Mobility training for increasing mobility and functioning in older people with frailty

Daniel Treacy; Leanne Hassett; Karl Schurr; Nicola J Fairhall; Ian D Cameron; Catherine Sherrington

doi:10.1002/14651858.CD010494.pub2

. 2022 Jun 30;2022(6):CD010494. doi: 10.1002/14651858.CD010494.pub2

Mobility training for increasing mobility and functioning in older people with frailty

Daniel Treacy ¹, Leanne Hassett ², Karl Schurr ³, Nicola J Fairhall ^4,^✉, Ian D Cameron ⁵, Catherine Sherrington ⁴

Editor: Cochrane Musculoskeletal Group

PMCID: PMC9245897 PMID: 35771806

Abstract

Background

Frailty is common in older people and is characterised by decline across multiple body systems, causing decreased physiological reserve and increased vulnerability to adverse health outcomes. It is estimated that 21% of the community‐dwelling population over 65 years are frail. Frailty is independently predictive of falls, worsening mobility, deteriorating functioning, impaired activities of daily living, and death. The World Health Organization's International Classification of Functioning, Disability and Health (ICF) defines mobility as: changing and maintaining a body position, walking, and moving. Common interventions used to increase mobility include functional exercises, such as sit‐to‐stand, walking, or stepping practice.

Objectives

To summarise the evidence for the benefits and safety of mobility training on overall functioning and mobility in frail older people living in the community.

Search methods

We searched CENTRAL, MEDLINE, Embase, AMED, PEDro, US National Institutes of Health Ongoing Trials Register, and the World Health Organization International Clinical Trials Registry Platform (June 2021).

Selection criteria

We included randomised controlled trials (RCTs) evaluating the effects of mobility training on mobility and function in frail people aged 65+ years living in the community. We defined community as those residing either at home or in places that do not provide rehabilitative services or residential health‐related care, for example, retirement villages, sheltered housing, or hostels.

Data collection and analysis

We undertook an 'umbrella' comparison of all types of mobility training versus control.

Main results

This review included 12 RCTs, with 1317 participants, carried out in 9 countries. The median number of participants in the trials was 97. The mean age of the included participants was 82 years. The majority of trials had unclear or high risk of bias for one or more items. All trials compared mobility training with a control intervention (defined as one that is not thought to improve mobility, such as general health education, social visits, very gentle exercise, or "sham" exercise not expected to impact on mobility).

High‐certainty evidence showed that mobility training improves the level of mobility upon completion of the intervention period. The mean mobility score was 4.69 in the control group, and with mobility training, this score improved by 1.00 point (95% confidence interval (CI) 0.51 to 1.51) on the Short Physical Performance Battery (on a scale of 0 to 12; higher scores indicate better mobility levels) (12 studies, 1151 participants). This is a clinically significant change (minimum clinically important difference: 0.5 points; absolute improvement of 8% (4% higher to 13% higher); number needed to treat for an additional beneficial outcome (NNTB) 5 (95% CI 3.00 to 9.00)). This benefit was maintained at six months post‐intervention.

Moderate‐certainty evidence (downgraded for inconsistency) showed that mobility training likely improves the level of functioning upon completion of the intervention. The mean function score was 86.1 in the control group, and with mobility training, this score improved by 8.58 points (95% CI 3.00 to 14.30) on the Barthel Index (on a scale of 0 to 100; higher scores indicate better functioning levels) (9 studies, 916 participants) (absolute improvement of 9% (3% higher to 14% higher)). This result did not reach clinical significance (9.8 points). This benefit did not appear to be maintained six months after the intervention.

We are uncertain of the effect of mobility training on adverse events as we assessed the certainty of the evidence as very low (downgraded one level for imprecision and two levels for bias). The number of events was 771 per 1000 in the control group and 562 per 1000 in the group with mobility training (risk ratio (RR) 0.74, 95% CI 0.63 to 0.88; 2 studies, 225 participants) (absolute difference of 19% fewer (9% fewer to 26% fewer)).

Mobility training may result in little to no difference in the number of people who are admitted to nursing care facilities at the end of the intervention period as the 95% confidence interval includes the possibility of both a reduced and increased number of admissions to nursing care facilities (low‐certainty evidence, downgraded for imprecision and bias). The number of events was 248 per 1000 in the control group and 208 per 1000 in the group with mobility training (RR 0.84, 95% CI 0.53 to 1.34; 1 study, 241 participants) (absolute difference of 4% fewer (8% more to 12% fewer)).

Mobility training may result in little to no difference in the number of people who fall as the 95% confidence interval includes the possibility of both a reduced and increased number of fallers (low‐certainty evidence, downgraded for imprecision and study design limitations). The number of events was 573 per 1000 in the control group and 584 per 1000 in the group with mobility training (RR 1.02, 95% CI 0.87 to 1.20; 2 studies, 425 participants) (absolute improvement of 1% (12% more to 7% fewer)).

Mobility training probably results in little to no difference in the death rate at the end of the intervention period as the 95% confidence interval includes the possibility of both a reduced and increased death rate (moderate‐certainty evidence, downgraded for bias). The number of events was 51 per 1000 in the control group and 59 per 1000 in the group with mobility training (RR 1.16, 95% CI 0.64 to 2.10; 6 studies, 747 participants) (absolute improvement of 1% (6% more to 2% fewer)).

Authors' conclusions

The data in the review supports the use of mobility training for improving mobility in a frail community‐dwelling older population. High‐certainty evidence shows that compared to control, mobility training improves the level of mobility, and moderate‐certainty evidence shows it may improve the level of functioning in frail community‐dwelling older people. There is moderate‐certainty evidence that the improvement in mobility continues six months post‐intervention. Mobility training may make little to no difference to the number of people who fall or are admitted to nursing care facilities, or to the death rate. We are unsure of the effect on adverse events as the certainty of evidence was very low.

Keywords: Aged; Aged, 80 and over; Humans; Exercise; Exercise Therapy; Exercise Therapy/methods; Frailty; Independent Living; Quality of Life

Plain language summary

Mobility training for increasing mobility and functioning in older people who are frail

Background

Frailty is common in older people. Frailty leads to an increased likelihood of falling, difficulty moving (e.g. walking) and functioning, admissions to hospital, and mortality. It is estimated that 21% of the community‐dwelling population over 65 years are frail. Mobility training involves controlled movements of your body to perform specific tasks. Examples of mobility training included practicing standing up and sitting down, walking along a walking track, or going up and down stairs. Mobility training can be used when people have difficulties performing these tasks.

Study characteristics

This Cochrane Review is current to June 2021 and includes 12 studies with a total of 1317 participants. The studies were conducted in nine countries. The average participant age in the included studies was 82 years old; 73% of the participants were women. Six trials reported funding by government and research institutions, and one study reported funding from a commercial advocacy group.

Key results

‐ Mobility improved by 8% (4% higher to 13% higher) upon completion of the training period (12 studies, 1151 participants). People with no mobility training scored 4.69 points (out of 12 on the Short Physical Performance Battery scale; a higher score indicates better mobility). People with mobility training scored 5.69 points.

