WHAT IS ALREADY KNOWN ON THIS TOPIC

People with disabilities are at least twice as likely to be physically inactive as those without disabilities across their adult lifespan.

Current evidence indicates health and psychosocial benefits of sport and physical recreation for the general population, but there has not been a synthesis of evidence using meta-analysis evaluating these two types of physical activities for people living with disabilities from a range of health conditions.

WHAT THIS STUDY ADDS

Physical recreation may provide small to large effects on mobility and quality of life (and the secondary outcomes of fatigue, depression and anxiety) for people living with disabilities.

The certainty of the evidence was low or very low and predominantly for people living with a physical disability.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

Several gaps in the evidence were identified where evidence generation is urgently needed: sport and recreation interventions delivered in inclusive community settings, younger adults and adults living with intellectual disabilities.

Introduction

About 16% of the global population experience disability; this rate is increasing with an ageing population and growth in the number of people living with non-communicable diseases and surviving serious injury.1 The United Nations Convention on the Rights of Persons with Disability defines people with disabilities as ‘those who have long-term physical, mental, intellectual or sensory impairments which in interaction with various barriers may hinder their full and effective participation in society on an equal basis with others’.2 People with disabilities face health inequities and are at greater risk of morbidity and mortality. There is strong evidence that physical activity reduces risk of morbidity and mortality and benefits many health outcomes relevant for people with disabilities.3–5 National physical activity guidelines from countries including the UK and the USA and international guidelines from the WHO now include guidance for people living with a disability.3–6 The WHO guidelines recommend the same amount of physical activity for adults with disabilities as the general population. However, people with disabilities are at least twice as likely to be physically inactive as those without disabilities,7 across their adult lifespan.8 Strategies to increase physical activity for people with disabilities are a global priority area identified in the WHO Global Physical Activity Action Plan 2018-2030.9

Sport and physical recreation are physical activities suitable for people of all ages and abilities.9 Sport is ‘an activity involving physical exertion, skill and/or hand–eye co-ordination as the primary focus of the activity, with elements of competition where rules and patterns of behaviour governing the activity exist formally through organisations’10 and physical recreation is ‘an activity or experience that involves varying levels of physical exertion, prowess and/or skill, which may not be the main focus of the activity, and is voluntarily engaged in by an individual in leisure time for the purpose of mental and/or physical satisfaction’.10 Sport and physical recreation can be modified or use equipment to enable the participation of people with disabilities in either mainstream or disability-specific activities.11 While facilitators including enjoyment and social contact often encourage participation in sport and physical recreation by people with disabilities,12–14 data shows lower participation rates compared with adults without disabilities.8

High-quality evidence supports the health and psychosocial benefits of sport participation for the general population,15–17 with growing evidence of effectiveness of physical recreation such as yoga18 and Tai Chi.19 One could assume participation in sport or physical recreation would have equal if not greater health benefits for people with disabilities given they are more inactive and often more socially isolated than their non-disabled peers. Although there is some evidence of effectiveness of particular physical recreation activities in people living with specific health conditions (eg, yoga for stroke20) and benefits from any type of physical activity for people living with disability,3 4 there has not been a rigorous systematic review with meta-analysis of trial-level evidence for the interventions of any sport or physical recreation for people with disabilities. While within and between trial variability poses challenges for analysis, research synthesis from a range of health conditions (rather than for specific diagnoses) would facilitate scalability and provide evidence to support advocacy and service development. Identifying and synthesising the current evidence may provide support for the provision of inclusive and community-based sporting and physical recreation opportunities for people with disabilities and identify where evidence is lacking and research funding is needed.

This review aims to evaluate the effects of sport or physical recreation for adults with physical or intellectual disabilities. The primary outcomes were participation, activity (mobility) and quality of life, and the secondary outcomes were impairment (fatigue, anxiety or depression), care needs, socioeconomic benefit and adverse events. These outcomes were selected because they are likely to be considered important by people with disability and/or represent a sample of the domains in the International Classification of Functioning, Disability and Health.21

MethodDesign

Systematic review with meta-analysis that follows methods described by Cochrane22 and is reported according to the Preferred Reporting Items for Systematic Reviews and Meta-Analysis guidelines.23 Data are available in a public, open access repository (Sydney eScholarship Repository, https://hdl.handle.net/2123/31569.2, Creative Commons Attribution V.4.0 Licence).24

