Facilitators and Barriers and Self-management in Adults With Multiple Sclerosis

Background: An estimated 2.3 million individuals affected worldwide are affected by multiple sclerosis. Due to advancements in treatment and earlier diagnosis, there are higher numbers of people living with MS than ever before. There is evidence linking optimal self-management to better outcomes among patients with chronic illnesses; however, regarding adults with MS specically, the results are inconclusive. The purpose of this study was to determine whether self-management mediates the relationship between facilitators and barriers (comorbidity, condition severity, pain, fatigue, and cognitive decits) and quality of life. Methods: To complete this cross-sectional correlational quantitative study, the researcher recruited a sample of 196 U.S. adults diagnosed with MS and administered a survey consisting of the Comorbidity Questionnaire for MS, the Patient-Determined Disease Steps, the Numeric Pain Intensity Scale, the Fatigue Severity Scale, the Multiple Sclerosis Neuropsychological Screening Questionnaire, the Multiple Sclerosis Self-Management Scale-Revised, the Multiple Sclerosis International Quality of Life, and a demographic questionnaire. The results were analyzed using multiple linear regression analysis. Results: The studied facilitators and barrier had statistically signicant relationship(s) with the dependent variable of QoL; however, the relationship between fatigue, pain, and QoL was not supported. A relationship exists between comorbidity, condition severity, and symptomatology and self-management. Finally, there is a signicant relationship between self-management and QoL; however, self-management did not mediate the relationship between the comorbidity, condition severity, and symptomatology (pain, fatigue, and cognitive decits) and QoL. Conclusions: By providing an understanding of the factors that hinder or facilitate MS self-management, the ndings can inform the development of interventions and the improvement of health outcomes for adults with MS.

Factors Related to Self-Management Several factors may in uence self-management in MS. Despite evidence for improved clinical outcomes with the effective medical treatment (e.g., better medications) of chronic conditions such as diabetes and asthma, suboptimal self-management and health outcomes continue among many chronic illnesses, including MS ( Comorbidity, condition severity, and symptomatology (pain, fatigue, and cognition) may affect one's ability to self-manage chronic disease, increasing the complexity of care needed for individuals to successfully achieve self-management and improve health outcomes (Grey et al., 2015). Regarding other chronic conditions, researchers have found self-management connects to comorbidity, condition severity, and symptoms (Grey et al., 2015; Schulman-Green, Jaser, Park, & Whittemore, 2015). More experts are recognizing self-management as a key component in improving the overall health of populations with multiple chronic conditions and comorbidities (Liddy, 2014). The compelling evidence, therefore, favors self-management as an effective strategy for the management of MS.

Self-Management and Health Outcomes
Evidence shows a positive relationship between the self-management of MS and improvements in clinical outcomes, including improved QoL, reduced disability, and lower treatment costs. Additionally, experts recognize self-management as an effective strategy that facilitates the achievement of highquality, long-term care, improved outcomes, and the appropriate utilization of services while ensuring increased patient con dence, reduced anxiety and unplanned admissions, and improved adherence to treatment and medication regimens (Reidy et al., 2016).
According to the self-and family management model, disease severity, symptoms (pain, fatigue, and cognition), and comorbidities in uence self-management needs (Grey et al., 2006). Comorbidities in MS, the severity of the disease, and the associated symptomatology (pain, fatigue, and cognitive problems) are potential factors (facilitators and barriers) affecting the transition to the process of self-management (Grey et al., 2015).
Gap in the Literature Throughout the literature, optimal self-management appears connected to better outcomes among patients with chronic illnesses, including reported QoL ( . Despite the compelling emerging evidence, the evidence regarding self-management in adults with MS is still inconclusive. It is not possible to conclude that the process of self-management directly affects health outcomes (e.g., QoL) when self-management undergoes in uence presumably by other factors (e.g., condition severity, comorbidity, and symptomatology). Although the self-management of pain, fatigue, and cognition is necessary with other chronic diseases, and the emerging literature on MS shows the relationships between such factors, there are remaining knowledge gaps regarding self-management's mediating effects on distal health outcomes (Bosworth et al., 2010;Fraser et al., 2013;Gao & Yuan, 2011).
Current researchers in self-management have not clearly demonstrated the links among MS symptoms, condition severity and comorbidities, self-management abilities, and QoL. Despite emerging studies on self-management in MS, experts know little about how disease-speci c factors predict self-management ability in adults with MS, and subsequently how self-management ability predicts QoL. Elucidating these issues is critical for improving overall QoL, and previous researchers indicated that critical factors affect self-management in individuals with other types of chronic conditions, subsequently affecting QoL. Thus, there is a need for further research on the associations among these factors in adults with MS.

