Stress management interventions for multiple sclerosis: A meta-analysis of randomized controlled trials

Abstract

Stress management interventions can help to improve mental health for adults living with multiple sclerosis. However, uncontrolled study designs may overestimate intervention effects. A systematic search of the Embase, PsycINFO, PubMed, and Scopus databases identified eight randomized controlled trials evaluating cognitive behavioral approaches for a pooled sample of 568 adults with multiple sclerosis. Both group and individual-based stress management interventions appear to be effective in promoting self-management of stress. Further research is needed to confirm the optimal timing of stress management interventions across the MS spectrum and strategies to maintain intervention effects.

Keywords

adults anxiety cognitive behavior therapy depression mindfulness

Multiple sclerosis (MS) is a chronic and often debilitating neurological disease. The unpredictability of the MS disease course in addition to variation in symptom type and severity can significantly disrupt daily functioning, contributing to tension, strain, and stress (Benito-León, 2011; Ziemssen, 2011). Moreover, there is evidence that the pathogenesis and progression of MS can be exacerbated by chronic, unremitting stress (Briones-Buixassa et al., 2015). It follows that evidence-based psychological techniques to support the reduction of stress responses offer a promising avenue for MS symptom management.

However, there remains a lack of consensus regarding the specific effects of stress management for persons with MS. Indeed, the MS literature is characterized by inconsistencies in the operationalization of “stress,” defined as the frequency of negative life events, perceived stress, emotional distress, and even physiological stress (e.g. Bogosian et al., 2016; Mohr et al., 2012). Stress prevention protocols targeted to individuals also vary in their treatment components—some of which may incorporate “elements of stress coping training” (e.g. Tesar et al., 2003: 394), in addition to their delivery methods—including the use of individual and group programs.

A useful conceptual framework for evaluating and synthesizing the stress management and MS literature is the Transactional Model of Stress. Developed by Lazarus and Folkman (1984), this model emphasizes the transaction between a person (including cognitive, physiological, affective, psychological, and neurological systems) and their environment. Lazarus and Folkman (1984) also implied that stress responses are influenced by a person’s appraisal or cognitive response to a stressor more so than the stressful life event itself. Indeed, stress management techniques which target identity change and illusory thinking styles in persons with MS have demonstrated effectiveness. For example, mindfulness practice alongside methods to induce relaxation has been shown to enhance self-awareness and regulation of negative emotions (e.g. Artemiadis et al., 2012; Blankespoor et al., 2017). Even basic information and education on methods for coping with stress can positively impact stress appraisal (McGuire et al., 2015).

In summary, psychological-based stress management interventions targeted to MS have been developed and evaluated, however, these data are characterized by variability in the operationalization of stress in addition to quasi-experimental study designs. Importantly, randomized controlled studies specifically focused on reducing stress in the MS cohort have been published in recent years and can extend on the findings of previous reviews (Munoz et al., 2016; Reynard et al., 2014; Simpson et al., 2014; Thomas et al., 2006). The current paper quantitatively evaluates the evidence-base for stress management in MS to determine whether: (1) stress management interventions effectively reduce perceived stress, and improve mental health, in persons with MS compared to standard care or no treatment, and (2) noted intervention effects are maintained over time. The impact of intervention format (individual vs group-based) as a potential moderator of intervention effectiveness was additionally considered.

Method

Literature search

A systematic search of the Embase, PsycINFO, and PubMed databases (from database inception to January 2019) was conducted to identify articles that examined the effectiveness of stress management interventions for an adult MS population (see Table S1 supplemental material for search strategy). Citation searching of eligible studies via Scopus identified one additional study (Sanaeinasab et al., 2017). Finally, a manual search of the reference lists of included studies and previous reviews on this topic (Malcomson et al., 2007; Munoz et al., 2016; Pagnini et al., 2014; Reynard et al., 2014; Sesel et al., 2018; Simpson et al., 2014) was conducted to identify eligible articles that may have been missed in the initial search.

Selection criteria

As specified in our pre-registered protocol (PROSPERO CRD42019119589), eligible studies needed to examine an adult population (i.e. ⩾18 years) with a self-reported or clinically confirmed diagnosis of MS (McDonald et al., 2001; Polman et al., 2011, 2005). In accordance with Lazarus and Folkman’s (1984) conceptualization of stress, cognitive behavioral interventions designed to alter stress responses (i.e. physiological responses, dysfunctional thinking, maladaptive life-style patterns) were considered. This included individual and group-based techniques. The intervention had to include “stress management” or a synonymous term (i.e. “stress reduction,” “stress inoculation,” “transactional model of stress”; Linden, 2005) as a descriptor. In addition, perceived stress had to be assessed using a standardized self-report measure, administered at baseline and post-intervention. Secondary outcomes included depression and anxiety severity, commonly used as proxies for measuring psychological stress (Suzuki and Ito, 2013). The stress management intervention had to be administered by a health professional or specialist trainer and involve some face-to-face component. In addition, only studies with an independent groups design were eligible, whereby persons with MS were randomly assigned to either an experimental group who received the stress management intervention under evaluation, a comparison group (controls) which received a conventional (active) treatment (e.g. routine medical or psychosocial care), or no treatment (e.g. wait-list control). Uncontrolled designs were ineligible as these designs are known to introduce validity concerns and raise uncertainty about intervention effects (Maida et al., 2014; Peinemann et al., 2014; Schweizer et al., 2016). Finally, studies had to be published in the English language, or with English translation (Jüni et al., 2002), and provide sufficient data to allow the calculation of group mean differences in the form of Hedges’ g (e.g. means, standard deviations).

