Health benefits, safety and cost of physical activity interventions for mental health conditions: A meta-review to inform translation efforts

https://doi.org/10.1016/j.mhpa.2018.11.001Get rights and content

Highlights

  • There is a large evidence base supporting the efficacy of physical activity interventions for many mental health conditions.

  • There is less research in other areas that may be important in informing routine clinical practice.

  • This review found no evidence of cost data and inconsistent reporting on adverse events.

  • The research agenda should expand to include translational studies that report on outcomes useful in real-world settings.

Abstract

Background

Mental illness is a leading cause of the global burden of disease. Physical activity (PA) can improve physical and mental health outcomes for people with mental illness, yet routine implementation of PA within standard care remains ad-hoc. The reasons for this are unclear, although the dissonance between the evidence produced in research settings and that needed in real-world environments may be key.

Purpose

To explore the effectiveness of PA interventions as a treatment for mental illness. We synthesised past systematic reviews and meta-analyses.

Methods

We conducted a systematic review of reviews from database inception to 09/2017. Reviews were included that considered any mental health condition (diagnosed via standardised criteria) and where PA interventions were a stand-alone or adjunctive treatment. Effectiveness was defined as outcomes that are important in real-world healthcare (i.e. expected clinical outcomes, intervention safety and cost).

Results

From 4008 hits, 33 reviews (including 155 unique studies) were included and 32 reported that PA has a positive effect on at least one main outcome of interest (symptoms of mental illness, quality of life and/or physical health). There was inconsistent reporting of adverse events and no cost data was identified. The AMSTAR quality rating suggests inconsistencies in review quality.

Conclusions

The research agenda must expand to report on outcomes that can support evidence translation efforts (i.e. cost and adverse events). Without such a shift, research in PA and mental health may fail to achieve translation to routine care and may have limited impact on patient outcomes.

Introduction

Mental illness represents a growing and significant burden to individuals, communities and the health system. On average, just under 30% of the population will experience a mental illness at some point in their life (Steel et al., 2014). Many people live long periods of their life affected by mental illness, as evidenced through the contribution mental illness makes to years-lived with some form of disability (i.e. disability adjusted life years) (Moussavi et al., 2007; Vigo, Thornicroft, & Atun, 2016). For example, the global cost of mental illness was estimated at $2.5 trillion in 2010, with this figure expected to rise to $6 trillion by 2030 (Marquez & Saxena, 2016). Further, high-prevalence disorders (anxiety, depression, substance use) cost the Australian healthcare system $A974 million and over $11 billion annually when economic costs to society are included (i.e. productively loss) (Lee et al., 2017).

Mental illness is highly co-morbid with physical health problems, such as diabetes and cardiovascular disease (Vancampfort et al., 2015b, 2016). In severe and persistent mental disorders, such as schizophrenia, major depression and/or bipolar (referred to as serious mental illness) (Vancampfort, Rosenbaum, et al., 2017), the co-existence of comorbid physical illness is associated with increased mortality (Hjorthøj, Stürup, McGrath, & Nordentoft, 2017; D.; Lawrence, Hancock, & Kisely, 2013). Critically, life expectancy in this population is ∼15 years less than in people without a mental illness (Hjorthøj et al., 2017; D.; Lawrence et al., 2013) primarily due to the increased prevalence of physical health conditions, such as cancer and cardiometabolic disease (Mental Health Commission of NSW, 2016).