Aims of the Study
The researcher's objective in conducting this cross-sectional correlational quantitative study was to determine whether self-management mediates the relationship between facilitators and barriers (comorbidity, condition severity, pain, fatigue, and cognitive de cits) and QoL in adults with MS.
Speci cally, the researcher designed the study to achieve know whether comorbidity, condition severity, and symptomatology (pain, fatigue, and cognitive de cits), are negatively associated with quality of life in adults with MS are investigated. In addition, the researcher was interested in determining whether these factors are signi cantly associated with self-management ability in adults with MS. In addition, the researcher explored whether self-management ability can affect the quality of life in adults with MS and serve as a mediator on the effects of comorbidity, condition severity, and symptomatology on the quality of life in adults with MS.
With an improved understanding of self-management in MS, the researcher obtained much-needed knowledge to begin development of an appropriate o ce-based assessment tool speci c to adults with MS. In addition, the ndings of this investigation contributed to existing theoretical and practical understanding of self-management in MS nursing, utilizing self-and family-management theories.

Methods Design
The researcher selected a cross-sectional correlational design to examine the relationships among the variables of interest in a sample of adults with MS. In this type of design, the researcher takes measurements corresponding to the independent and dependent variables for each participant at a single point in time and performs analyses to reveal statistically signi cant associations.

Participants
The target population included U.S. adults with MS aged 18 to 50 years. Recruitment involved a nonprobability sampling method for a minimum of 98 participants through Qualtrics, an online platform for collecting survey data. The inclusion criteria for participants were as follows: (a) medical diagnosis of MS for a minimum of 6 months, (b) aged 18 to 50 years, (c) able to speak and write in English, and (d) access to and pro ciency in the use of a computer. The exclusion criteria for the study included the following: (a) individuals with severe MS who were physically or cognitively unable to function independently, and (b) individuals with moderate to severe psychological impairments based on the initial cognitive screenings conducted by Qualtrics. This study was quantitative in nature; therefore, the researcher paid attention to minimizing threats to objectivity. This included making all efforts to follow a precise data collection plan. Once a minimum of 160 completed surveys met the criteria for completeness and underwent validation for eligibility, the researcher ended data collection.

Description of Materials
The surveys in the study included the following: Demographics questionnaire. The researcher used a demographics questionnaire to collect sociodemographic data. The questionnaire required participants to indicate their age, gender, educational level, and marital status. The researcher also collected additional clinical variables, including duration of the illness and MS subtype.
Comorbidity Questionnaire for Multiple Sclerosis. Use of a portion of the Comorbidity Questionnaire for Multiple Sclerosis probed the presence of the following conditions: depression, anxiety, hypertension, hyperlipidemia, and chronic lung disease. The next step was combining the responses from this survey into a single numeric value for each participant to represent the variable comorbidity. As such, the researcher represented comorbidity as an ordinal variable with possible scores ranging from 0 (no other conditions besides MS) to 5 (participant has all ve comorbid conditions).