The initial search identified 1958 potential studies which were screened by the second and third author using Covidence software (Veritas Health Innovation, Melbourne, Australia, www.covidence.org), with unanimous agreement. A total of 1474 articles remained after removal of duplicates and unrelated topics. Re-applying the inclusion and exclusion criteria to the titles and abstracts reduced this number to 149. The full-text of these articles was subsequently examined against the eligibility criteria. Authors of seven articles were emailed for additional detail, with six responding. This resulted in a final sample of eight independent studies (Figure S1 supplemental material).

Data collection and preparation

Consistent with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement for health care interventions (Hutton et al., 2015; Liberati et al., 2009), the following information was extracted from each study: sample description (e.g. age, gender, MS subtype), intervention characteristics (e.g. therapy framework, therapy format), and outcome data (e.g. means, standard deviations for each repeated measure, for both intervention and control groups, extracted at baseline, immediately post-intervention, or at follow-up). For ease of data interpretation, outcomes were classified according to the construct they represented: stress, depression, or anxiety. Data extraction was performed by the first author and checked by the second and third authors.

Risk of bias assessment

The Cochrane Collaboration’s Risk of Bias tool (revised version 5.1.0, Higgins et al., 2016) was utilized to assess each study on five methodological biases (“domains”) common to randomized controlled trials (RCTs): insufficient random sequence or inadequate concealment of group allocation (selection bias), deviations in the delivery of the intended intervention—which can be minimized via therapist training and supervision (performance bias), missing outcome data (attrition bias), measurement validity and the potential for detection bias, and selective reporting of study results (reporting bias). Bias is assessed as “high,” “low,” or “unclear” (Higgins et al., 2016). The second and third authors performed this assessment with moderate agreement (kappa = .54). Disagreements were resolved by consensus-based discussion with all authors.

Statistical analyses

Data were entered into Comprehensive Meta-Analysis software (Version 3.9; Biostat Inc, Englewood, NJ). Hedges’ g, which measures effect size in standard deviation units and has better small sample properties than Cohen’s d (Ellis, 2010), was the primary estimate in this review. The larger the g value the greater the difference between the intervention and control groups in terms of outcome. For example, a g value of 0 indicates that on average, participants in each group performed the same, whereas a g value of 1.0 indicates that they differed by one standard deviation (Ellis, 2010). Effect sizes were calculated for each psychological measure of stress/distress reported by a study. Both short-term (baseline to immediately post-intervention) and longer-term intervention effects (baseline to follow-up) were examined. Interpretation of effect size results was based on Cohen’s (1988) guidelines: .2, .5, and ⩾.8 representing small, medium, and large to very large estimates, respectively.

When computing g from studies with a two-group repeated measures design, the pre-post correlation is required. As studies did not routinely report this correlation, a conservative estimate of r = .7 was used based on existing within-group test–retest data for measures of stress, depression, and anxiety in MS research (r: .65–.98; Briones-Buixassa et al., 2015; Hind et al., 2016; Marrie et al., 2018; Watson et al., 2014).

Individual g’s for each outcome measure utilized by a study were categorized according to the construct they represented (stress, depression, anxiety) and pooled (g_w), with weighting based on each study’s inverse variance (Lipsey and Wilson, 2001). If a study contributed more than one outcome measure per construct, an average g was calculated prior to pooling to ensure data independence (Lipsey and Wilson, 2001). Ninety-five percent confidence intervals (CIs) were additionally calculated to quantify the uncertainty of individual and pooled g’s while p values (<.05) were used to confirm the statistical significance of g (Cumming, 2012).

Between-study heterogeneity was estimated with three indices. The Q statistic, which analyses the ratio of observed variation to within-study error (significant Q suggests that true effect sizes vary); T, analogous to a SD for the overall g_w effect; and the I² statistic - a proportional estimate of true effect variance over sampling error observed (Borenstein et al., 2017). A random effects model was used for these analyses to allow for variation in the “true” effect size between each study due to sampling and methodological differences as well as random error (Cumming, 2012).

Potential evidence of publication bias was examined by calculating Orwin’s (1983) Fail-safe N (N_fs) for each effect estimate. This statistic determines how many hypothetical publications would be required to reverse an overall g_w to a statistically unimportant effect size (i.e. g = .2). A fail-safe N value was considered adequate if it exceeded the overall number of studies included in this review (N_fs > N_studies).

Sensitivity and subgroup analyses

The data extraction process highlighted potential outliers, or influential effect estimates, among the included intervention trials. A one-study removed sensitivity analysis was subsequently conducted for each symptom construct to determine the robustness of the observed findings. Results were considered meaningful if there was a change in the effect size magnitude (Cohen, 1988) or if the statistical significance of an effect estimate changed (Borenstein et al., 2009).

There was sufficient power to undertake a subgroup analysis of intervention format (i.e. individual vs group) and its impact on stress outcomes in the short-term. This analysis used a mixed-effect model, consisting of a random effects model within subgroups and a fixed effect model across subgroups (Borenstein et al., 2009).

Results

Study characteristics

As seen in Table 1, all studies were published within the last 30 years. Four standardized self-report measures of chronic stress were utilized to assess intervention effectiveness, typically the Depression Anxiety and Stress Scales and Perceived Stress Scale. Depressed mood and state anxiety were screened with tools that have been validated in MS research (e.g. Fischer et al., 1999; Hind et al., 2016; Senders et al., 2014; Watson et al., 2014).

Table 1.

Characteristics of included studies.