Modifiable risk factors such as poor diet, smoking and a lack of physical activity (PA) are primary drivers of the increased cardiometabolic disease in this population (P. Smith, Mazure, & McKee, 2014; Teasdale, Samaras, Wade, Jarman, & Ward, 2017; Vancampfort, Firth, et al., 2017). As such, efforts to reduce the burden of comorbid cardiometabolic disease often focus on improving these risk factors. In particular, there is increasing international recognition of the importance of addressing physical inactivity among people with mental illness (Rosenbaum et al., 2018). Promisingly, evidence seems to suggest increasing levels of PA may have a dual effect on health outcomes for people with mental illness (Curtis et al., 2016; Schuch, Vancampfort, Richards, et al., 2016c; Stubbs et al., 2017). PA not only improves physical health but can have a direct impact on mental health symptoms across the spectrum of mental illness (Rosenbaum, Tiedemann, Sherrington, Curtis, & Ward, 2014). For example, PA can improve symptoms of anxiety and depression (Schuch, Vancampfort, Richards, et al., 2016c; Stubbs et al., 2017) and together with dietary interventions and life-skills training, can mitigate weight-gain associated with anti-psychotic medication (Curtis et al., 2016). This weight-gain contributes to the development of cardiometabolic disease and early mortality in this population (Álvarez-Jiménez, Hetrick, González-Blanch, Gleeson, & McGorry, 2008).

Coupled with calls for integrated physical and mental healthcare (National Mental Health Commission, 2016; Naylor et al., 2016; Rodgers et al., 2018; Vancampfort, Stubbs, Ward, Teasdale, & Rosenbaum, 2015c), it is reasonable that PA is viewed as a promising intervention to help address the mental ill-health disease burden. However, widespread uptake of PA as part of routine care for people with a mental illness is sub-optimal (Lederman et al., 2017; Pratt et al., 2016; Vancampfort et al., 2015c). This is despite the evidenced-base moving to a stage where government and non-government organisations have endorsed the role of PA in mental health treatment (Mental Health Commission of NSW, 2016; National Institute for Health and Care Excellence (NICE), 2014; Ravindran et al., 2016; The Royal Australian and New Zealand College of Psychiatrists, 2015).

The reasons for poor uptake of PA as part of routine treatment in mental health care are not clear. Although more broadly, the uptake of evidenced-based interventions in routine practice is widely recognised as being complex, problematic and slow (Balas & Boren, 2000; Brownson, Colditz, & Proctor, 2012). Consequently, new fields of research have evolved, dedicated solely to bridging the gap between research and practice (Eccles & Mittman, 2006). For example, implementation and dissemination scientists have consistently identified factors associated with the intervention, the people using the intervention, the setting that hosts the intervention and broader environmental influences that contribute to this gap (implementation) (Durlak & DuPre, 2008). This information is coupled with strategies that actively help to spread effective interventions, so that population-wide benefits are achieved (dissemination) (Rabin, Brownson, Haire-Joshu, Kreuter, & Weaver, 2008). Notably, implementation and dissemination should be predicated on the evidence base of the intervention of interest, wherein only interventions of proven effectiveness should be considered for wider use (Brownson et al., 2012; Frieden, 2013). Further, because there are more effective interventions than can reasonably be funded (Frieden, 2013), consideration of clinical impacts alone is insufficient to demand change (Baltussen & Niessen, 2006; Lavis, Davies, Gruen, Walshe, & Farquhar, 2006). Many factors are often considered when decisions are made to invest in an intervention, including but not limited to; clinical effectiveness, cost-effectiveness, burden of disease analysis and potential harms (Baltussen & Niessen, 2006; Lavis et al., 2006).

Systematic reviews and meta-analyses are typically viewed as the most appropriate method to summarise an interventions’ effectiveness (Centre for Reviews and Dissemination, 2009; Mulrow, 1994) primarily because they are considered comprehensive and reliable summaries of “what works”. However, there is growing acknowledgement that traditional methods used by researchers to synthesise a body of evidence are incongruent with the needs of end-users (Greenhalgh & Russell, 2006; Murthy et al., 2012; Tricco et al., 2015). For example, most systematic reviews are designed to critically analyse the evidence for a specific question (Moher, Liberati, Tetzlaff, Altman, & The Prisma Group, 2009). They often focus on a discrete sample, outcome/s or disease-type in order to be comprehensive and manageable. Unfortunately, this approach can be incongruent with the needs of decision makers aiming to inform practice in complex health organisations and environments.