Patient-Determined Disease
Steps. This survey quanti es the severity of neurological impairment, with patients asked to rate the severity of their disease, mainly focusing on how well they walk. The PDDS is a self-report version of the physician-reported Disease Steps developed by Hohol, Oray, and Weiner (1995), modi ed by NARCOMS. A rating of 0 indicates mild symptoms of MS, ratings of 1 to 4 indicate moderate symptoms, where patients are still fully ambulatory, and a rating of 8 indicates that the patient is bedridden. The PDDS has well-established validity, with scores strongly correlated with EDSS scores (p = .783). PDDS and EDSS scores both strongly correlated with Pyramidal (p = .578 and p = .647, respectively) and Cerebellar (p = .501 and p = .528, respectively) Functional System (FS) scores, as well as 6-minute walk (MW) distance (p = .704 and p = .805, respectively), MSWS-12 scores (p = .801 and p = .729, respectively), and accelerometer steps/day (p = -.740 and p = -.717, respectively; Learmonth, Motl, Sandroff, Pula, & Cadavid, 2013). For the present study, an individual's PDDS score represented the ordinal variable condition severity. Overall, a strong correlation between EDSS and PDDS scores (p = .783, 95% CI = .691, .850, p = .0001) existed (Learmonth et al., 2013).
Numeric Pain Intensity Scale. The NPIS asks participants to rate the average pain intensity experienced from their point of view in the last 24 hours (Michalski et al., 2011). Respondents indicated their pain level using a numerical scale from 0 (no pain) to 10 (severe pain). The NPIS has good sensitivity, producing data for statistical analysis and having undergone testing in various populations, including chronic pain patients, acute pain patients, older adults, and individuals with rheumatoid arthritis (Williamson & Hoggart, 2005). Experts have observed high test-retest reliability in both literate and illiterate patients with RA (r = 0.96 and 0.95, respectively) before and after medical consultation (Ferraz et al., 1990). Construct validity highly correlates with the Visual Analog Scale (a measurement instrument for individual characteristics of pain) in patients with rheumatic and other chronic pain conditions (pain > 6 months) before and after medical consultation, con rming it as an effective strategy for measuring pain (Ferraz et al., 1990). Responses to the NPIS represented the ordinal variable pain, an aspect of symptomatology.
Fatigue Severity Scale. The FSS is a 9-item, self-report summary scale designed to assess disabling fatigue in all individuals, with common use in patients with MS and systemic lupus erythematosus . For the present study, the total score variable served as a continuous score and was computed from an average of the nine survey questions. This score represented the variable fatigue, a subcomponent of symptomatology.
Multiple Sclerosis Neuropsychological Screening Questionnaire. The MSNQ is a 15item self-report inventory that measures cognitive function (Benedict & Cox, 2004). Patients rate themselves from 0 (ne) to 4 (v) regarding 15 speci c cognitive and behavioral problems that may arise in daily life, such as "Are you easily distracted?" The higher the score, the greater amount and severity of cognitive problems. Cronbach's alpha coe cients were 0.93 and 0.94 upon analyses for the reliability of the MSNQ and correlations between both patient-and informant-report forms, respectively (Benedict et al., 2003). The MSNQ is a reliable self-administered screening test yielding a sensitivity of 0.83 and a speci city of 0.97 (Benedict et al., 2003). The researcher calculates a total score for each participant by taking the average of responses for all 15 questions; this total score corresponded to the continuous variable cognitive de cits (a subcomponent of symptomatology). Multiple Sclerosis Self-Management Scale-Revised. The MSSM-R addresses the multidimensional nature of self-management for individuals with MS, including the self-management processes conceptualized in the self-and family management framework. Such processes comprise factors focusing on illness, the activation of resources, and living with the condition. This instrument asks participants to rate 24 statements about their ability to self-manage, like "I have a good understanding of why I take my medications and what they are supposed to do" and "I try to take a break when I feel myself getting tired" on a scale from 1 (I ) to 5 (I). Psychometrics are well-established. The test-retest reliability, using the intraclass correlation coe cient (ICC) with 95% con dence interval (CI), revealed a coe cient of 0.83 between two scores over time; the overall ICC ranged from 0.64 to 0.88, suggesting satisfactory test-retest reliability (Ghahari, Khoshbin, & Forwell, 2014). Criterion validity through comparison of the MSSM-R to two generic self-management scales (Health Education Impact Questionnaire and Partners in Health) indicated moderate to high criterion validity within the subscales of the MSSM-R, p = .01 (Ghahari et al., 2014). For the present study, each participant had a calculated total score that the researcher derived by taking an average of responses to each question. This total score represented the continuous variable self-management.
Multiple Sclerosis International Quality of Life. The MusiQoL is a multidimensional scale, selfadministered, 31-item questionnaire available in 14 languages, serving as a disease-speci c QoL scale appropriate for international application. The MusiQoL features 31 questions in nine dimensions (subscales): activities of daily living, psychological well-being, symptoms, relationships with friends, relationships with family, relationships with healthcare system, sentimental and sexual life, coping, and rejection. The global index score is the mean of the nine subscales, linearly transformed and standardized on a scale 0-100 scale; where 0 indicates the worst possible level of QoL and 100 indicates the best level.