‐ Function improved by 9% (3% higher to 14% higher) upon completion of the training period (9 studies, 916 participants). People with no mobility training scored 86.1 points (out of 100 on the Barthel Index; a higher score indicates better functioning). People with mobility training scored 94.68 points.

‐ Unwanted or harmful effects of the training decreased by 19% (9% fewer to 26% fewer) (2 studies, 225 participants). If 1000 people were followed over 1 year, 771 people with no mobility training would experience unwanted or harmful effects, whereas 562 people with mobility training would experience unwanted or harmful effects.

‐ Admissions to nursing care facilities decreased by 4% (8% more to 12% fewer) (1 study, 241 participants). If 1000 people were followed over 1 year, 248 people with no mobility training would have an admission to a nursing care facility, whereas 208 people with mobility training would have an admission to a nursing care facility.

‐ Falls increased by 1% (12% more to 7% fewer) (2 studies, 425 participants). If 1000 people were followed over 1 year, 573 people with no mobility training would have a fall, whereas 584 people with mobility training would have a fall.

‐ Death rate increased by 1% (6% more to 2% fewer) (6 studies, 747 participants). If 1000 people were followed over 1 year, 51 people with no mobility training would die, whereas 59 people with mobility training would die.

Certainty of the evidence

In people with frailty, high‐certainty evidence shows mobility training improves mobility. Moderate‐certainty evidence shows mobility training is likely to improve function. Mobility training may result in little to no difference to the number of admissions to nursing care facilities (low‐certainty evidence), number of people who fall (low‐certainty evidence), and the death rate (moderate‐certainty evidence). Unwanted or harmful effects of the training were not well reported, and where reported, the overall evidence was of very low certainty.

Summary of findings

Summary of findings 1. Mobility training compared to control.

Intervention targeting mobility compared with control for community‐dwelling older people with frailty
Patient or population: older people living in the community with frailty Settings: community, either at home or in places of residence that, on the whole, do not provide residential health‐related care Intervention: intervention that aims to improve mobility Comparison: usual care (no change in usual activities) or a control (non‐active) intervention
Outcomes	*Anticipated absolute effects (95% CI)**		Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments^c
Outcomes	Control	Mobility training	Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments^c
Mobility (SPPB from 0 to 12; a higher score indicates better mobility.) Follow‐up: range 3 to 52 weeks	The mean mobility score was 4.69.^a	The mean mobility score was 1.00 point higher (0.51 higher to 1.51 higher).^b,c	‐	1151 (12)	⊕⊕⊕⊕ High^d	Mobility training increases mobility levels: SMD 0.47 higher (0.24 higher to 0.71 higher); absolute difference 8% higher (4% higher to 13% higher); relative difference 17% higher (9% higher to 26% higher).^e Clinically significant change. NNTB = 5 (95% CI 3.00 to 9.00)^f
Function (Barthel Index from 0 to 100; a higher score indicates better functioning.) Follow‐up: range 3 to 52 weeks	The mean function score was 86.1.^a	The mean function score was 8.58 points higher (3 higher to 14.3 higher).^b,g	‐	916 (9)	⊕⊕⊕⊝ Moderate^h	Mobility training likely increases function levels: SMD 0.60 better (0.21 better to 1.00 better); absolute difference 9% higher (3% higher to 14% higher); relative difference 9% higher (3% higher to 15% higher).^e Not clinically significant
Adverse events Follow‐up: range 36 to 52 weeks	771 per 1000	562 per 1000 (448 to 625)	0.74 (0.63 to 0.88)	225 (2)	⊕⊝⊝⊝ Very low^i,j,k	The evidence is very uncertain about the effect of mobility training on adverse events: absolute difference 19% fewer (9% fewer to 26% fewer); relative difference 26% fewer (12% fewer to 37% fewer).
Number of people who experienced 1 or more admissions to a nursing care facility Follow‐up: 52 weeks	248 per 1000	208 per 1000 (131 to 332)	0.84 (0.53 to 1.34)	241 (1)	⊕⊕⊝⊝ Low^i,j	Mobility training may result in little to no difference in the number of people who experience admissions to a nursing care facility: absolute difference 4% fewer (8% more to 12% fewer); relative difference 16% fewer (34% more to 47% fewer).
Number of fallers Follow‐up: 52 weeks	573 per 1000	584 per 1000 (498 to 687)	1.02 (0.87 to 1.20)	425 (2)	⊕⊕⊝⊝ Low^i,j	Mobility training may result in little to no difference in the number of people who experience falls;:absolute difference 1% more (12% more to 7% fewer); relative difference 2% more (20% more to 13% fewer).
Death Follow‐up: range 3 to 52 weeks	51 per 1000	59 per 1000 (33 to 107)	1.16 (0.64 to 2.10)	747 (6)	⊕⊕⊕⊝ Moderateⁱ	Mobility training probably results in little to no difference to the death rate: absolute difference 1% higher (6% more to 2% fewer); relative difference 16% more (110% more to 36% fewer).
The risk in the intervention group* (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; SD: standard deviation; SMD: standardised mean difference; SPPB: Short Physical Performance Battery; NNTB: number needed to treat for an additional beneficial outcome
GRADE Working Group grades of evidence High certainty: Further research is very unlikely to change our confidence in the estimate of effect. Moderate certainty: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low certainty: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low certainty: We are very uncertain about the estimate.

Open in a new tab

^aMean values for mobility (4.69) and function (86.1) were estimated from the follow‐up control group from Fairhall 2012. ^bSMD back translated to typical scales by multiplying SMD by the standard deviation at baseline in the control group as reported in Fairhall 2012: mean (SD) for SPPB (0 to 12), 5.74 (2.12); mean (SD) for Barthel Index (0 to 100), 92.5 (14.3) ^cSensitivity analysis found little difference when the following were removed: studies that only provided home‐based programmes (SMD 0.50, 95% CI 0.20 to 0.81; participants = 766; studies = 10; I² = 76%); studies with the longest and shortest intervention period (SMD 0.46, 95% CI 0.15 to 0.78; participants = 727; studies = 9; I² = 76%); and studies with the largest and smallest dose of intervention per week (SMD 0.41, 95% CI 0.08 to 0.74; participants = 645; studies = 8; I² = 75%). ^dWhen we undertook the GRADE assessment, we did not downgrade for risk of bias as results were essentially unchanged with the removal of the trials with a high risk of bias on one or more items or with the removal of studies with high or unclear risk for selection bias or detection bias. We did not downgrade for inconsistency (heterogeneity was not considerable). ^eRelative changes calculated relative to baseline in control group (i.e. absolute change (mean difference) divided by mean at baseline for mobility (5.74) and function (92.5) in the control group from Fairhall 2012) ^fCalculations based on the control group baseline mean (SD) SPPB: 5.74 (2.12) points on 0 to 12 scale (from Fairhall 2012) and an assumed minimal clinically important difference of 0.5 points ^gSensitivity analysis found little difference when the following were removed: studies that only provided home‐based programmes (SMD 0.75, 95% CI 0.20 to 1.29; participants = 533; studies = 7; I² = 88%); studies with the longest and shortest intervention period (SMD 0.67, 95% CI 0.09 to 1.25; participants = 494; studies = 9; I² = 89%); and studies with the largest and smallest dose of intervention per week (SMD 0.64, 95% CI ‐0.03 to 1.30; participants = 422; studies = 5; I² = 89%). ^hDowngraded due to inconsistency of results. Function (I² = 87%) ⁱDowngraded due to risk of bias ^jDowngraded due to imprecision ^kDowngraded additional level due to risk of bias

Summary of findings 2. Mobility training compared to control: 6 months post‐intervention.