Acknowledging these challenges, researchers are exploring strategies (Lavis et al., 2005; Murthy et al., 2012; Tricco et al., 2015) and adapting existing methodologies (Cochrane Library, 2014; V.; Smith, Devane, Begley, & Clarke, 2011) to enhance evidence use in healthcare (Anderson et al., 2013; Petticrew et al., 2013; V.; Smith et al., 2011). There is no single methodology that is best suited to achieving this aim, however a systematic review of reviews is one adaption (Aromataris et al., 2015). Reviews of reviews are used to consolidate the results of individual systematic reviews into one document. The benefit of this method is the ability to comprehensively summarise the evidence-base and offer generalised conclusions (Cochrane Library, 2014) to better inform decision making needs (Aromataris et al., 2015; V.; Smith et al., 2011).

In the mental health field this methodology has been used to determine the efficacy of pharmacological versus non-pharmacological interventions for the treatment of major depression (Gartlehner et al., 2017), to investigate the effects of exercise on depressive symptoms in older people (Catalan-Matamoros, Gomez-Conesa, Stubbs, & Vancampfort, 2016) and to determine the effectiveness of drug and physical treatments in mild to moderate or severe depression (Cipriani, Barbui, Butler, Hatcher, & Geddes, 2011). This methodology is used here to explore the effectiveness of PA as a treatment for mental illness. Pointedly, we defined effectiveness as more than just the expected clinical outcomes given earlier acknowledgement of the many factors that can influence adoption of a new intervention (Baltussen & Niessen, 2006; Lavis et al., 2006). As such, effectiveness is defined as: 1) what are the main clinical outcomes expected of the intervention (i.e. changes in mental health symptoms, physical health metrics such as cardiorespiratory risk and/or quality of life); 2) is the intervention safe (i.e. reporting on adverse events); and 3) what are the costs associated with the intervention. We selected cost and adverse events over other factors because they become particularly important when considering the ongoing challenges healthcare systems face in providing safe (Leape et al., 2009), evidence-based care in fiscally demanding environments (Garber et al., 2014; Papanicolas, Woskie, & Jha, 2018). For example, the USA has a higher healthcare spending than other high incomes countries with no resulting improvement to health outcomes (Schneider, Sarnak, Squires, Shah, & Doty, 2017). As such, efforts to maximise efficiencies (through optimising health at the lowest possible cost) is a priority (Garber et al., 2014). Recent studies in mental health have collectively explored these three constructs (Arnberg, Linton, Hultcrantz, Heintz, & Jonsson, 2014; Jonsson et al., 2016), reflecting this priority. Further internationally, cost and safety feature across many guiding documents designed to improve the delivery of evidence-based care (Agency for Healthcare Research and Quality, 2018; Department of Health, 2018; National Institute for Health and Care Excellence (NICE), 2018). Finally, we purposely included both physical and mental health outcomes as a main intervention effect, to address calls from the sector to adopt integrated care practices (National Mental Health Commission, 2016; Naylor et al., 2016; Rebar, Stanton, & Rosenbaum, 2017; Vancampfort et al., 2015c).

This review of reviews aims to:

  • 1)

    Summarise the current literature on the effectiveness of PA interventions as a treatment for mental health conditions.

  • 2)

    Synthesise past systematic reviews and meta-analyses to improve practical application and inform healthcare decisions.

  • 3)

    Identify gaps in the evidence base and provide recommendations for priority research, to support evidence-translation efforts.

Section snippets

Protocol and registration

This systematic review was registered with the PROSPERO database (CRD42017065789). The review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement (Moher et al., 2009). An electronic data-base search was conducted from earliest record to September 2017 using Medline, Cochrane Central Register of Systematic Reviews, Web of Science, Sport Discus, CINAHL, Scopus, PsycINFO. The search strategy is provided in Supplemental Table 1.