Lead author (date)	Sample						Stress management intervention						Control or comparison
	N			Country	Recruitment source	Measure(s) of stress/distress	Attrition^a	Framework	Therapist	Format	No. sessions	No. weeks
	Total	SM	C	Country	Recruitment source	Measure(s) of stress/distress	Attrition^a	Framework	Therapist	Format	No. sessions	No. weeks
Agland (2018)	100	50	50	Australia	MS clinic	DASS-21 VAS	–	MiCBT	Nurse and Psychologist	Individual	4	4	Wait-list
Artemiadis (2012)	61	31	30	Greece	MS clinic	PSS BDI STAI	SM = 16.2% C = 16.7%	PMR	Psychologist	Individual	8	8	Education
Foley (1987)	41	–	–	United States	MS clinic	Hassles Scale BDI STAI	Total = 12.1%	CBT	Psychologist	Individual	6	6–8	Usual care
Kolahkaj (2015)	48	24	24	Iran	MS society	DASS-21	SM = 16.7% C = 16.7%	MBSR	Psychologist	Group	8	8	Usual care
Mohr (2012)	121	60	61	United States	MS clinics MS society	BIPS	SM = 21.6% C = 9.8%	CBT	Psychologists and Social Worker	Individual	16	20–24	Wait-list
Sanaeinasab (2017)	80	40	40	Iran	MS society	PSS	SM = 0% C = 0%	Psychoeducation	Health Educator	Group	6	6	Usual care
Senders (2018)	67	33	34	United States	MS society MS clinics	PSS PROMIS	SM = 0% C = 14.7%	MBSR	Health Educator	Group	8	8	Education
Simpson (2017)	50	25	25	Scotland	Physicians	PSS MHI	SM = 16% C = 16%	MBSR	General Practitioners	Group	8	8	Wait-list

N: number of participants at baseline; SM: stress management intervention; C: control or comparison group; DASS: Depression Anxiety Stress Scales (42- and 21-item); VAS: Visual Analogue Scale; PSS: Perceived Stress Scale; BDI: Beck Depression Inventory; STAI: Spielberger State-Trait Anxiety Inventory; BIPS: Brief Inventory of Perceived Stress; PROMIS: Patient-Reported Outcomes Measurement Information System; MHI: Mental Health Inventory; MiCBT: mindfulness-integrated cognitive behavioral therapy; PMR: progressive muscle relaxation; CBT: cognitive behavior therapy; MBSR: mindfulness-based stress reduction.

The symbol “–” denotes data/detail not reported.

Attrition based on number of participants that completed the intervention.

Participant characteristics

The pooled sample primarily included females living with chronic relapsing–remitting MS (Table S2 supplemental material). Four studies provided critical information related to participants’ pharmacological care (i.e. interferon beta medications to reduce MS disease activity) prior to study recruitment. Studies commonly excluded persons with high levels of disability (Kurtzke Expanded Disability Status Scale (EDSS) > 8.0; Kurtzke, 1983), with most participants able to walk without an aid (EDSS mean range: 2.4–6). Individuals with a prior history of psychotherapy, current psychiatric disorder (e.g. major depression, anxiety disorder), or cognitive impairment were typically excluded.

Risk of bias assessment

Two studies were identified with low risk of bias (Mohr et al., 2012; Senders et al., 2018; Figure S2 supplemental material). The remainder had some concerns in one or more RoB 2.0 domains. Most (75%) studies minimized selection bias (Domain 1) by identifying their method of randomization (e.g. permuted block), although detail relating to allocation concealment was not always provided and concealment of intervention allocation not feasible, as is commonplace in psychotherapy research (Hróbjartsson et al., 2014). There was risk of performance bias (Domain 2), with only three studies (37.5%) explicitly evaluating intervention delivery in addition to providing therapists with supervised practice. Attrition bias was low, with outcome data available for most (i.e. ⩾85%) randomized participants (Domain 3). There were some concerns about detection bias due to the reliance on self-assessed symptoms (Domain 4). Finally, four studies (50%) minimized reporting bias (Domain 5) by publishing a pre-specified statistical plan or documented clinical trial registry use.

Intervention characteristics

Eight individual interventions sought to improve stress management in persons with MS. Cognitive components commonly included: challenging and confronting negative or pessimistic thoughts, redefining life goals post MS diagnosis, and mindfulness acceptance (e.g. mindfulness-based stress reduction; MBSR) to train self-awareness and maximize coping ability. Behavioral techniques to relieve stress included progressive muscle relaxation and controlled breathing.

Stress management interventions were time-limited, averaging eight sessions (range: 4–16) delivered over 9 weeks (range: 3–24 weeks) with a mean session duration of 90 minutes (range: 50 minutes to 2 hours; Table 1). Four studies favored group therapy, led by one or two experienced facilitators. In-session learning was supplemented with daily practice (range: 20–50 minutes) to reinforce relaxation skills and techniques taught. Artemiadis et al. (2012) minimized therapist time by offering an initial face-to-face session followed by weekly telephone communication to encourage compliance. Programs were also shortened to enhance intervention feasibility and effectiveness: Foley et al. (1987) targeted progressive muscle relaxation exercises to functional muscle groups and Simpson et al. (2017) did not include a day retreat in their MBSR program.

Intervention adherence was monitored by five studies; all of which adopted standardized interventions (e.g. MBSR, cognitive behavior therapy; CBT). The majority (76%) of participants were classified as “completers” by attending at least 50% of classes. However, adherence to home practice of relaxation techniques (including activity logs, daily symptom diary) varied (range: 50%–75%). Therapist competence and fidelity to manualized treatment was explicitly monitored by Simpson et al. (2017), using guidelines for health behavior change research developed by the National Institutes of Health (Bellg et al., 2004). Similarly, Mohr et al. (2012) and Foley et al. (1987) audio recorded therapist sessions which were subsequently reviewed in weekly supervision.

Control conditions typically included a wait-list with participants offered the stress management intervention on study completion, or usual medical care (i.e. routine pharmacological treatment and symptom monitoring, N_studies = 3). Two studies provided an attentional control condition involving written information and education about MS care (Artemiadis et al., 2012; Senders et al., 2018).