Processes and Interventions
The researcher entered the collected data into SPSS version 22.0 for Windows. Administered through Qualtrics, the survey required participants to provide a response to every item. The researcher also screened the data for outliers in order to ensure that extreme values were not skewing the data distribution. The researcher identi ed outliers as values outside of the range z = + 3.29 standard deviations away from the mean (Tabachnick & Fidell, 2012) and removed these values from further analysis.

Statistical Analyses
To determine whether comorbidity, condition severity, and symptomatology (pain, fatigue, and cognitive de cits) are inversely predictive of quality of life in adults with multiple sclerosis, the researcher conducted a multiple linear regression analysis. To determine whether comorbidity, condition severity, and symptomatology (pain, fatigue, and cognitive de cits) are signi cantly associated with self-management in adults with multiple sclerosis, the researcher conducted another multiple linear regression analysis.
To assess whether self-management predictive of quality of life in adults with multiple sclerosis, the researcher conducted a linear regression with self-management as the independent predictor variable and QoL as the dependent outcome variable. The researcher performed a nal regression analysis to determine whether self-management acts as a mediator between the independent variables comorbidity, condition severity, pain, fatigue, and cognitive de cits and the dependent variable QoL. In this analysis, the researcher regressed QoL on self-management and the independent variables of comorbidity, condition severity, pain, fatigue, and cognitive de cits.

Descriptive Statistics
Frequencies and percentages. A majority of the participants were female (n = 165, 84.2%). The age of participants heavily varied between 36 to 40 years (n = 40, 20.4%) and 41 to 45 years (n = 48, 24.5%). The ethnicity of the participants was predominantly White (n = 159, 81.1%). Most of the participants were married (n = 101, 51.5%). The highest level of education varied between primary/vocational education (n = 49, 25.0%), secondary education (n = 53, 27.0%), and higher education (n = 94, 48.0%). Many of the participants were employed (n = 85, 43.4%). Roughly one quarter of the sample could not work or had physical challenges (n = 53, 27.0%). Table 1 shows the frequencies and percentages of the demographics. Table 1 Frequency    Table 3 presents the ndings of the descriptive statistics for the variables of interest. Normality. The researcher examined the skewness and kurtosis values to test the normality assumption. The researcher used the following criteria to assess normality: skew between -2.0 and 2.0 and kurtosis between -7.0 and 7.0 (Kline, 2010). The values for skew and kurtosis fell within the acceptable range. Further assessment of the normality assumption occurred through use of the Kolmogorov-Smirnov test and the Shapiro-Wilk test. The researcher concluded that the ndings of the tests were not statistically signi cant for self-management and QoL; therefore, these variables met the assumption. The ndings of the Kolmogorov-Smirnov test and Shapiro-Wilk test, however, were not signi cant for the remaining variables (p < .05). Comorbidity, condition severity, pain intensity, fatigue, and cognitive de cits did not meet the assumption. Stevens (2009) suggested that samples with sums of 50 or more observations often approximate to normality. With a total sample of 196 participants, the current researcher assumed that the distributions approximated toward normality. Table 4 presents the ndings of the Kolmogorov-Smirnov and Shapiro-Wilk tests.

Bivariate Analysis
The researcher performed bivariate analyses to identify which predictor variables have a statistically signi cant relationship with the dependent variable (see Table 5). The ndings revealed that comorbidity, condition severity, symptomatology (pain, fatigue, and cognitive de cits), and self-management had signi cant relationships with QoL. Comorbidity, condition severity, and symptomatology were inversely associated with QoL, while self-management was positively associated with QoL.