Search results

7050 studies were retrieved from the database search and were exported to Endnote (X8) for deduplication. After deduplication, the remaining 4098 articles were imported into Covidence for screening. A second deduplication was completed on importing to Covidence. 4008 articles remained for title and abstract screening with zero additional studies identified through hand-screening. After screening, a total of 33 reviews met the inclusion criteria (Fig. 1). Supplemental Table 4 provides a list of

Symptoms of mental illness

Eight (Blake et al., 2009; Eriksson & Gard, 2011; Heinzel et al., 2015; Krogh et al., 2011; Kvam et al., 2016; Schuch, Vancampfort, Richards, et al., 2016c; Schuch et al., 2016b; Sukhato et al., 2017) of the nine identified reviews reported a positive effect favouring PA to reduce symptoms of depression. The magnitude of this effect ranged from small to large, somewhat dependent upon the methodological approach (Supplemental Table 2). For example, reviews where the methodology included another

Symptoms of mental illness

All eight reviews examining changes in mental health symptoms in people diagnosed with schizophrenia reported a positive effect favouring PA. The positive effects appear subject to moderators associated with the PA intervention. Keller-Varady and colleagues (Keller-Varady et al., 2017) reported the positive effects favouring PA included interventions that combined both aerobic and strength training. Firth et al.(Firth et al., 2015) reported a greater treatment effect when low-intensity PA

Symptoms of mental illness

Three reviews (Klatte et al., 2016; Rosenbaum et al., 2014, 2015a) explored the effect of PA interventions on reducing symptoms of depression in people with a mental illness, where the population included people with different diagnoses. All three reviews reported positive effects favouring PA interventions. Of these, two reviews (Klatte et al., 2016; Rosenbaum et al., 2014) reported moderate to large effect sizes favouring the PA interventions. One of these reviews (Rosenbaum et al., 2014) was

Symptoms of mental illness

Three reviews reported a positive effect favouring PA to reduce symptoms of anxiety, two of these reviews (Jayakody et al., 2014; Stubbs et al., 2017) stated this effect was significant.

Quality of life

One review (Jayakody et al., 2014) reported a non-significant, positive effect favouring PA to improve quality of life outcomes.

Physical health metric

None of the included reviews reported on physical health outcomes for people diagnosed with an anxiety disorder.

Adverse events and cost

One review (Jayakody et al., 2014) reported on adverse events, stating no

Post-traumatic stress disorder

Two reviews (S. Lawrence et al., 2010; Rosenbaum et al., 2015) met our inclusion criteria, however one of these reviews (S. Lawrence et al., 2010) did not find any individual studies that met their inclusion criteria. The two reviews were conducted five years apart, with the earlier review (S. Lawrence et al., 2010) failing to find any individual studies that met their inclusion criteria. As such, the evidence reported below demonstrates the most recent data for patients with PTSD and is

Symptoms of mental illness

Two reviews (Hallgren et al., 2017; Wang et al., 2014) reported positive effects favouring PA to reduce symptoms of depression in people diagnosed with AUD/SUD. One review (Wang et al., 2014) suggested a positive effect favouring PA to improve abstinence in SUD. A second review (Hallgren et al., 2017) showed no impact on consumption of alcohol in AUD. These findings suggest that the different diagnosis (SUD versus AUD) and the different outcomes measured may influence the effectiveness of

Evidence summary

Table 2 provides a summary of the evidence from the review. Notwithstanding the inconsisent reporting of adverse events and lack of cost data, the clinical outcomes suggest that positive gains could be achieved in people diagnosed with depression, schizophrenia, multiple mental health diagnoses and anxiety through participation in PA internventions. Preliminary evidence is promising in PTSD and AUD/SUD, however, due to the limited number of systematic reviews and included studies, more research