Stress intervention effects

Short-term effects

Table 2 lists the effects reported by individual studies, grouped by construct, measure, intervention framework, and format. Of the eight stress management interventions evaluated, six identified significant and immediate reductions in levels of perceived stress based on five self-report measures. The pooled effect estimate for this domain was robust. However, significant between-study variation in effect estimates was noted. This included medium to very large improvements following MBSR (Kolahkaj and Zargar, 2015; Simpson et al., 2017), an educational program based on the transactional model (Sanaeinasab et al., 2017), progressive muscle relaxation training (Artemiadis et al., 2012), and CBT (Foley et al., 1987; Mohr et al., 2012). In comparison, Agland et al.’s (2018) multimodal intervention (i.e. combination strategies; mindfulness, meditation, CBT) program had no significant effect on stress outcomes, while Senders et al.’s (2018) pilot MBSR trial produced non-significant improvements. These individual findings were, however, susceptible to publication bias (N_fs ⩽ 1).

Table 2.

Short-term effects associated with stress management interventions.

Construct	Measure	N_studies	N_participants	Framework	Format	g_w	95% CI		p	N_fs	Heterogeneity				Lead author (date)
Construct	Measure	N_studies	N_participants	Framework	Format	g_w	L	U	p	N_fs	Q	p	I ²	T	Lead author (date)
Stress	DASS	1	48	MBSR	Group	2.36	.16	3.09	<.01	11					Kolahkaj (2015)
		1	100	MiCBT	Individual	.03	–.36	.41	.90	1					Agland (2018)
	PSS	1	80	Education	Group	1.51^a	1.02	2.01	<.01	7					Sanaeinasab (2017)
		2	109	MBSR	Group	.65	–.12	1.43	.10	5					Senders (2018) Simpson (2017)
		1	61	PMR	Individual	.61^a	.10	1.12	.02	2					Artemiadis (2012)
	Hassles	1	36	CBT	Individual	.78^a	.12	1.45	.02	3					Foley (1987)
	BIPS	1	121	CBT	Individual	.44^a	.08	.80	.02	1					Mohr (2012)
	VAS	1	100	MiCBT	Individual	.05	–.34	.44	.79	1					Agland (2018)
	Total g_w	8	544			.83^a	.38	1.29	<.01	25	58.38	<.01	88.01	.59
Depression	DASS	1	40	MBSR	Group	2.52^a	1.70	3.34	<.01	12					Kolahkaj (2015)
	BDI	1	36	CBT	Individual	1.40^a	.69	2.12	<.01	6					Foley (1987)
		1	61	PMR	Individual	.84^a	.32	1.35	<.01	3					Artemiadis (2012)
	MHI	1	50	MBSR	Group	.58^a	.02	1.14	.04	2					Simpson (2017)
	PROMIS	1	59	MBSR	Group	.35	–.16	.86	.18	1					Senders (2018)
	Total g_w	5	210			1.08^a	.44	1.73	.01	22	22.72	<.01	82.39	.66
Anxiety	DASS	1	40	MBSR	Group	2.86^a	1.98	3.72	<.01	13					Kolahkaj (2015)
	STAI	1	36	CBT	Individual	.83^a	.17	1.50	.01	3					Foley (1987)
		1	61	PMR	Individual	.47	–.03	.98	.07	1					Artemiadis (2012)
	MHI	1	50	MBSR	Group	.74^a	.18	1.31	.01	3					Simpson (2017)
	PROMIS	1	59	MBSR	Group	.40	–.11	.91	.12	1					Senders (2018)
	Total g_w	5	246			.99^a	.32	1.67	<.01	20	25.37	<.01	84.23	.70

N_studies: number of studies providing these data; N_participants: number of participants; g_w: weighted Hedges’ g 42; CI: confidence interval; DASS: Depression Anxiety Stress Scales (42- and 21- item); VAS: Visual Analogue Scale; PSS: Perceived Stress Scale; Hassles: Hassles Scale; BIPS: Brief Inventory of Perceived Stress; BDI: Beck Depression Inventory; MHI: Mental Health Inventory; PROMIS: Patient-Reported Outcomes Measurement Information System; STAI: Spielberger State-Trait Anxiety Inventory; MBSR: mindfulness-based stress reduction; MiCBT: mindfulness-integrated cognitive behavioral therapy; PMR: progressive muscle relaxation; CBT: cognitive behavior therapy.

Significant finding: CI ≠ 0, p < .05.

Five of the aforementioned programs examined depressed mood and state anxiety as secondary outcomes. Four studies reported significant improvements in one or both of these symptom domains following MBSR (Kolahkaj and Zargar, 2015; Simpson et al., 2017), CBT (Foley et al., 1987), or relaxation training (Artemiadis et al., 2012).

Longer-term effects

Five studies provided sufficient data to evaluate the maintenance effects of stress management (Table 3). An overall trend toward improvement on stress, depression, and anxiety measures was reported by four studies, with statistically significant and very large effects associated with two MBSR programs at 1-month (Kolahkaj and Zargar, 2015) and 3-month follow-up (Sanaeinasab et al., 2017).

Table 3.

Longer-term effects associated with stress management interventions.