Search strategy

Six electronic databases were searched from inception to May 2022: Medline (Ovid), CINAHL (EBSCO), Embase (Ovid), SPORTDiscus (EBSCO), PsycINFO (Ovid) and Physiotherapy Evidence Database (PEDro). The search strategy was developed for Medline and adapted for each database in consultation with a university librarian. Search terms were a mix of database-controlled keywords, Medical Subject Headings and text words. The search strategy combined the ‘population’ AND ‘intervention’ search terms (see Appendix 1, supplementary file24). We also citation tracked relevant systematic reviews and searched the International Clinical Trials Registry Platform (apps.who.int/trialsearch). No language or publication date restrictions were imposed.

Selection criteriaStudy type

Randomised controlled trials were included.

Population

Trials involving adults with intellectual and/or physical disabilities from any congenital or acquired health condition resulting in physical impairments eligible for parasports25 and/or intellectual impairments eligible for Special Olympics.26 Examples of eligible health conditions include acquired brain injury, spinal cord injury, cerebral palsy, Parkinson’s disease, Down syndrome and amputation. Trials including participants with sensory impairments (hearing, vision), psychiatric or acute health conditions were excluded.

Intervention

Trials examining the effects of sport or physical recreation interventions conducted in community or clinical settings. Physical recreation activities where most of the activity was a structured exercise programme (aerobic, strength or neuromotor exercises) or routine rehabilitation or conducted in participants’ home, education or employment setting were excluded.

Comparator

The comparator was either no intervention, both groups receiving usual care or an intervention that did not include participation in sport or physical recreation.

Outcome measures

We included health-related outcomes under the following domains: participation, activity and quality of life (primary outcomes) and impairment, care needs, socioeconomic benefit and adverse events (secondary outcomes). Participation was defined as involvement in life situations measured as employment, connectedness and engagement in community21 (including attendance or involvement).27 Activity was defined as ‘the execution of a task or action by an individual’,21 measured by self-reported or performance-based tests of mobility. Quality of life was defined as ‘optimum levels of mental, physical, role and social functioning, including relationships, and perceptions of health, fitness, life satisfaction and well-being’,28 measured with multiattribute generic or health condition specific instruments. Impairments were defined as ‘problems in body function or structure as a significant deviation or loss’21 measured by self-reported or performance-based tests of fatigue, anxiety or depression. Care needs or socioeconomic benefit were measured from any relevant outcome measures or data included in the trials. Adverse events included injuries while participating in the sport or physical recreation or during the intervention.

Data management and selection procedure

Searches were combined in EndNote and duplicates were removed. One reviewer (LH or AMM) did an EndNote search to identify potential trials (‘rand’ in the title or abstract field). Two reviewers independently screened the title and abstract for eligibility. Any disagreements were resolved by discussion or arbitration by a third reviewer (LH or NS). Full text was obtained for potentially eligible trials and independently assessed for eligibility by two reviewers with disagreements resolved through team discussion.

Data extraction

Data extraction was completed independently by two reviewers using a customised Microsoft Excel spreadsheet. Data were compared and any discrepancies resolved via discussion. Data extraction included trial information, population, intervention, comparator and outcomes measured at baseline, end of intervention and follow-up. Where multiple outcome measures were used in a single trial to assess similar constructs, the most suitable measure was selected based on a prioritisation list developed by the authors using the International Classification of Functioning, Disability and Health (Appendix 2, supplementary file24).21

Quality appraisal

The PEDro Scale29 was used for quality appraisal, with scores downloaded from PEDro (pedro.org.au). Each trial was assigned a score (0–10), with a lower score indicating a greater risk of bias, while a score of ≥7 represents a trial of moderate to high quality.30 No trials were excluded based on the quality appraisal score.