Discussion

From the 33 systematic reviews/meta-analyses included in this review, 32 reported outcomes that support the effectiveness of PA interventions (including yoga, tai chi, general PA and structured exercise) on at least one of the main clinical outcomes of interest (symptoms of mental illness, quality of life and/or physical health metric). There was large variation in the effect size and significance of the results (including non-significant findings); however, the overall trend was positive and

Conclusion

A large body of evidence reports the positive impact PA interventions have on clinical outcomes in people with a mental illness. However, adverse events are not routinely reported and to date no cost data exists in systematic reviews and/or meta-analyses. To build an evidence base that is applicable in real-world health services it is essential to collect and report this, and other relevant data in research studies. While reviewing the full range of evidence needed to influence translation and

Funding

BS is part supported by the National Institute for Health Research (NIHR) Collaboration for Leadership in Applied Health Research and Care South London at King's College Hospital NHS Foundation Trust. BS is also part funded by the National Institute for Health Research (NIHR) Biomedical Research Centre at South London and Maudsley NHS Foundation Trust and King's College London. The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR or the Department of

All authors declare they have no conflict of interest in relation to this study

This article does not contain any studies with human participants performed by any of the authors.

This study does not involve human participants and informed consent was therefore not required.

This article does not contain any studies with animals performed by any of the authors.

References (119)

  • M. Petticrew et al.

    Synthesizing evidence on complex interventions: How meta-analytical, qualitative, and mixed-method approaches can contribute

    Journal of Clinical Epidemiology

    (2013)
  • A. Rebar et al.

    Comorbidity of depression and anxiety in exercise research

    The Lancet Psychiatry

    (2017)
  • S. Rosenbaum et al.

    Online physical activity interventions for mental disorders: A systematic review

    Internet Interventions

    (2015)
  • S. Rosenbaum et al.

    Physical activity in the treatment of post-traumatic stress disorder: A systematic review and meta-analysis

    Psychiatry Research

    (2015)
  • R. Rosenthal

    The file drawer problem and tolerance for null results

    Psychological Bulletin

    (1979)
  • F. Schuch et al.

    Exercise as a treatment for depression: A meta-analysis adjusting for publication bias

    Journal of Psychiatric Research

    (2016)
  • F. Schuch et al.

    Exercise improves physical and psychological quality of life in people with depression: A meta- analysis including the evaluation of control group response

    Psychiatry Research

    (2016)
  • B. Shea et al.

    AMSTAR is a reliable and valid measurement tool to assess the methodological quality of systematic reviews

    Journal of Clinical Epidemiology

    (2009)
  • B. Stubbs et al.

    Exercise improves cardiorespiratory fitness in people with depression: A meta-analysis of randomized control trials

    Journal of Affective Disorders

    (2016)
  • B. Stubbs et al.

    An examination of the anxiolytic effects of exercise for people with anxiety and stress-related disorders: A meta-analysis

    Psychiatry Research

    (2017)
  • A.H. Taylor et al.

    Evidence and theory into practice in different health care contexts: A call for more translational science

    Mental Health and Physical Activity

    (2014)
  • Agency for Healthcare Research and Quality

    Mission and budget

  • M. Álvarez-Jiménez et al.

    Non-pharmacological management of antipsychotic-induced weight gain: Systematic review and meta-analysis of randomised controlled trials

    The British Journal of Psychiatry

    (2008)
  • American Diabetes Association

    Standards of medical care in diabetes—2015 abridged for primary care providers

    Clinical Diabetes : A Publication of the American Diabetes Association

    (2015)
  • F.K. Arnberg et al.

    Internet-delivered psychological treatments for mood and anxiety disorders: A systematic review of their efficacy, safety, and cost-effectiveness

    PLoS One

    (2014)
  • E. Aromataris et al.

    Summarizing systematic reviews: Methodological development, conduct and reporting of an umbrella review approach

    International Journal of Evidence-based Healthcare

    (2015)
  • M. Bak et al.