Patient or population: older people living in the community with frailty Settings: community, either at home or in places of residence that, on the whole, do not provide residential health‐related care Intervention: intervention that aims to improve mobility Comparison: usual care (no change in usual activities) or a control (non‐active) intervention
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
	Control	Mobility training
Intervention targeting mobility compared with control for community‐dwelling older people with frailty (studies including participants with cognitive impairment)
Mobility (A higher score indicates better mobility.)	The mean mobility score was 4.69.^a	The mean mobility score was 0.72 points higher (0.21 higher to 1.14 higher).^b	‐	451 (5)	⊕⊕⊕⊝ Moderate^c	Mobility training probably increases mobility levels at 6 months post‐intervention: SMD 0.32 higher (0.10 higher to 0.54 higher); absolute difference 6% higher (2% higher to 10% higher); relative difference 13% higher (4% higher to 20% higher).^d NNTB = 7 (95% CI 4.00 to 22)^e
Function (A higher score indicates better functioning.)	The mean function score was 86.1.^a	The mean function score was 18.45 points higher (5.43 worse to 42.33 higher).^b	‐	278 (3)	⊕⊕⊝⊝ Low^c,e	Mobility training may result in little to no difference in function levels at 6 months post‐intervention: SMD 1.29 better (0.38 worse to 2.96 better); absolute difference 18% higher (5% lower to 42% higher); relative difference 20% higher (6% lower to 46% higher).^d
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; SD: standard deviation; SMD: standardised mean difference; SPPB: Short Physical Performance Battery; NNTB: number needed to treat for an additional beneficial outcome
GRADE Working Group grades of evidence High certainty: Further research is very unlikely to change our confidence in the estimate of effect. Moderate certainty: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low certainty: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low certainty: We are very uncertain about the estimate.

Open in a new tab

Summary of findings 3. Mobility training compared to control: subgroups with or without cognitive impairments.

Patient or population: older people living in the community with frailty Settings: community, either at home or in places of residence that, on the whole, do not provide residential health‐related care Intervention: intervention that aims to improve mobility Comparison: usual care (no change in usual activities) or a control (non‐active) intervention
Outcomes	*Illustrative comparative risks (95% CI)**		Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
Outcomes	Control	Mobility training	Relative effect (95% CI)	Number of participants (studies)	Certainty of the evidence (GRADE)	Comments
Intervention targeting mobility compared with control for community‐dwelling older people with frailty (studies including participants with cognitive impairment)
Mobility (A higher score indicates better mobility.) Follow‐up: range 3 to 52 weeks	The mean mobility score was 4.69.^a	The mean mobility score was 1.93 points higher (0.09 higher to 3.48 higher).^b	‐	171 (4)	⊕⊕⊝⊝ Low^c,d	Mobility training may increase mobility levels: SMD 0.91 higher (0.19 higher to 1.64 higher); absolute difference 16% higher (1% higher to 29% higher); relative difference 34% higher (2% higher to 61% higher)^e NNTB = 3 (95% CI 2.00 to 11.00)^f
Function (A higher score indicates better functioning.) Follow‐up: range 3 to 52 weeks	The mean function score was 86.1.^a	The mean function score was 41.76 points higher (3.15 worse to 86.80 higher).^b	‐	80 (2)	⊕⊝⊝⊝ Very low^c,d,g	The evidence is very uncertain about the effect of mobility training on function: SMD 2.92 better (0.22 worse to 6.07 better); absolute difference 42% higher (3% lower to 87% higher); relative difference 45% higher (3% lower to 94% higher).^e
Intervention targeting mobility compared with control for community‐dwelling older people with frailty (studies excluding participants with cognitive impairment)
Mobility (A higher score indicates better mobility.) Follow‐up: range 12 to 52 weeks	The mean mobility score was 4.69.^a	The mean mobility score was 0.78 points higher (0.30 higher to 1.27 higher).^b	‐	967 (8)	⊕⊕⊕⊕ High^h	Mobility increases mobility levels: SMD 0.37 higher (0.14 higher to 0.60 higher); absolute difference 7% higher (3% higher to 11% higher); relative difference 14% higher (5% higher to 22% higher).^e NNTB = 6 (95% CI 4.00 to 16.00)^f
Function (A higher score indicates better functioning.) Follow‐up: range 12 to 52 weeks	The mean function score was 86.1.^a	The mean function score was 4.43 points higher (1.00 higher to 8.00 higher).^b	‐	839 (7)	⊕⊕⊕⊝ Moderate^d	Mobility training probably increases function levels: SMD 0.31 better (0.07 better to 0.56 better); absolute difference 4% higher (1% higher to 8% higher); relative difference 5% higher (1% higher to 9% higher).^e
The basis for the assumed risk* (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; SD: standard deviation; SMD: standardised mean difference; SPPB: Short Physical Performance Battery; NNTB: number needed to treat for an additional beneficial outcome
GRADE Working Group grades of evidence High certainty: Further research is very unlikely to change our confidence in the estimate of effect. Moderate certainty: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low certainty: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low certainty: We are very uncertain about the estimate.

Open in a new tab

Background

Description of the condition

Frailty is a common syndrome in older people; it is characterised by decline across multiple body systems, causing decreased physiological reserve and increased vulnerability to adverse health outcomes (Fried 2004). It is estimated that 21% of the community‐dwelling population over 65 years are frail, and a further 48% are pre‐frail (Thompson 2018). The percentage of older adults with frailty is increasing (Ofori‐Asenso 2019). Frailty is independently predictive of falls, worsening mobility, deteriorating functioning, impaired activities of daily living, and death (Fried 2004; Pel‐Littel 2009). Frailty is therefore costly to the individual, their families, and society. The economic and healthcare burden will increase as the population ages (Howe 1997).

Commonly associated with ageing, frailty is not a requisite of the ageing process, with many people reaching older ages without becoming frail (Clegg 2013). People can transition between frailty state over time, and there is potential for people with frailty to improve or reverse their levels of frailty (Gill 2006).

Although there is no agreed method for diagnosing frailty, the most commonly used methods include the physical phenotype (i.e. Fried’s Frailty Phenotype (Fried 2001)) and accumulated deficit models (i.e. Frailty Index (Mitnitski 2001)). Other common diagnostic tools include the Frail Scale (Woo 2012), Clinical Frailty Index (Rockwood 2005), and Edmonton Frail Scale (Rolfson 2006). Individual factors underlying frailty (i.e. walking speed and grip strength) can also be used to screen for frailty (Dent 2017; Woolford 2020).