Data synthesis

We used RevMan V.5.4 for meta-analysis.31 Mean difference (MD; outcomes measured using same scale) or standardised MD (SMD; Hedges’ g; outcomes measured using different scales) and 95% CI were calculated using a random effects model because of the clinical diversity of the participants and interventions (fixed effects models are included in the supplementary file24 for interested readers). Means and SD were used; if necessary, these were imputed from other variables using Cochrane guidance22 or extracted from figures using WebPlotDigitizer.32 Effect sizes were categorised as small (0.1–0.4), medium (0.5–0.7) or large (≥0.8).33 Heterogeneity was determined by visual inspection of the forest plots and with consideration of the I2 test. Certainty of evidence was determined using the Grading of Recommendations Assessment, Development and Evaluation system (GRADE) approach.22 Funnel plots were drawn to assess publication bias for meta-analysis with ≥10 included trials. Sensitivity analyses were conducted for health condition specific considerations (ie, measures taken ‘on’ and ‘off’ medications for Parkinson’s disease), where data were presented in ways that could not be pooled (ie, walking measured in speed vs time, state vs trait anxiety), where there was a methodological concern with an included trial(s) and where omitting trials reported change scores.

Subgroup analysis

Subgroup analysis was conducted for type of activity and disability.

Equity, diversity and inclusion statement

This review was initiated by the chief executive officer of a national sporting organisation for people with disability to inform policy and advocacy work. Our review included trials from any country that evaluated sport or physical recreation in the marginalised group of people living with physical or intellectual disabilities. Our research team, with assistance from bilingual colleagues, were able to screen and extract data for articles written in languages other than English (including from low-income and middle-income countries) to ensure all eligible trials were included. Members of our research team were all women and included senior, mid-career and early-career academics, a chief executive officer of a national sporting organisation for people living with a disability and research students from three countries. Equity was paramount to our justification for this review and at the forefront of the discussion of our results and implications for research and practice.

ResultsSearch results and overview

Searches identified 26 034 records (figure 1). After filtering for randomised controlled trials and duplicates, title and abstract screening was performed on 3859 records and full-text screening on 357 reports. There were 133 reports (see supplementary table24) for 74 included trials with 79 intervention and 74 control groups. In total, 60 trials (64 intervention groups) were included in the meta-analyses.

Figure 1

Preferred Reporting Items for Systematic Reviews and Meta-Analysis flow diagram.

Trial characteristics

In total, 4 included trials had multiple intervention groups and a control group,34–37 1 article reported the results of two separate trials38 and 69 trials had one intervention and one control group (supplementary table24). Average trial quality was below moderate quality (mean (SD), PEDro Score 5.4 (1.3)). The quality items with the lowest achievement were intention to treat analysis (27%), concealed allocation (30%), adequate follow-up (51%) and assessor blinding (54%). The included trials were conducted across 18 countries, with the majority from the USA (22/74), Iran (11/74) and China (10/74).

Participant characteristics

There were 2954 participants with an average age of 55 years (range 30–76) and 54% were women (supplementary table24). Ten trials included women only,35 37 39–46 four trials did not report gender47–50 and the remaining trials included a mix of genders. All but one trial included participants with a single health condition (table 1).48 Most trials included participants with physical disabilities (70/74), with only four trials including adults with intellectual disabilities.39 51–53 Of those trials including participants living with a physical disability, the majority had a degenerative health condition (Parkinson’s disease 27/70; multiple sclerosis 21/70). There were fewer trials for an acquired disability (stroke 16/70; traumatic brain injury 2/70; spinal cord injury 2/70). Most trials reported participants had a mild to moderate disability (45/74 trials). Average time post onset of diagnosis was 7.5 years.

Table 1

Characteristics of the included participants and interventions: disability type, health condition, intervention, number (N) of comparisons (compar) and participants (partic); duration and frequency of the intervention; training and/or experience of the activity leader; health professional involvement in the intervention; and control interventions

Intervention characteristics

No included trial evaluated sport. Physical recreation interventions evaluated were traditional Chinese exercise (eg, Tai Chi, Qigong; 28/79), yoga (21/79), dance (14/79), aquatic exercise (9/79), gym training (3/79) and boxing, horse riding, Nordic walking and start to run (1/79 each) (table 1, supplementary table24). The setting was the community (26/79), clinic (27/79), university (8/79) or not reported (18/79). Most interventions were conducted in a group 72/79, however only 18 interventions reported the number of participants per group (range 5–16) and only 11 reported the supervision ratio (ranging from 2–4 personnel for each participant54 to 1 personnel for 16 participants55).

