well-being in individuals with MS and potential pathways for increasing activity engagement.

The authors wish to thank Dotty Wakefield, MS, for coding and management of the REDCap data collection; Albert C. Lo, MD, PhD, for his review of the paper; and Jennifer A. Ruiz, PT, DPT, and Thomas Agresta, MD, MBI, for their input on the initial grant.