Mobility is a broad term, which is defined as the ability to move around and change positions, such as walking, rising from a chair, and maintaining balance while standing (WHO 2001). Reducing the decline of mobility and functioning should be considered a key goal of interventions for frail older people. Many hospital admissions in frail older people relate to decreased mobility levels (Keeble 2019). High rates of mobility and functional decline are evident in frail community‐dwelling older people (Gill 1995). Interventions that slow the decline of mobility and functioning in the frail population will impact upon morbidity and mortality (Gill 2004).

The International Conference of Frailty and Sarcopenia research guidelines recommend that multicomponent physical activity should be prescribed for all older adults with frailty. However, there was insufficient evidence to note if a specific type of exercise is more effective in a frail population (Dent 2019).

In the International Classification of Functioning, Disability and Health (ICF) (WHO 2001), 'functioning' refers to all body functions, activities, and participation. The evidence to guide interventions for increasing functioning in frail older people is both sparse and conflicting. Many trials systematically exclude frail older adults, largely due to their comorbidities and difficulties associated with their recruitment (Ridda 2008).

Description of the intervention

We examined the effect of interventions that aimed to improve mobility. In accordance with the ICF definition of mobility in the activity and participation domains, in this review, we define mobility as changing and maintaining a body position, walking, and moving. Common interventions used to increase mobility include task‐specific mobility training and balance‐challenging exercises. This includes whole task practice, such as sit‐to‐stand; walking; or part practice, such as stepping practice. In addition, we included multifactorial interventions if the majority of the intervention aimed to improve mobility, and we included interventions that involved performance of mobility tasks if participants' performance was monitored and reviewed during the period of intervention delivery.

How the intervention might work

Studies have shown that frail older people can improve their degree of frailty (Gill 2006), functioning, and health, and the benefits of exercise on health status in community‐dwelling frail older people may be greatest in those who are more frail at baseline (Hubbard 2009). Although some interventions delivering exercise to frail older people have demonstrated improved mobility (Fiatarone 1994), physical functioning (Binder 2002; Worm 2001), and reduced rate of functional decline (Gill 2002), other trials have found no effect on functioning or disability (Chin 2001; Latham 2003; Rydwik 2004), and inconsistent effects on falls outcomes (Clegg 2014; Faber 2006; Ng 2014). A systematic review of interventions targeting disability in community‐dwelling frail older people indicated that the effect on performance of activities of daily living was inconclusive (Daniels 2008). It is unclear which parameters of a mobility training intervention (e.g. dose, degree of difficulty) are necessary to improve mobility and functioning in frail older people. In addition, the safety of such interventions is unclear, with reports of increased musculoskeletal injury in frail older people (Latham 2003).

Cochrane Reviews that assessed studies of exercise delivered to older people have implications for the frail older population. Recent Cochrane Reviews indicate that exercise interventions can improve strength and balance and prevent falls in older people (Cameron 2018; Hopewell 2020; Howe 2011; Liu 2009; Sherrington 2019; Tricco 2017). These outcomes have potential benefits for mobility and functioning; however, the impact of exercise on mobility and functioning is less clear (Ashworth 2005; De Morton 2007).

This Cochrane Review differs from other Cochrane Reviews that have examined the effects of various exercise interventions in older people. While a proportion of participants in previous Cochrane Reviews were frail (Ashworth 2005; Cameron 2018; De Morton 2007; Howe 2011; Liu 2009; Sherrington 2019), our review is restricted to frail older people. We focused on frail older people living in the community, rather than the hospital setting studied by de Morton and colleagues (De Morton 2007). This review builds upon previous Cochrane Reviews that evaluated the effect of exercise intervention on measures of functioning (Howe 2011; Liu 2009), by including a broader range of interventions that may affect mobility and functioning outcomes. Finally, in addition to mobility, this review also includes function as a major outcome, as opposed to specific outcomes such as balance and falls (Cameron 2018; Howe 2011; Sherrington 2019).

Why it is important to do this review

This Cochrane Review has important clinical implications for the frail population. An intervention that increases mobility and functioning in this population has the potential to reduce hospitalisation and admission to residential care facilities, as well as costs to the individual, government, and society.

A systematic review is needed to identify the trials in this field, and a meta‐analysis is required to synthesise the evidence for researchers, professionals, policymakers, and others with an interest in this important and evolving area.

Objectives

To summarise the evidence for the benefits and safety of mobility training on overall functioning and mobility in frail older people living in the community.

Methods

Criteria for considering studies for this review

Types of studies

We included randomised controlled trials and controlled clinical trials that used quasi‐randomisation methods (e.g. allocation by date of birth or alternation).

Types of participants

Eligible trials included men and women aged 60 and older who lived in the community and were described as being frail.

We only included trials that described the participants as frail and justified the use of the term 'frail' (for example, using the Cardiovascular Health Study criteria (Fried 2001) or the Frailty Index (Mitnitski 2004), when participants had difficulty performing specified tasks, or using other measures of vulnerability).

We only included trials that described participants as older adults, aged, seniors, geriatric, elderly, or where participants were all over the age of 60 and their mean or median age was over 65 years.

We included trials where the majority of participants lived, or planned to live, in the community. We defined the community as residing either at home or in places that do not provide rehabilitative services or residential health‐related care, for example, we included retirement villages, sheltered housing, or hostels. In the event that a trial included participants living both in the community and in higher‐dependency residences, for example, nursing care facilities or hospitals, we included participants in this review only if data were provided for the subgroup living in the community.

The review focused on trials delivering interventions to frail people with a wide range of comorbidities. We excluded trials of a single diagnostic group (as listed in the International Statistical Classification of Diseases and Related Health Problems 10th Revision (WHO 2010)).

Types of interventions

To be eligible, trial interventions must have targeted improvement in mobility, as evidenced by measurement of mobility as an outcome. We defined mobility according to the International Classification of Functioning, Disability and Health (ICF) (WHO 2001): changing and maintaining a body position, walking, and moving. Interventions used to improve mobility commonly include task‐specific mobility training and balance‐challenging exercises, which are prescribed based upon individualised assessment and ongoing reassessment.

If an intervention contained multiple elements, we only included trials where there was an aim (explicit or implicit) to improve mobility for the majority of the intervention. Training that involved performance of a task must have had a monitored component, with review of participants' performance during the period of intervention delivery. The mobility training intervention may have been augmented by cognitive interventions, such as goal setting, education, and motivational interviewing.

We included trials where the intervention was compared with a control group that received sham exercise (the exercise appeared to be of insufficient intensity and progression to have beneficial effects on balance and mobility), no intervention, usual care, or a social visit. We included trials comparing two or more interventions if the difference between the intervention and control groups was exercise.

We grouped the intervention arms by exercise modality into six categories (Appendix 1), using the Prevention of Falls Network Europe (ProFaNE) taxonomy (Lamb 2007; Lamb 2011) (Appendix 2).

Types of outcome measures

Major outcomes

Mobility
Functioning
Adverse effects of the interventions
Admission to a nursing care facility
Falls
Death

The main outcome of interest was mobility, defined in the ICF as changing and maintaining a body position, walking, and moving (WHO 2001). We assessed mobility using physical performance measures.

The second main outcome of interest was functioning, measured in terms of the activity or participation levels of the ICF (WHO 2001). We assessed functioning as a continuous variable where possible.