Construct	Measure	N_studies	N_participants	Time	Framework	Format	g_w	95% CI		p	N_fs	Heterogeneity				Lead author (date)
Construct	Measure	N_studies	N_participants	Time	Framework	Format	g_w	L	U	p	N_fs	Q	p	I ²	T	Lead author (date)
Stress	DASS	1	48	1	MBSR	Group	2.32^a	1.53	3.11	<.01	11					Kolahkaj (2015)
		1	100	6	MiCBT	Individual	.03	–.36	.41	.90	1					Agland (2018)
	PSS	1	77	3	Education	Group	2.11^a	1.56	2.66	<.01	10					Sanaeinasab (2017)
		1	56	3	MBSR	Group	.31	–.24	.86	.27	1					Simpson (2017)
		1	100	12	MBSR	Group	.33	–.19	.85	.22	1					Senders (2018)
	Total g_w	5	323				.99^a	.09	1.89	.03	20	56.42	<.01	92.91	.99
Depression	DASS	1	40	1	MBSR	Group	2.76^a	1.91	3.62	<.01	13					Kolahkaj (2015)
	MHI	1	50	3	MBSR	Group	.05	–.50	.59	.87	1					Simpson (2017)
	PROMIS	1	56	12	MBSR	Group	.15	–.36	.67	.56	0					Senders (2018)
	Total g_w	3	246				.95	–.47	2.34	.19	11	31.10	<.01	93.57	1.20
Anxiety	DASS	1	40	1	MBSR	Group	2.93^a	2.04	3.81	<.01	14					Kolahkaj (2015)
	MHI	1	50	3	MBSR	Group	.56	.00	1.11	.05	2					Simpson (2017)
	PROMIS	1	56	12	MBSR	Group	.03	–.49	.54	.92	1					Senders (2018)
	Total g_w	3	246				1.13	–.34	2.59	.13	14	31.15	<.01	93.58	1.22

N_studies: number of studies providing these data; N_participants: number of participants; g_w: weighted Hedges’ g; CI: confidence interval; DASS: Depression Anxiety Stress Scales (42- and 21-item); PSS: Perceived Stress Scale; MHI: Mental Health Inventory; PROMIS: Patient-Reported Outcomes Measurement Information System; time: post-intervention assessment interval (in months); MBSR: mindfulness-based stress reduction; MiCBT: mindfulness-integrated cognitive behavioral therapy.

Significant finding: CI ≠ 0, p < .05.

Sensitivity and subgroup analyses

Excluding Kolahkaj and Zargar’s (2015) study, whose very large effects (g > 2.0) differed substantially from the medium to large short-term changes noted by the remaining studies, did not significantly change stress or depression ratings. However, exclusion of this study reduced the immediate improvements noted for anxiety from a large to medium overall effect (g_w = .58, CI: .30 to .85, p < .01). Exclusion of this same RCT from meta-analysis of the follow-up stress data also altered the results to a non-significant effect (g_w = .68, CI: –.21 to 1.58, p = .13).

Group-based interventions appeared to have a greater short-term impact on stress ratings (g_w = 1.27 (CI: .47, 2.08), p < .01, I² = 87.65; T = .77; N_studies = 4) in comparison to individually delivered programs (g_w = .36 (CI: .07, .66); p = .02, I² = 59.12; T = .22; N_studies = 4) — a difference which was statistically significant (Q_B(1) = 4.31, p = .04). A similar pattern was noted for the longer-term stress data (group g_w = 1.25 (CI: .19, 2.29), p = .02, I² = 92.16; T = 1.0, N_studies = 4 vs. individual g_w = .03 (CI: .36, .41), p = .90; N_studies = 1), although the limited follow-up data precluded further between-group analysis (Fu et al., 2011).

Discussion

The pooled findings of eight RCTs suggest that stress management interventions incorporating elements of CBT, mindfulness and/or psychoeducation, can effectively reduce perceived stress in addition to targeting trans-diagnostic symptoms of depression and anxiety. Equally important in MS management is the provision of information and education to help increase understanding of the disease process and enhance self-confidence in symptom management (Sanaeinasab et al., 2017).

Group-based interventions produced some of the most promising results. The shared experience of peers has been identified as a critical strategy to promote self-care in persons with MS (Fraser et al., 2013; Schwartz, 1999). Group members provide alternative ways of viewing a situation, help to identify and challenge self-defeating beliefs in each other, and encourage the practice of homework tasks (e.g. relaxation, visualization). However, the benefit of group therapy depends on the person seeking support. Individual-based therapy may be better suited to those unable to participate in complex, multi-session programs or those who may be overwhelmed by the experiences of group members.

Regardless of the intervention format, ongoing relational support from a health care professional is recommended to enhance the success of a stress management program (Artemiadis et al., 2012). Even those who feel capable of self-managing their MS may need professional support when their symptoms flare up (Fraser et al., 2013). Active follow-up should also be accommodated into current stress management interventions in order to promote the maintenance of intervention effects over time. Follow-up might include the use of “booster” sessions focusing on single behavioral components (e.g. progressive muscle relaxation, focused breathing; Artemiadis et al., 2012).

However, little is known about the application and effectiveness of stress management techniques across the MS disease spectrum. Indeed, the majority of studies in this review focused on persons with mild to moderate physical impairment. Although this strict sample screening ensured that participants could actively engage in (modified) progressive muscle relaxation exercises (e.g. Agland et al., 2018), those living with severe MS symptoms have reported a strong need for psychological support and education services (McCabe et al., 2015). Future stress management trials might include various MS subtypes in order to differentiate participants who respond to the intervention from non-responders (Greenstein, 2002; Ziemssen and Thomas, 2017).

The benefits of stress management for those with MS could be optimized if offered soon after diagnosis. This focus on early preventive intervention is consistent with the Transactional Model of Stress (Lazarus and Folkman, 1984), which emphasizes the importance of enhancing one’s coping resources and capabilities prior to the onset of clinically significant symptoms. Earlier implementation of interventions is critical given established links between stress duration, severity, and frequency with autonomic nervous system reactivity (i.e. heart rate variability; Briones-Buixassa et al., 2015) in addition to risk of relapse (Mohr, 2007; Mohr et al., 2004). Stress may even increase suscep-tibility to the MS disease process by impairing the immune system and damaging health (Artemiadis et al., 2011). Furthermore, regular, prolonged practice of stress management techniques has been shown to promote brain activation and preserve nerve integrity (Mohr et al., 2012).