    Almost all antipsychotics result in weight gain: A meta-analysis

    PLoS One

    (2014)
  • E. Balas et al.

    Managing clinical knowledge for health care improvement

    IMIA Yearbook

    (2000)
  • R. Baltussen et al.

    Priority setting of health interventions: The need for multi-criteria decision analysis

    Cost Effectiveness and Resource Allocation

    (2006)
  • H. Blake et al.

    How effective are physical activity interventions for alleviating depressive symptoms in older people? A systematic review

    Clinical Rehabilitation

    (2009)
  • R. Brownson et al.

    Dissemination and implementation research in health

    (2012)
  • J. Caird et al.

    Mediating policy-relevant evidence at speed: Are systematic reviews of systematic reviews a useful approach?

    Evidence and Policy: A Journal of Research, Debate and Practice

    (2015)
  • Centre for Reviews and Dissemination

    CDR's Guidance for undertaking reviews in healthcare

    (2009)
  • A. Cipriani et al.

    Depression in adults: Drug and physical treatments

    BMJ clinical evidence

    (2011)
  • Cochrane Library

    Special collection - Cochrane overview

    (2014)
  • J. Cohen

    Statistical power analysis for the behavioral sciences

    (1988)
  • J. Curtis et al.

    Evaluating an individualized lifestyle and life skills intervention to prevent antipsychotic-induced weight gain in first-episode psychosis

    Early Intervention in Psychiatry

    (2016)
  • A. Daley

    Exercise and depression: A review of reviews

    Journal of Clinical Psychology in Medical Settings

    (2008)
  • L. Damschroder et al.

    Fostering implementation of health services research findings into practice: A consolidated framework for advancing implementation science

    Implementation Science

    (2009)
  • M. Dauwan et al.

    Exercise improves clinical symptoms, quality of life, global functioning, and depression in schizophrenia: A systematic review and meta-analysis

    Schizophrenia Bulletin

    (2016)
  • Department of Health

    Definitions

  • Department of Health

    Medical services advisory committee

    (2018)
  • J. Durlak et al.

    Implementation matters: A review of research on the influence of implementation on program outcomes and the factors affecting implementation

    American Journal of Community Psychology

    (2008)
  • M. Eccles et al.

    Welcome to implementation science

    Implementation Science

    (2006)
  • S. Eriksson et al.

    Physical exercise and depression

    Physical Therapy Reviews

    (2011)
  • J. Firth et al.

    A systematic review and meta-analysis of exercise interventions in schizophrenia patients

    Psychological Medicine

    (2015)
  • T. Frieden

    Six components necessary for effective public health program implementation

    American Journal of Public Health

    (2013)
  • S. Garber et al.

    Redirecting innovation in U.S. Health care: Options to decrease spending and increase value

    Rand Health Quarterly

    (2014)
  • G. Gartlehner et al.

    Pharmacological and non-pharmacological treatments for major depressive disorder: Review of systematic reviews

    BMJ Open

    (2017)
  • T. Greenhalgh et al.

    Evidence based medicine: A movement in crisis?

    BMJ British Medical Journal

    (2014)
  • Cited by (63)

    • Cardiorespiratory fitness and self-reported physical activity levels of referring mental healthcare professionals, and their attitudes and referral practices related to exercise and physical health

      2022, Journal of Psychiatric Research
      Citation Excerpt :

      A personal interview and referral to an exercise professional were perceived as the most effective ways to advise about PA. These findings are consistent with Fibbins et al. (2018a) and correspond with previous evidence showing larger effects and fewer dropouts in interventions led by qualified exercise professionals (Czosnek et al., 2019; Schmitt et al., 2018; Stubbs et al., 2016, 2018; Vancampfort et al., 2016b, 2019, 2021). Referrals to the gym or community services were found less effective.

    View all citing articles on Scopus
    View full text