We excluded trials if they reported outcomes at the 'body structure/function' level of the ICF (such as proprioception or strength), without outcomes at the level of activity or participation. Balance as an outcome was treated as an activity and therefore included. We classified instruments containing multiple ICF concepts, for example, the EuroQol‐5D (EQ‐5D) and 12‐item Short Form (SF‐12), as health‐status instruments and did not consider these as measures of functioning.

As trials measured mobility and functioning using different instruments, in our meta‐analysis, we used standardised mean difference to facilitate quantitative pooling. To enable inclusion of as many trials as possible in the meta‐analysis, we used the method reported by Allen 2011 to pool results across multiple outcomes. For trials that reported results for more than one outcome measure, we used the highest priority mobility outcome and the highest priority functioning measure in the analysis. The highest priority mobility measure is that which most reflects the components of everyday mobility. The highest priority functioning measure is that which encompasses the most activities and participation domains of the ICF. The analysis involved the pooling of the highest priority mobility measure from each trial and pooling of the most comprehensive functioning measure from each trial.

The order of priority is as follows.

Mobility: Short Physical Performance Battery, Timed Up and Go test, walking distance, walking speed, timed sit to stand, Berg Balance Scale, single leg stand time.
Functioning: Barthel Index, Physical Performance Test, Lawton Independent Activities of Daily Living Index, Functional Independence Measure, Performance Sum Score, Activity of Daily Living (ADL) score, Instrumental Activities of Daily Living.

Minor outcomes

Costs

Timing of outcome assessment

The primary time period of outcome assessment was the completion of the intervention programme. We also planned to extract outcomes up to 6 months and up to 12 months post‐completion of the intervention programme. As no studies reported outcome measures post‐six months, we only extracted outcomes up to six months.

Search methods for identification of studies

Electronic searches

We searched the following:

Cochrane Central Register of Controlled Trials (CENTRAL; 2021, Issue 6) in the Cochrane Library (searched 1 June 2021);
MEDLINE Ovid (1946 to 1 June 2021);
Embase Ovid (1947 to 1 June 2021);
AMED Ovid (Allied and Complementary Medicine; 1985 to 1 June 2021);
PEDro (Physiotherapy Evidence Database; 1999 to 1 June 2021);
US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov (www.clinicaltrials.gov; searched 1 June 2021); and
World Health Organization International Clinical Trials Registry Platform (apps.who.int/trialsearch; searched 1 June 2021).

There were no language restrictions. The search strategy is in Appendix 3.

Searching other resources

We checked the reference lists of all included studies and other systematic reviews.

Data collection and analysis

Selection of studies

One review author (DT) screened the title, abstract, and descriptors of all trials that potentially met the inclusion criteria. From the full text, three review authors (DT, NF, KS) independently assessed potentially eligible trials for inclusion and resolved any disagreement through discussion. We contacted study authors for additional information as necessary.

Data extraction and management

Three review authors (DT, NF, KS) independently extracted data using a pre‐tested data extraction form. We resolved disagreements by consensus or third‐party adjudication. We were not blinded to authors and sources.

We used a standardised data extraction form to record the following items.

Methods: study design, total duration of study, details of any 'run in' period, number of study centres and locations, study setting, withdrawals, and date of study.
Participants: sample size, mean age, age range, sex, disease duration, severity of condition, diagnostic criteria, important baseline data describing degree of frailty, inclusion criteria, and exclusion criteria.
Interventions: type of intervention, content of intervention, dose of intervention (intensity, frequency, duration, progression), supervision (self, individual, groups), supervisor (self, peer, healthcare professional, exercise specialist), number of supervisors, setting (home, community or gym, healthcare provider), and adherence rate.
Comparison outcomes: primary and secondary outcomes specified and collected, and time points reported. Number of events and number of participants per treatment group for dichotomous outcomes. Means and standard deviations and number of participants per treatment group for continuous outcomes.
Characteristics of the design of the trial as outlined in Assessment of risk of bias in included studies.
Notes: funding for trial and notable declarations of interest of trial authors.
The primary time period of outcome assessment was the completion of the intervention programme. We collected time periods less than 12 months after the intervention ceased.

Assessment of risk of bias in included studies

Three review authors (DT, NF, KS) independently assessed risk of bias using RoB 1 as described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2020). Review authors were not blinded to authors and sources. We resolved disagreement by consensus or third‐party adjudication. We assessed the risk of bias according to the following domains.

Random sequence generation.
Allocation concealment.
Blinding of participants and personnel.
Blinding of outcome assessment (subjective outcomes).
Blinding of outcome assessment for assessor‐reported outcomes (mobility and function).
Incomplete outcome data.
Selective reporting.
Other bias

We rated risk of bias as either low, high, or unclear for each domain and provided a quote from the study report together with a justification for our judgment in each risk of bias table. We summarised the risk of bias judgements across different studies for each of the domains listed. We considered blinding separately for different key outcomes where necessary (e.g. for unblinded outcome assessment, risk of bias for all‐cause mortality may be different than for a patient‐reported pain scale), and we considered the impact of missing data by key outcomes. Where information on risk of bias relates to unpublished data or correspondence with a trialist, we noted this in the risk of bias table. We presented the figures generated by RoB 1 to provide summary assessments of the risk of bias.

Measures of treatment effect

Data relating to mobility

All mobility scales used were continuous measures. We used the standardised mean difference (SMD) as a summary statistic for the pooled intervention effect as different outcome measures were reported across trials. Where available, we used the means and standard deviations of change scores and their respective standard deviations. We used the standard deviation of change score when we could reliably obtain this from other statistics (such as the confidence interval (CI) or P value). We used unadjusted values in the meta‐analysis. Where scales decreased with increasing function, we multiplied the mean values by ‐1 to correct for differences in the directions of scales. We combined studies in the meta‐analysis to calculate the SMD, which we then back‐transformed to clinically meaningful units using the Short Physical Performance Battery (SPPB). We chose the SPPB as it is widely used in practice and has been shown to be predictive of an increased risk of falling, loss of independence in activities of daily living, decreased mobility, disability, decline in health, re‐hospitalisation, increased hospital length of stay, nursing home admission, and death (Treacy 2018). As no representative observational study was identified using this scale, we calculated the mean difference by multiplying the SMD by the control group baseline standard deviation (SD) from the most representative trial (Higgins 2020a). We identified Fairhall 2012 as the most representative trial as it was the largest trial; we used the SD from this trial's control group.

A difference of 0.5 points on the SPPB is considered a clinically significant difference (Perera 2006).

Data relating to functioning

All function scales used were continuous measures. We used the SMD to summarise the size and direction of the effect (using the method described above), then back‐transformed the SMD to clinically meaningful units (Barthel Index). We chose the Barthel Index as it is widely used in practice. As no representative observational study was identified using this scale, we calculated the mean difference by multiplying the SMD by the control group baseline SD from the most representative trial (Higgins 2020a). We identified Fairhall 2012 as the most representative trial as it was the largest trial; we used the SD from this trial's control group.

A difference of 9.8 points on the Barthel Index is considered a clinically significant difference (Unnanuntana 2018).