Methodological limitations

These findings need to be considered in the context of several limitations. This included our focus on stress as a negative construct. Later developments of the Transactional Model emphasized positive emotions (e.g. happiness, hope, gratitude, compassion) associated with stress (or eustress) which help to determine one’s coping potential (Lazarus, 1991, 1999, 2000). The study of positive psychological functioning is only recent in the MS literature. Nonetheless, there is some evidence that optimism and hope can lower emotional worry and improve mood (Hart et al., 2008; Schiavon et al., 2017). Similarly, our focus on stress as a psychological process limited our examination to interventions which targeted the teaching of coping skills and reduction of stress responses. Interestingly, the association between psychosocial work environments and the health of employees with a chronic disease has been examined in the organizational behavior literature (see Siegrist and Li, 2018), yet has received less attention in the stress management and MS literature.

The self-reported and self-administered assessments of psychological functioning reported in this review may have also biased effect estimates. For example, the Beck Depression Inventory—one of the most commonly utilized depression screening tools for MS—can artificially inflate depression levels due to its focus on neurovegetative symptoms (e.g. fatigue, sleep; Mohr et al., 1997; Skokou et al., 2012). Future trials might consider supplementing self-assessments with objective evaluation of stress-related symptoms (e.g. physiological or neuroimaging markers).

Our reliance on the fail-safe N as an estimate of publication bias also warrants attention. This statistic is highly dependent on an arbitrary threshold set by the reviewer and, as such, can lead to varying estimates of the number of additional studies (Pham et al., 2001). Notably, alternative methods (e.g. Egger’s regression test, funnel plot analysis) are not recommended with fewer than 10 studies (Egger et al., 1997; Sterne et al., 2011). Importantly, we adopted several criteria for an effect size to be significant (i.e. g > .2; CIs, p p > .05) in order to minimize the number of false positive effects—a common problem in small meta-analyses (Sterne et al., 2011).

Finally, statistical analysis of potential critical differences in the examined stress management protocols was limited. This included the degree to which treatment “dose” (i.e. session number, duration) moderates symptom change. Indeed, standardized mindfulness interventions such as MBSR require relatively intensive training in mindfulness meditation (Kabat-Zinn, 1996), as opposed to the brief relaxation exercises utilized by Artemiadis et al. (2012). Insufficient data were also available in relation to group size and composition—factors which may explain the significant improvements noted in group treatment at an aggregate level. It is important that future MS research includes these details in order to determine the optima format in which stress management interventions should be delivered.

Conclusion

The psychological benefits of stress management interventions for persons with MS are promising, producing significant reductions in self-reported stress and distress. These findings warrant further exploration with a larger cohort of persons at different stages of the MS disease process, in order to ascertain which aspects of stress management are most suitable and effective for this patient group under which contexts.

Supplemental Material

JHP_supplementary_updated_2019NEW – Supplemental material for Stress management interventions for multiple sclerosis: A meta-analysis of randomized controlled trials

Supplemental material, JHP_supplementary_updated_2019NEW for Stress management interventions for multiple sclerosis: A meta-analysis of randomized controlled trials by Paul Taylor, Diana S Dorstyn and Elise Prior in Journal of Health Psychology

Footnotes

Acknowledgements

The authors would like to acknowledge Maureen Bell and Vikki Langton, Research Librarians at the University of Adelaide, for assistance with the database search terms. We are also grateful to authors of included studies who kindly responded to requests for additional data.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Diana S Dorstyn

Supplemental material

Supplemental material for this article is available online.

References

*Agland

Lydon

Shaw

, et al. (2018) Can a stress management programme reduce stress and improve quality of life in people diagnosed with multiple sclerosis? Multiple Sclerosis Journal: Experimental, Translational and Clinical 4(4): 1–4.

Artemiadis

Anagnostouli

Alexopoulos

(2011) Stress as a risk factor for multiple sclerosis onset or relapse: A systematic review. Neuroepidemiology 36(2): 109–120.

*Artemiadis

Vervainioti

Alexopoulos

, et al. (2012) Stress management and multiple sclerosis: A randomized controlled trial. Archives of Clinical Neuropsychology 27(4): 406–416.

Bellg

Borrelli

Resnick

, et al. (2004) Enhancing treatment fidelity in health behavior change studies: Best practices and recommendations from the NIH Behavior Change Consortium. Health Psychology 23(5): 443–451.

Benito-León

(2011) Stress and multiple sclerosis: What’s new? Neuroepidemiology 36(2): 121–122.

Blankespoor

Schellekens

MPJ

Vos

, et al. (2017) The effectiveness of mindfulness-based stress reduction on psychological distress and cognitive functioning in patients with Multiple Sclerosis: A pilot study. Mindfulness 8(5): 1251–1258.

Bogosian

Hughes

Norton

, et al. (2016) Potential treatment mechanisms in a mindfulness-based intervention for people with progressive multiple sclerosis. British Journal of Health Psychology 21(4): 859–880.

Borenstein

Hedges

Higgins

, et al. (2009) Introduction to meta-analysis. Chichester, U.K: John Wiley & Sons.

Borenstein

Higgins

JPT

Hedges

, et al. (2017) Basics of meta-analysis: I2 is not an absolute measure of heterogeneity. Research Synthesis Methods 8: 5–18. DOI:10.1002/jrsm.1230.

10.

Briones-Buixassa

Milà

Aragonès

, et al. (2015) Stress and multiple sclerosis: A systematic review considering potential moderating and mediating factors and methods of assessing stress. Health Psychology Open 2(2): 1–16.