Facilitators/Barriers and Quality of Life
The researcher used a multiple linear regression to examine the predictive relationship between comorbidity, condition severity, and symptomatology (pain, fatigue, and cognitive de cits) on QoL. The predictor variables corresponded to comorbidity, condition severity, and symptomatology (pain, fatigue, and cognitive de cits). The continuous criterion variable corresponded to QoL.
Prior to analysis, the researcher assessed the assumptions of normality, homoscedasticity, and absence of multicollinearity. The results met the assumption of normality for the residuals due to the data closely following the normality trend line (see Figure 2). The assumption of homoscedasticity underwent visual assessment through a residuals plot. The results met the assumption due to a nonrecurring pattern appearing in the plot (see Figure 3). The results met the assumption of multicollinearity due to the variance in ation factors (VIF) being below 10 (see Table 6).

Facilitators/Barriers and Self-Management
The researcher conducted a multiple linear regression to examine the predictive relationship of comorbidity, condition severity, and symptomatology (pain, fatigue, and cognitive de cits) on selfmanagement. A criterion in this regression model was the mediating variable.
Prior to analysis, the researcher assessed assumptions of normality for the residuals, homoscedasticity, and absence of multicollinearity. The assessment met assumption of normality for the residuals due to the data closely following the normality trend line (see Figure 4). This visual assessment of assumption of homoscedasticity featured a residuals plot. The assessment met the assumption due to a nonrecurring pattern appearing in the plot (see Figure 5). The assessment met assumption of multicollinearity due to the VIF being below 10 (see Table 7).
The results of the overall model of the multiple linear regression were statistically signi cant, (F[5, 190] = 7.99, p < .001, R 2 = .174). This suggests that a signi cant relationship exists between comorbidity, condition severity, and symptomatology (pain, fatigue, and cognitive de cits) and self-management. The R 2 value suggests that approximately 17.4% of the variance in self-management can be explained by the predictors.
Comorbidity (t = -3.23, p < .001) was a signi cant predictor in the model, suggesting that with every oneunit increase in comorbidity scores, self-management scores decreased by approximately 0.09 units. Cognitive de cits (t = -4.28, p < .001) was a signi cant predictor in the model, suggesting that with every one-unit increase in cognitive de cits scores, self-management scores decreased by approximately 0.17 units. Table 7 depicts the results of the multiple linear regression.

Self-Management and Quality of Life
The researcher conducted a linear regression to examine the predictive relationship of self-management on QoL. Prior to analysis, the researcher assessed the assumptions of normality for the residuals and homoscedasticity. The assessment met the assumption of normality for the residuals due to the data closely following the normality trend line (see Figure 6). The assessment met the assumption of homoscedasticity with a visual assessment through a residuals plot. The assessment met assumption due to a nonrecurring pattern appearing in the plot (see Figure 7).
The results of the overall model of the linear regression indicated a lack of statistical signi cance (F [1,194] = 58.56, p < .001, R 2 = .232), suggesting that a signi cant relationship between self-management and QoL exists. The R 2 value suggests that approximately 23.2% of the variance in QoL can be explained by self-management. Self-management (t = 7.65, p < .001) was a signi cant predictor in the model, suggesting that with every one-unit increase in self-management, QoL scores increased by approximately 18.10 units. Table 8 presents the results of the linear regression.

Mediating Effects of Self-Management
The researcher conducted a Baron and Kenny mediation analysis to assess whether self-management signi cantly mediates the effects of comorbidity, condition severity, and symptomatology (pain, fatigue, and cognitive de cits) on QoL in adults with MS. The researcher conducted three regressions to assess for mediation. For mediation to be supported, the results were required to meet four conditions: The results did not meet the fourth item of the Baron and Kenny method. Therefore, the results support partial mediation for self-management on the relationship between comorbidity, condition severity, and cognitive de cits on QoL. The results did not support mediation for self-management on the relationship between pain intensity and fatigue on QoL. Table 9 shows the results of the regressions.