Data relating to death, falls, adverse outcomes, and movement to nursing care facilities

We treated these outcomes as dichotomous. For each trial with data available, we calculated the risk ratio (a comparison of the number of participants in each group with one or more event) and 95% CI using the generic inverse variance option. We used unadjusted values in the meta‐analysis; we entered the number of participants contributing data to each group in the analysis. In addition, we also measured the rate of falls using a rate ratio (for example, incidence rate ratio or hazard ratio for all falls) with 95% CI if the paper reported this. If studies reported both adjusted and unadjusted rate ratios, we used the unadjusted estimate unless the adjustment was for clustering.

In Effects of interventions and the 'Comments' column of the summary of findings tables, we provided the absolute per cent difference, the relative per cent change from baseline, and the number needed to treat for an additional beneficial outcome (NNTB) or the number needed to treat for an additional harmful outcome (NNTH) (we provided the NNTB or NNTH only when the outcome showed a clinically significant difference).

For dichotomous outcomes, we calculated the absolute per cent change from the difference in risks between the intervention and control group using GRADEpro GDT and expressed as a percentage (GRADEpro GDT). For continuous outcomes, we calculated the absolute benefit as the improvement in the intervention group minus the improvement in the control group, in the original units, expressed as a percentage. We calculated the relative per cent change for dichotomous data as the risk ratio ‐ 1 and expressed as a percentage. For continuous outcomes, we calculated the relative difference in the change from baseline as the absolute benefit divided by the baseline mean of the control group, expressed as a percentage. For dichotomous outcomes, we calculated the NNTB or NNTH from the control group event rate and the relative risk using the Visual Rx NNT calculator (Cates 2008). We calculated the NNTB or NNTH for continuous measures using the Wells calculator (available at the Cochrane Musculoskeletal editorial office).

Unit of analysis issues

Where possible, we used outcomes from all randomised participants in the meta‐analysis. Where intention‐to‐treat results were not reported, we used per‐protocol results.

In the event of multiple intervention groups, we combined the groups relevant to the review and containing similar interventions to create a single pair‐wise comparison.

Dealing with missing data

We only included in the analysis data from participants whose outcomes were available. We addressed the potential influence of missing data in the risk of bias assessment.

If we could reliably obtain the standard deviation of change score from other statistics (such as CI or P value), we did this. Where the study sample size exceeded 70 and the study reported the median and interquartile range, we used the median as an approximation of the mean and estimate of the interquartile range as 1.35 standard deviations (Deeks 2020; Hozo 2005).

Assessment of heterogeneity

For studies that we judged as clinically homogeneous, we assessed and quantified the possible magnitude of inconsistency (i.e. heterogeneity) across studies. We assessed statistical heterogeneity using the I² statistic (Deeks 2020). We interpreted the I² statistic using the following as an approximate guide.

0% to 40% might not be important.
30% to 60% may represent moderate heterogeneity.
50% to 90% may represent substantial heterogeneity.
75% to 100% considerable heterogeneity.

Assessment of reporting biases

We reported bias using the Egger's test and visual inspection of the symmetry of a funnel plot: we constructed and visually inspected funnel plots for outcomes that included more than 10 data points.

To assess the possible presence of small sample bias in the published literature (i.e. in which the intervention effect is more beneficial in smaller studies), we performed meta‐analysis using random‐effects modelling. We undertook sensitivity analysis to explore whether fixed‐effect modelling gave similar results. In the presence of small sample bias, the random‐effects estimate of the intervention is more appropriate to use than the fixed‐effect estimate (Page 2020).

Data synthesis

We pooled the results of studies with comparable participant characteristics and interventions using the generic inverse variance method in Review Manager 5. We used the random‐effects model to calculate SMD with 95% CI.

Following analysis, we converted outcomes back to clinically meaningful units, such as walking speed and minimum chair height from which a person can stand up. We multiplied the SMD by a typical among‐person standard deviation for each particular scale (derived from a representative observational study). Where no representative observational study existed, we used the control group baseline SD from the most representative trial (Deeks 2020).

Subgroup analysis and investigation of heterogeneity

We performed subgroup analyses for the mobility and function outcomes to compare effects in trials that excluded participants with cognitive impairment versus studies that included participants with cognitive impairments. People with cognitive impairment, such as dementia, are not routinely offered rehabilitation services and are frequently excluded from studies (Cations 2017; Laver 2020).

We used the test for subgroup differences available in RevMan 5.4 to determine whether there was evidence for a difference in treatment effect between subgroups. We used caution in the interpretation of subgroup analyses as advised in Chapter 10 of the Cochrane Handbook for Systematic Reviews of Interventions (Deeks 2020).

Sensitivity analysis

We carried out sensitivity analyses to examine the robustness of our results. To explore the possible impact of risk of bias on the primary pooled estimates of treatment effect, we examined the effects of removing the following.

Trials with a high risk of bias in any item.
Trials at high or unclear risk of selection bias.
Trials at high or unclear risk of detection bias.

To explore the possible impact of intervention delivery strategies on the primary pooled estimates of treatment effect, we examined the effects of removing the following.

Studies that only provided home‐based programmes (i.e. did not provide centre‐based programmes).
Studies with the longest and shortest intervention period.
Studies with the largest and smallest dose of intervention per week (centre‐based programmes only).

Summary of findings and assessment of the certainty of the evidence

We presented our six major outcomes (mobility, function, adverse effects, admission to nursing care facilities, falls, death) in summary of findings tables (SOFs) for the primary comparison mobility training versus control. SOFs summarise the certainty of evidence, the magnitude of effect of the interventions examined, and the sum of available data on the outcomes as recommended by Cochrane (Schünemann 2020). The summary of findings table includes an overall grading of the evidence related to each of the main outcomes, using the GRADE approach. Our primary time point for the summary of findings tables was the completion of the intervention programme.

Three people (DT, NF, KS) independently assessed the certainty of the evidence. We used the five GRADE considerations (study limitations, consistency of effect, imprecision, indirectness, and publication bias) to assess the certainty of a body of evidence as it related to the studies that contributed data to the meta‐analyses for the prespecified outcomes, and we reported the certainty of evidence as high, moderate, low, or very low. We used methods and recommendations described in Chapter 14 of the Cochrane Handbook for Systematic Reviews of Interventions (Schünemann 2020). We used GRADEpro GDT software to prepare the SoF tables (GRADEpro GDT). We justified all decisions to downgrade the certainty of studies using footnotes and made comments to aid the reader's understanding of the review where necessary. We provided the number needed to treat for an additional beneficial outcome (NNTB) or the number needed to treat for an additional harmful outcome (NNTH) absolute and relative per cent change in the 'Comments' column of the SoF tables as described in the Measures of treatment effect.

Interpreting results and reaching conclusions

We followed the guidelines in Chapter 15 of the Cochrane Handbook for Systematic Reviews of Interventions (Schünemann 2020a) for interpreting results and were aware of distinguishing a lack of evidence of effect from a lack of effect. We based our conclusions only on findings from the quantitative or narrative synthesis of included studies for this review. We avoided making recommendations for practice, and our implications for research suggests priorities for future research and outlines the remaining uncertainties in the area.

Results

Description of studies

Results of the search

We identified a total of 13,473 records from the following databases: Cochrane Central Register of Controlled Trials (4089), MEDLINE (4834), Embase (4176), AMED (95), and PEDro (279).