11.

Cohen

(1988) Statistical Power Analysis for the Behavioral Sciences (2nd edn). Hillsdale, NJ: Erlbaum.

12.

Cumming

(2012) Understanding the New Statistics: Effect Sizes, Confidence Intervals, and Meta-Analysis. New York: Routledge.

13.

Egger

Smith

Schneider

, et al. (1997) Bias in meta-analysis detected by a simple, graphical test. British Medical Journal 315(7109): 629–634.

14.

Ellis

(2010) The Essential Guide to Effect Sizes: Statistical Power, Meta-Analysis, and the Interpretation of Research Results. Cambridge: Cambridge University Press.

15.

Fischer

LaRocca

Miller

, et al. (1999) Recent developments in the assessment of quality of life in multiple sclerosis (MS). Multiple Sclerosis 5(4): 251–259.

16.

*Foley

Bedell

LaRocca

, et al. (1987) Efficacy of stress-inoculation training in coping with multiple sclerosis. Journal of Consulting and Clinical Psychology 55(6): 919–922.

17.

Fraser

Ehde

Amtmann

, et al. (2013) Self-management for people with multiple sclerosis: Report from the first international consensus conference, november 15, 2010. International journal of MS care 15(2): 99–106. doi:10.7224/1537-2073.2012-044

18.

Gartlehner

Grant

, et al. (2011) Conducting quantitative synthesis when comparing medical interventions: AHRQ and the effective health care program. Journal of Clinical Epidemiology 64: 1187–1197.

19.

Greenstein

(2002) Problems in clinical trial design in multiple sclerosis. International Journal of MS Care 4(3): 125–138.

20.

Hart

Vella

Mohr

(2008) Relationships among depressive symptoms, benefit-finding, optimism, and positive affect in multiple sclerosis patients after psychotherapy for depression. Health Psychology 27: 230–238.

21.

Higgins

JPT

Sterne

JAC

Savović

, et al. (2016) A revised tool for assessing risk of bias in randomized trials. Cochrane Database of Systematic Reviews 10(Suppl. 1): 29–31.

22.

Hind

Kaklamanou

Beever

, et al. (2016) The assessment of depression in people with multiple sclerosis: A systematic review of psychometric validation studies. BMC Psychiatry 16: 278.

23.

Hróbjartsson

Emanuelsson

Skou Thomsen

, et al. (2014) Bias due to lack of patient blinding in clinical trials. A systematic review of trials randomizing patients to blind and nonblind sub-studies. International Journal of Epidemiology 43(4): 1272–1283.

24.

Hutton

Salanti

Caldwell

, et al. (2015) The PRISMA extension statement for reporting of systematic reviews incorporating network meta-analyses of health care interventions: Checklist and explanations. Annals of Internal Medicine 162(11): 777–784.

25.

Jüni

Holenstein

Sterne

, et al. (2002) Direction and impact of language bias in meta analyses of controlled trials: Empirical study. International Journal of Epidemiology 31(1): 115–123.

26.

Kabat-Zinn

(1996) Mindfulness meditation: What it is, what it isn’t, and it’s role in health care and medicine. In: Haruki

Ishii

Suzuki

(eds) Comparative and Psychological Study on Meditation. Delft, NL: Eburon Publishers, pp. 161–169.

27.

*Kolahkaj

Zargar

(2015) Effect of mindfulness-based stress reduction on anxiety, depression and stress in women with multiple sclerosis. Nursing and Midwifery Studies 4(4): e29655.

28.

Kurtzke

(1983) Rating neurologic impairment in multiple sclerosis: An Expanded Disability Status Scale (EDSS). Neurology 33(11): 1444.

29.

Lazarus

(1991) Progress on a cognitive-motivational-relational theory of emotion. American Psychologist 46: 819–834.

30.

Lazarus

(1999) Stress and Emotion: A New Synthesis. New York: Springer.

31.

Lazarus

(2000) Cognitive-motivational-relational theory of emotion. In: Hanin

(ed.) Emotions in Sport. Champaign, IL: Human Kinetics, pp. 39–63.

32.

Lazarus

Folkman

(1984) Stress, Appraisal and Coping. New York: Springer.

33.

Liberati

Altman

Tatzlaff

, et al. (2009) The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: Explanation and elaboration. British Medical Journal 339: b2700.

34.

Linden

(2005) Stress Management: From Basic Science to Better Practice. Thousand Oaks, CA: SAGE, pp. 83–140.

35.

Lipsey

Wilson

(2001) Practical meta-analysis. Thousand Oaks: SAGE Publications.

36.

McCabe

Ebacioni

Simmons

, et al. (2015) Unmet education, psychological and peer support needs of people with multiple sclerosis. Journal of Psychosomatic Research 78(1): 82–87.

37.

McDonald

Compston

Edan

, et al. (2001) Recommended diagnostic criteria for multiple sclerosis: Guidelines from the international panel on the diagnosis of multiple sclerosis. Annals of Neurology 50: 121–127.

38.

McGuire

Stojanovic-Radic

Strober

, et al. (2015) Development and effectiveness of a psychoeducational wellness program for people with multiple sclerosis: Description and outcomes. International Journal of MS Care 17: 1–8.

39.

Maida

Dalla Costa

Rodegher

, et al. (2014) Overcoming recruitment challenges in patients with multiple sclerosis: Results from an Italian survey. Clinical Trials 11(6): 667–672.

40.

Malcomson

Dunwoody

Lowe-Strong

(2007) Psychosocial interventions in people with multiple sclerosis: A review. Journal of Neurology 254: 1–13.

41.