Discussion
The studied facilitators and barrier had statistically signi cant relationship(s) with the dependent variable (QoL); however, the relationship between fatigue, pain, and QoL was not supported. The ndings supported the theoretical proposition that health status facilitators and barriers such as comorbidities, condition severity, and symptomatology (cognitive de cits) affect distal individual outcomes such as QoL (Grey et al., 2015). The results showed that a greater number of comorbidities, a worse condition severity, and a greater number of cognitive de cits correlated with a lower quality of life. The inverse relationship between the predictors and QoL supported the notion of the predictors functioning as barriers.
Moreover, the ndings revealed that a relationship exists between comorbidity, condition severity, and symptomatology and self-management. The ndings supported the theoretical proposition that health status facilitators and barriers, such as comorbidities and symptomatology (cognitive de cits), affect The results indicated that self-management did not mediate the relationship between the comorbidity, condition severity, and symptomatology (pain, fatigue, and cognitive de cits) and QoL in adults with MS.
To date, no researchers have investigated the mediation effect of self-management on the relationship between comorbidity, condition severity, and symptomatology (pain, fatigue, and cognitive de cits) and QoL for adults with MS. Researchers studying individuals with other chronic diseases such as type 2 diabetes, however, have demonstrated the mediating effects of self-management on health outcomes, both proximal and distal (i.e., self-management behaviors, cost of health services, health status, and The conclusions that the researcher has drawn from this study must be considered in terms of limitations related to the adequacy of the indicators of the constructs and the population under study. Although the researcher took efforts to address potential limitations with the use of an online survey related to sampling frames, response rates, participation deception, and access to the desired sample, data collection occurred by way of self-report, which could have led to potential bias in the results. Although the researcher sought a nationally representative sample, the participants were homogeneous, with the majority being White and nearly half employed with higher education. The timing of patient-reported outcomes, although helpful in gaining insight, also has limitations. There remains no consensus on how to de ne severe MS, but it is well established from observational studies that a small minority of patients with MS follow a rapidly disabling disease course (Charleson, Herbert, & Kister, 2016). The assignment of condition severity in the setting of relapses may potentially have affected study participants' responses. No gathered data were speci c to whether participants felt they were currently having relapse symptoms. There is instability during relapse (Charleson et al., 2016).
In view of the instability during and following relapse, the researcher recommends that in the future, severity grade determination ideally should not be made within 6 months of reported relapse and within the rst year of diagnosis. Unfortunately, the researcher did not obtain or analyze data on the timing of the last relapse.
The design of this study did not allow for test-retest reliability or comparison of the FSS, NPIS, and PDDS to other comparable scales or between MS patients and health controls. This cross-sectional study lacked any basis for causality since all variables were assessed simultaneously. There are three requirements for causality: correlation, temporal precedence, and removal of confounding variables. This cross-sectional study indicated correlation, but not temporal precedence. Control of key plausible confounding variables may be considered; however, in order to better understand key in uential variables that affect the process of SM and QoL, a longitudinal design may provide bene ts, especially given its ability to assess change over time.

Conclusions
Although the current researcher failed to identify any mediating effects of self-management between predictor variables (facilitators and barriers) and distal outcomes (QoL), the ndings provided valuable insight. Self-management was positively associated with QoL and, as the researcher hypothesized, predictor variables (comorbidity, condition severity, and symptomatology) in uence the process of selfmanagement in adults with MS. There remains a need for ongoing biopsychosocial research incorporating physiological, psychological, and psychosocial factors in understanding the mediating effects of self-management between facilitators and barriers and QoL. While a cure remains out of reach, individuals with MS should be capable of managing the day-to-day effects of the disease on their lives.
Understanding the factors that facilitate or hinder MS self-management among adults with MS is essential to improving interventions and health outcomes. The expansion and understanding of self-management research in MS are critical to the development of -and ultimately, the translation of-evidence-based interventions to consumers in a sustainable manner.
Through this study, the researcher provided additional evidence in support of the growing body of literature regarding self-management in adults with MS, including several opportunities to improve health, advance the delivery of healthcare services, and in uence policy.

Declarations
Ethics Approval and Consent to Participate: Permission was obtained to conduct the study from The Catholic University Committee for the Protection of Human Subjects. The respondents were required to indicate their informed consent to participate before they were able to complete the survey. The risk to the participants was minimal.
Consent for Publication: The participants were aware that their data would be published in aggregate and without identi ers; any further permissions were not required.
Availability of Data and Materials: The datasets generated and analyzed in the current study are available from the corresponding author upon reasonable request.