After removing duplicates, we screened 12,183 articles, excluded 12,154, and identified 29 full texts for consideration. We excluded a further 17 papers after careful review following the inclusion criteria. We included the remaining 12 studies in this review (see Figure 1 for a flow diagram of the search results).

Included studies

We included 12 trials, with 1317 participants. We provide details in the Characteristics of included studies and provide a brief summary below.

We contacted authors of eight included studies to request additional details regarding study design and outcome data, and we received responses from four studies; this resulted in additional information used in the review for Bonnefoy 2003; Chin A Paw 2002; Langlois 2013; and Rydwik 2008. The authors of four studies did not supply additional information (Binder 2002; Gill 2004; Mugueta‐Aguinaga 2017; Worm 2001).

Trial design

All included studies were randomised controlled trials. All trials were individually randomised. The included trials had a total of 32 groups. Most trials (n = eight) had two groups (usually intervention and control) (Binder 2002; Fairhall 2012; Gill 2004; Gine‐Garriga 2010; Langlois 2013; Mugueta‐Aguinaga 2017; Tarazona‐Santabalbina 2016; Worm 2001). The other studies (n = four) had four groups (Bonnefoy 2003; Chin A Paw 2002; Kim 2015; Rydwik 2008). Eleven of the 12 studies had a follow‐up length of 12 weeks or greater (Binder 2002; Bonnefoy 2003; Chin A Paw 2002; Fairhall 2012; Gill 2004; Gine‐Garriga 2010; Kim 2015; Langlois 2013; Rydwik 2008; Tarazona‐Santabalbina 2016; Worm 2001) (Table 4).

1. Study design, length of follow‐up, setting, and trial size.

Study ID	Study design	No. arms (clusters)	Length of follow‐up	Setting	No. randomised	No. analysed	% lost to follow‐up
Binder 2002	Parallel	2	9 months	USA	119	115	3%
Bonnefoy 2003	Parallel	4	9 months	France	57	42	26%
Chin A Paw 2002	Parallel	4	17 weeks	The Netherlands	217	157	27%
Fairhall 2012	Parallel	2	12 months	Australia	241	216	10%
Gill 2004	Parallel	2	12 months	USA	188	178	5%
Gine‐Garriga 2010	Parallel	2	36 weeks	Spain	51	41	20%
Kim 2015	Parallel	4	7 months	Japan	131	126	4%
Langlois 2013	Matched control	2	12 weeks	Canada	34	29	15%
Mugueta‐Aguinaga 2017	Parallel	2	6 weeks	Spain	40	39	3%
Rydwik 2008	Parallel	4	9 months	Sweden	93	62	33%
Tarazona‐Santabalbina 2016	Parallel	2	24 weeks	Spain	100	82	18%
Worm 2001	Parallel	2	12 weeks	Denmark	46	44	4%

Open in a new tab

Trial size and setting

The median number of participants randomised per trial was 134 (interquartile range: 49 to 145). Included trials ranged in sample size from 40 participants, Mugueta‐Aguinaga 2017, to 241 participants, Fairhall 2012. The included trials were conducted in nine countries: Spain (three trials: Gine‐Garriga 2010; Mugueta‐Aguinaga 2017; Tarazona‐Santabalbina 2016), USA (two trials: Binder 2002; Gill 2004), Australia (one trial: Fairhall 2012), Canada (one trial: Langlois 2013), Denmark (one trial: Worm 2001), France (one trial: Bonnefoy 2003), Japan (one trial: Kim 2015), the Netherlands (one trial: Chin A Paw 2002), and Sweden (one trial: Rydwik 2008) (Appendix 4).

Participants

There were a total of 1317 participants randomised and 1151 with data at follow‐up. Overall, 73% of the included participants were women. All participants were women in one trial (Kim 2015). On average, the mean age of participants in the included trials was 82 years (Appendix 4). The inclusion and exclusion criteria and other participant details are listed for each study in the Characteristics of included studies.

Eight trials excluded participants with cognitive impairment, either defined as an exclusion criterion or implied by the stated requirement to be able to give informed consent or follow instructions, or both (Binder 2002; Chin A Paw 2002; Fairhall 2012; Gill 2004; Kim 2015; Langlois 2013; Rydwik 2008; Tarazona‐Santabalbina 2016) (see Appendix 4).

Interventions

All trials compared mobility training with a control intervention (defined as one that is not thought to improve mobility, such as general health education (Gill 2004; Gine‐Garriga 2010; Rydwik 2008), social visits (Chin A Paw 2002), cognitive therapy (Bonnefoy 2003), no intervention (Fairhall 2012; Kim 2015; Langlois 2013; Mugueta‐Aguinaga 2017; Tarazona‐Santabalbina 2016; Worm 2001), or very gentle exercise not expected to impact on mobility (Binder 2002)). Four trials compared mobility training in addition to nutritional supplements to nutritional supplements alone (Bonnefoy 2003; Chin A Paw 2002; Kim 2015; Rydwik 2008).

The length of the intervention period varied from 6 weeks to 12 months; 11 of the 12 studies had an intervention period of between 3 and 12 months (Binder 2002; Bonnefoy 2003; Chin A Paw 2002; Fairhall 2012; Gill 2004; Gine‐Garriga 2010; Kim 2015; Langlois 2013; Rydwik 2008; Tarazona‐Santabalbina 2016; Worm 2001). Ten studies (83%) delivered the intervention in a community setting or gym (Binder 2002; Bonnefoy 2003; Chin A Paw 2002; Gine‐Garriga 2010; Kim 2015; Langlois 2013; Mugueta‐Aguinaga 2017; Rydwik 2008; Tarazona‐Santabalbina 2016; Worm 2001); the other two studies delivered the intervention in the home setting (Fairhall 2012; Gill 2004). The length of training for the studies that delivered the intervention in a community setting or a gym varied between 60 minutes to 325 minutes per week. Eight studies (67%) delivered the exercise in a group setting (Binder 2002; Chin A Paw 2002; Gine‐Garriga 2010; Kim 2015; Langlois 2013; Rydwik 2008; Tarazona‐Santabalbina 2016; Worm 2001); 1 study (8%) delivered the exercise individually (Mugueta‐Aguinaga 2017); 2 studies (17%) delivered the exercise as self‐monitored home exercise (Fairhall 2012; Gill 2004); and for one intervention arm, the method of delivery was not reported (Bonnefoy 2003). Four studies included home programmes (Fairhall 2012; Gill 2004; Rydwik 2008; Worm 2001) (Table 5).