Average intervention duration was 65 min/session (range 30–240) conducted at a frequency of 2 sessions/week (range 0.17–10) for 13 weeks (range 6–52). Few trials reported intervention intensity (19/79). Most interventions were led by a person with experience and/or training and/or accreditation in the activity (68/79), but fewer reported disability experience and/or training (29/79). Extra personnel to support the intervention were reported for 18/79 interventions, including students as dance partners and lifeguards for aquatic activities. Health professional involvement was reported for 30/79 interventions, with physiotherapists most common (12), and included supervising the activity or activity leader (16), designing (6) or adapting (1) the programme, being the activity leader (9) and training the activity leader (3). Few trials reported including a home programme (21/79). Modifications were reported for 37 interventions, how programmes were progressed were reported in 42 interventions and what materials were used were reported in 40 interventions. Some activities were designed specifically for the disability group, and modifications included breaking down or simplifying or reducing the time or options for movements or poses, providing additional support (people, environmental, cognitive, behavioural) and modified or additional equipment (eg, blocks, straps, chairs, floatation aids).

Comparator characteristics

The included trials involved comparators of no intervention (38/74), wait-list control (22/74) or an alternative non-physical recreation intervention (14/74).

Outcome measures

Included trials measured participation (40/74), activity (49/74), quality of life (42/74), impairment (57/74) and adverse events (35/74) at the end of intervention. Care burden and socioeconomic benefit were not measured in any trial. Activity included five measures of mobility: combined mobility, walking endurance, walking speed, Timed Up and Go test56 and Berg Balance Scale.57 Quality of life included three measures: total score, physical health summary score and mental health summary score. Impairment included measures of fatigue (overall fatigue, physical fatigue, cognitive fatigue), depression and anxiety.

For participation, attendance was reported (43/79 experimental interventions) as either percentage of sessions attended or percentage of participants who completed a set proportion of sessions (supplementary table24). Average attendance was 81% (range 54%–100%; 30/79) and percentage completing a set proportion of sessions was 83% (range 50%–100%; 13/79). Three trials included a standardised measure of participation,58–60 all measures of attendance.27 The total number of reported adverse events (supplementary table24) were experimental groups 75 (21 related and 54 not related to the intervention) and control groups 68 (9 related and 59 not related to the intervention).

Meta-analysis

Meta-analyses for activity, quality of life and impairment are summarised in figure 2 (disability subgroups) and figure 3 (physical recreation subgroups). Effect sizes are given in table 2. Forest plots (random and fixed effects), funnel plots, GRADE ratings and sensitivity analyses are presented in the supplementary file.24

Figure 2

Summary of meta-analyses for any type of physical recreation, disability subgroups compared with control (no participation in sport or physical recreation). The number of comparisons is entered for each subgroup for each outcome, with data for all disabilities combined in the top row (empty cells indicate no data). The certainty of evidence is indicated for ‘all populations’. The shading indicates the direction of the result.

Figure 3

Summary of meta-analyses for all disability types, physical recreation subgroups compared with control (no participation in sport or physical recreation). The number of comparisons is entered for each subgroup for each outcome, with data for all interventions combined in the top row (empty cells indicate no data). The certainty of evidence is indicated for ‘all interventions’. The shading indicates the direction of the result.

Table 2

Number of participants, effect size, heterogeneity and analysis number for each outcome for the two comparisons: any type of physical recreation (disability subgroups) and all disability types (physical recreation subgroups)

Any type of physical recreation with disability subgroupsActivity

There was a small effect on combined mobility favouring physical recreation, driven mostly by the Parkinson’s disease subgroup. There was a positive effect for walking endurance favouring physical recreation, also with the biggest effect in the Parkinson’s disease subgroup. The entire 95% CI was smaller than the average minimal detectable change of 68.5 m for older adults,61 while much of the 95% CI was above the minimal important difference of 30 m for people with chronic respiratory disease.62 There was no effect of physical recreation for walking speed. A reduced time to complete the Timed Up and Go test favouring physical recreation was demonstrated. The entire 95% CI was smaller than the average minimal detectable change of −5.0 s for older adults.61 For the subgroup analyses, the biggest difference was demonstrated in acquired physical disability. Improved balance (increased score on Berg Balance Scale, range 0–56 points) favouring recreation was observed. The entire 95% CI was smaller than the average minimal detectable change of 5.0 points for older adults.61 The biggest effect was in the multiple sclerosis subgroup.