Marrie

Zhang

Lix

, et al. (2018) The validity and reliability of screening measures for depression and anxiety disorders in multiple sclerosis. Multiple Sclerosis and Related Disorders 20: 9–15.

42.

Mohr

(2007) Stress and multiple sclerosis. Journal of Neurology 254: II65–II68.

43.

Mohr

Goodkin

Likosky

, et al. (1997) Identification of Beck Depression Inventory items related to multiple sclerosis. Journal of Behavioural Medicine 20(4): 407–414.

44.

Mohr

Hart

Julian

, et al. (2004) Association between stressful life events and exacerbation in multiple sclerosis: A meta-analysis. British Medical Journal 328(7442): 731.

45.

*Mohr

Lovera

Brown

, et al. (2012) A randomized trial of stress management for the prevention of new brain lesions in MS. Neurology 79(5): 412–419.

46.

Munoz

SJA

Oreja -Guevara

Lorenzo

, et al. (2016) Psychotherapeutic and psychosocial interventions for managing stress in multiple sclerosis: The contribution of mindfulness-based interventions. Neurologia 31(2): 113–120.

47.

Orwin

(1983) A fail-safe N for effect size in meta-analysis. Journal of Educational Statistics 8: 157–159.

48.

Pagnini

Bosma

Phillips

, et al. (2014) Symptom changes in multiple sclerosis following psychological interventions: A systematic review. BMC Neurology 14: 222.

49.

Peinemann

Labeit

Thielscher

, et al. (2014) Failure to address potential bias in non-randomised controlled clinical trials may cause lack of evidence on patient-reported outcomes: A method study. BMJ Open 4: e004720.

50.

Pham

Platt

McAuley

, et al. (2001) Is there a “best” way to detect and minimize publication bias? An empirical evaluation. Evaluation & the Health Professions 24(2): 109–125.

51.

Polman

Reingold

Banwell

, et al. (2011) Diagnostic criteria for multiple sclerosis: 2010 revisions to the McDonald criteria. Annals of Neurology 69: 292–302.

52.

Polman

Reingold

Edan

, et al. (2005) Diagnostic criteria for multiple sclerosis: 2005 revisions to the “McDonald criteria.” Annals of Neurology 58: 840–846.

53.

Reynard

Sullivan

Rae-Grant

(2014) A systematic review of stress management interventions for multiple sclerosis patients. International Journal of MS Care 16(3): 140–144.

54.

*Sanaeinasab

Saffari

Hashempour

, et al. (2017) Effect of a transactional model education program on coping effectiveness in women with multiple sclerosis. Brain and Behavior 7: e00810.

55.

Schiavon

Marchetti

Gurgel

, et al. (2017) Optimism and hope in chronic disease: A systematic review. Frontiers in Psychology 7: 2022.

56.

Schwartz

(1999) Teaching coping skills enhances quality of life more than peer support: Results of a randomized trial with multiple sclerosis patients. Health Psychology 18: 211–220.

57.

Schweizer

Braun

Milstone

(2016) Research methods in healthcare epidemiology and antimicrobial stewardship-quasi-experimental designs. Infection Control & Hospital Epidemiology 37(10): 1135–1140.

58.

Senders

Hanes

Bourdette

, et al. (2014) Reducing survey burden: Feasibility and validity of PROMIS measures in multiple sclerosis. Multiple Sclerosis 20(8): 1102–1111.

59.

*Senders

Hanes

Bourdette

, et al. (2018) Impact of mindfulness-based stress reduction for people with multiple sclerosis at 8 weeks and 12 months: A randomized clinical trial. Multiple Sclerosis Journal 25: 1178–1188.

60.

Sesel

Sharpe

Naismith

(2018) Psychotherapy efficacy of psychosocial interventions for people with multiple sclerosis: A meta-analysis of specific treatment effects. Psychotherapy and Psychosomatics 87(2): 105–111.

61.

Siegrist

(2018) Work stress and the development of chronic diseases. International Journal of Environmental Research and Public Health 15(3): 536.

62.

Simpson

Booth

Lawrence

, et al. (2014) Mindfulness based interventions in multiple sclerosis—A systematic review. BMC Neurology 14: 15.

63.

*Simpson

Mair

Mercer

(2017) Mindfulness-based stress reduction for people with multiple sclerosis—A feasibility randomised controlled trial. BMC Neurology 17: 94.

64.

Skokou

Soubasi

Gourzis

(2012) Depression in multiple sclerosis: A review of assessment and treatment approaches in adult and pediatric populations. ISRN Neurology 2012: 427102.

65.

Sterne

JAC

Sutton

Ioannidis

JPA

, et al. (2011) Recommendations for examining and interpreting funnel plot asymmetry in meta-analyses of randomised controlled trials. British Medical Journal 343: d4002.

66.

Suzuki

Ito

(2013) Psychological stress. In: Gellman

Turner

(eds) Encyclopaedia of Behavioral Medicine. New York: Springer, pp. 1561.

67.

Tesar

Baumhackl

Kopp

, et al. (2003) Effects of psychological group therapy in patients with multiple sclerosis. Acta Neurology Scandinavia 107(6): 394–399.

68.

Thomas

Hillier

, et al. (2006) Psychological interventions for multiple sclerosis. Cochrane Database of Systematic Reviews 1: CD004431.

69.

Watson

Ford

Worthington

, et al. (2014) Validation of mood measures for people with multiple sclerosis. International Journal of MS Care 16(2): 105–109.

70.

Ziemssen

(2011) Symptom management in patients with multiple sclerosis. Journal of Neurological Science 311: S48–S52.

71.

Ziemssen

Thomas

(2017) Treatment optimization in multiple sclerosis: How do we apply emerging evidence? Expert Review of Clinical Immunology 13(6): 509–511.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.45 MB