Quality of life

No trials measured this variable in people with intellectual disabilities. A small effect favouring recreation was found for total score and physical and mental health summary scores, but the 95% CI included no effect for total score.

Impairment

There was a large effect on overall fatigue favouring physical recreation and, for people with multiple sclerosis only, medium to large effects for physical and cognitive fatigue. There was a medium effect on depression favouring physical recreation, with the biggest effect in those with multiple sclerosis. There was a small effect on anxiety favouring physical recreation, with the biggest effect in acquired physical disability.

All disability types with physical recreation subgroupsActivity

There was a small overall effect favouring physical recreation for combined mobility, but the 95% CI includes no effect. Moderate effects that do not include 0 in the 95% CI were evident for the dance subgroup. There was a positive effect favouring physical recreation for walking endurance, with the upper bound of the 95% CI exceeding the average minimal detectable change of 68.5 m for older adults61 and the minimal important difference of 30 m for people with chronic respiratory disease.62 An improved performance on the Timed Up and Go test favouring physical recreation was demonstrated, but this is smaller than the average minimal detectable change of 5.0 s for older adults.61 There was improved balance (Berg Balance Scale) favouring physical recreation, but this was smaller than the average minimal detectable change of 5.0 points for older adults.61 Across these four measures, largest effects were observed for dance, aquatic exercise, yoga and traditional Chinese exercise. There was no effect for physical recreation on walking speed.

Quality of life

There was no effect of physical recreation on the total score and a small effect favouring physical recreation for the physical and mental health summary scores. However, the 95% CI for the physical health summary score included no effect.

Impairments

There was a large effect favouring physical recreation on overall and physical fatigue. There was a moderate effect on cognitive fatigue favouring physical recreation, but the 95% CI included no effect. For depression, there was a moderate effect favouring physical recreation. For anxiety, there was a small effect favouring physical recreation. For all impairment measures, the largest effects were observed for aquatic exercise.

The certainty of evidence was either low or very low for all outcomes. For physical recreation and disability types subgroup comparisons, sensitivity analyses revealed that using walking time instead of walking speed meant that the 95% CI no longer included no effect and omitting trials with change scores for combined mobility or using trait (not state) scores for anxiety meant that the 95% CI included no effect.

Discussion

This large, rigorous systematic review of 74 trials from 18 countries with 2954 participants concludes that physical recreation may have positive effects on multiple outcomes important for people living with disabilities. There were small to large effects of physical recreation on mobility, quality of life, fatigue, depression and anxiety. However, the evidence was graded as low or very low certainty, mostly due to risk of bias and inconsistency. Subgroup analyses indicated stronger effects for people with Parkinson’s disease, multiple sclerosis or acquired disability and for the physical recreation activities of dance, aquatic exercise, yoga and traditional Chinese exercise across different outcomes. Given the range of positive outcomes and that fun and enjoyment are facilitators for participation,63 our findings align with previous coproduced physical activity recommendations4 that adults living with disability can choose a recreation activity they enjoy which they are likely to maintain in the long term.

Our positive findings across multiple outcomes agree with the health condition evidence summaries supporting the WHO physical activity guideline for people living with a disability6 as well as the evidence reviews informing the physical activity guidelines for the UK4 and the USA.3 Our review updates and builds on this evidence-base by synthesising trial-level evidence using meta-analysis to demonstrate effectiveness on a range of outcomes specifically for physical recreation activities. We demonstrated medium to large effects on fatigue and small to medium effects on depression and anxiety with physical recreation. Given these impairments are commonly reported interpersonal barriers to physical activity participation,63 activities that can improve these impairments are likely to facilitate ongoing participation.

All but one intervention was delivered in condition-specific groups, with no interventions delivered in inclusive community-based classes. While being active with people with similar health concerns may be a preference for some, for others this may restrict participation with family and friends in local opportunities63 and may increase the cost and distance of travel.12 If identified as a research priority by people with disabilities, large-scale and well-designed trials would be required to evaluate the benefits and harms of inclusive community-based classes. These trials could minimise within trial heterogeneity by using stratification to match groups for known prognostic factors, including stratification by impairment or health condition. This stratification would enable preplanned testing via subgroup analyses and provide subgroup data for future meta-analyses.

Providing physical recreation for people with disabilities required minor modifications (eg, seated, shorter duration poses) and relatively inexpensive equipment (eg, chairs, exercise bands). This would suggest that inclusive community-based recreation groups could accommodate the needs of people living with mild to moderate disability.11 By providing advice on suitable modifications and equipment, health professionals (eg, physiotherapists) could maximise safety and participation.8 Social care professionals (eg, social workers) could play a role in promoting participation.64 Importantly, only a small number of activity leaders were reported to have experience and/or training working with people living with a disability. This has been reported as a barrier for participation for people with disabilities, as has accessibility and organisational policies.63 These barriers highlight the need for disability education and better cross-sectorial relationships between health, disability and community to enable an inclusive environment6 as well as the need to challenge and reduce disablism and ableism more broadly.65

This review identified several evidence gaps. No trials evaluated sport. A recent review identified benefits of sport for older adults66 and there is strong evidence of benefits of sport in the general population.17 Sport is likely advantageous for people living with a disability and therefore should be encouraged, however the generation of rigorous evidence of benefit would support advocacy of appropriate (inclusive and disability specific) sporting opportunities being available in the community. Other gaps were for adults with intellectual disabilities (only four trials included)39 51–53 and younger adults (youngest average age of included trials was 30 years).44 Physical activity commonly declines in young adulthood as people (with and without disabilities) move away from structured activities in school and may become engaged in more sedentary pursuits like tertiary study or work,67 so may be a crucial time to intervene. Engaging and evaluating young adults in sport and physical recreation should be a priority area for future research. Lack of participation level measurement that focus on involvement was also evident.27 One reason for this may be the suitability of participation measures,27 however this situation is improving and should be revisited in order to provide greater justification for the provision of sport and physical recreation for people living with disabilities. Lastly, we used subgroup analyses to explore two sources of heterogeneity, health condition and recreation intervention. Future research is required to explore other sources of variability (eg, severity).

This review used integrated knowledge translation coproduction methods (a collaborative process where academic researchers work with knowledge users)68 as it was initiated by the chief executive officer of a national sporting organisation for people with disability in collaboration with academic partners. However, this work did not include any authors with lived experience of disability. Future work of this kind would be strengthened by using equitable and experientially informed coproduction methods,68 ensuring equitable partnerships with lived experience representatives and that they are involved in the project from study inception to authorship as has been demonstrated in previous work.4 Other limitations in this review mainly relate to the poor reporting of interventions in trials, which did not follow recommended reporting guidelines such as Template for Intervention Description and Replication (TIDieR)69 or implementing Prisma in Exercise, Rehabilitation, Sport medicine and SporTs science (PERSIST).70 For example, there was limited information on group sizes and on the supervision ratio for group activities, which limits the ability to confidently replicate interventions in practice. There was one deviation from the review protocol in which we stated we would use the Downs and Black appraisal tool71 as we planned to include both randomised and non-randomised studies. Given the large number of trials evaluating physical recreation, we decided to split the reporting of our review into two articles: this article reporting meta-analyses of randomised trials of physical recreation interventions (with appraisal using the PEDro Scale29) and a planned article synthesising non-randomised studies investigating sport interventions for people living with disabilities. This review of trials attempted to summarise evidence for a very broad population (adults living with physical or intellectual disabilities) and intervention (any physical recreation), with meta-analyses showing considerable heterogeneity that reduced the certainty of evidence. However, we chose to investigate a broad population and intervention with the view to scalability and sustainability. We provided subgroup analysis for different disability types and different physical recreation activities to demonstrate where the evidence was strongest.

Conclusion

Based on a rigorous systematic review with meta-analysis of trial-level data, physical recreation may confer multiple health benefits for people with disabilities and is a potentially scalable strategy to support policy and advocacy to address the inequity in their participation. Further research should investigate sport and physical recreation activities delivered in inclusive community settings to demonstrate feasibility and effectiveness of inclusive participation.

Data availability statement

Data are available in a public, open access repository.

Ethics statementsPatient consent for publication

Not applicable.

Ethics approval

Not applicable.

Acknowledgments

The authors would like to thank librarian, Disability Sports Australia, and also Fereshteh Pourkazemi, Qiuzhe Chen and Giane Caon Camara for assistance with data extraction for trials written in Persian, Chinese and Spanish.