2. Key characteristics of participants and intervention approach.

Study ID	Age (mean)	% Women	Dose of intervention	Length of intervention	Exercise setting	Supervision (self, individual, group)	Home exercise programme	Frequency of home programme	Home visits (number)
Binder 2002	83	52%	60 minutes, three times a week	9 months	Community or gym	Group	No	NA	No
Bonnefoy 2003	83	88%	60 minutes, three times a week	9 months	Community or gym	NR	No	NA	No
Chin A Paw 2002	79	71%	45 minutes, twice a week	17 weeks	Community or gym	Group	No	NA	No
Fairhall 2012	84	84%	Home exercise programme only	12 months	Home	Self	Yes	20 to 30 minutes, 3 to 5 times per week	Yes (up to 10)
Gill 2004	83	80%	Home exercise programme only	6 months	Home	Self	Yes	10 to 30 minutes, daily	Yes (averaged 15 over the 6 months)
Gine‐Garriga 2010	84	61%	45 minutes, twice a week	12 weeks	Community or gym	Group	No	NA	No
Kim 2015	81	100%	60 minutes, twice a week	12 weeks	Community or gym	Group	No	NA	No
Langlois 2013	75	74%	60 minutes, three times a week	12 weeks	Community or gym	Group	No	NA	No
Mugueta‐Aguinaga 2017	84	60%	20 minutes, three times a week	6 weeks	Community or gym	Individual	No	NA	No
Rydwik 2008	86	60%	60 minutes, twice a week	12 weeks	Community or gym	Group	Yes	Several times per week	No
Tarazona‐Santabalbina 2016	80	54%	65 minutes, five times a week	24 weeks	Community or gym	Group	No	NA	No
Worm 2001	81	57%	60 minutes, twice a week	12 weeks	Community or gym	Group	Yes	Daily	No

Open in a new tab

NA: not applicable

We grouped the intervention arms by exercise modality into six categories (Appendix 1), using the Prevention of Falls Network Europe (ProFaNE) taxonomy (Appendix 2).

Most intervention arms (n = 8, 67%) included gait, balance, and functional exercises as the primary intervention (ProFaNE taxonomy code gait, balance, co‐ordination, and functional task training) (Binder 2002; Chin A Paw 2002; Fairhall 2012; Gill 2004; Gine‐Garriga 2010; Mugueta‐Aguinaga 2017; Rydwik 2008; Worm 2001).
Strength or resistance training was the primary component of one (8%) intervention arm (Kim 2015).
Endurance training alone was the primary component of one (8%) intervention arm (Langlois 2013).
Multiple categories of ProFaNE taxonomy were the primary intervention in 2 (17%) intervention arms (Bonnefoy 2003; Tarazona‐Santabalbina 2016). In one study, the primary exercise category was gait, balance, and functional training; strength or resistance training; and flexibility training (Bonnefoy 2003). In the other study, the primary exercise category was strength or resistance training and endurance training (Tarazona‐Santabalbina 2016).

Five studies measured and reported adherence (Appendix 5). All five studies reported the proportion of scheduled sessions attended or completed.

Outcomes

Mobility

Seven different mobility outcome measures were reported across the 12 studies that measured mobility outcomes. These included the Short Physical Performance Battery (four studies) (Binder 2002; Fairhall 2012; Mugueta‐Aguinaga 2017; Tarazona‐Santabalbina 2016); Timed Up and Go test (three studies) (Kim 2015; Langlois 2013; Rydwik 2008); walking distance (one study) (Langlois 2013); walking speed (nine studies) (Bonnefoy 2003; Chin A Paw 2002; Fairhall 2012; Gill 2004; Gine‐Garriga 2010; Kim 2015; Langlois 2013; Rydwik 2008; Worm 2001); sit to stand (five studies) (Bonnefoy 2003; Chin A Paw 2002; Gill 2004; Gine‐Garriga 2010; Rydwik 2008); Berg Balance Scale (two studies) (Binder 2002; Worm 2001); and single leg standing time (three studies) (Binder 2002; Gine‐Garriga 2010; Rydwik 2008). All mobility outcome measures were performance based.

Functioning

Eight different functional outcome measures were reported across the nine studies that measured functional outcomes. These include the Barthel Index (four studies) (Fairhall 2012; Gine‐Garriga 2010; Mugueta‐Aguinaga 2017; Tarazona‐Santabalbina 2016); Lawton Scale (one study) (Tarazona‐Santabalbina 2016); Physical Performance Test (four studies) (Binder 2002; Gill 2004; Langlois 2013; Tarazona‐Santabalbina 2016); Performance Sum Score (one study) (Chin A Paw 2002); Activity of Daily Living (ADL) score (one study) (Chin A Paw 2002); Instrumental Activities of Daily Living (Instrumental Activities Measure) (one study) (Rydwik 2008); Activity Measure for Post Acute Care (AMPAC) (one study) (Fairhall 2012); and Functional Independence Measure (one study) (Rydwik 2008). Functional outcome measures were a combination of performance based or interview or self‐report. Of the nine studies that contributed to the meta‐analysis of the function outcome, eight were performance based (Binder 2002; Chin A Paw 2002; Fairhall 2012; Gill 2004; Gine‐Garriga 2010; Langlois 2013; Mugueta‐Aguinaga 2017; Tarazona‐Santabalbina 2016), and one was interview or self report (Rydwik 2008).

Death

Six studies reported on the number of deaths (Fairhall 2012; Gill 2004; Gine‐Garriga 2010; Kim 2015; Mugueta‐Aguinaga 2017; Rydwik 2008).

Admission to a nursing care facility

One study reported on the number of admissions to hospital or nursing care facility (Fairhall 2012).

Falls

Two studies reported on the number of falls (Fairhall 2012; Tarazona‐Santabalbina 2016), while two studies reported on the number of fallers (Fairhall 2012; Gill 2004).

Adverse events

Adverse events were reported to a degree in both the intervention and control groups in two trials (Gill 2004; Gine‐Garriga 2010) and in the intervention group only in a further three trials (Binder 2002; Chin A Paw 2002; Fairhall 2012). We describe the adverse events reported in Appendix 6.

Frailty measures

Three studies used the Cardiovascular Health Study Frailty Phenotype, Fried 2001, to categorise frailty (Fairhall 2012; Kim 2015; Tarazona‐Santabalbina 2016); five studies used the cut‐off score of an objective measure (Binder 2002; Gill 2004; Gine‐Garriga 2010; Mugueta‐Aguinaga 2017; Rydwik 2008); and three studies used difficulty performing specified tasks or other measures of vulnerability (Bonnefoy 2003; Chin A Paw 2002; Worm 2001). One study used a combination of a frailty criterion and the cut‐off score of an objective measure (Langlois 2013). Two studies reported the degree of frailty using the number of physiological deficits reported on the Cardiovascular Health Study Frailty Phenotype (Fairhall 2012; Kim 2015). One study reported the degree of frailty using the number of outcome criteria the participant met (Gill 2004) (see Appendix 7).

Excluded studies

On full‐text review, we excluded 17 studies for the following reasons (Beling 2009; Binder 2005; Clegg 2014; De Vries 2015; Faber 2006; Halvarsson 2013; Hiroyuki 2003; Ikezoe 2005; Latham 2003; Law 2018; Lord 2003; Ng 2014; Pahor 2014; Park 2012; Salem 2017; Seino 2017; Trombetti 2018):

12 trials did not meet the review's inclusion criteria for participants;
2 trials did not include mobility training as an intervention;
1 trial did not include an eligible comparison group;
1 trial included participants who were not community‐dwelling; and
1 trial measured and reported only one outcome (body fat‐free mass), which is not an outcome of interest in this review.

We describe these studies in the Characteristics of excluded studies.

Risk of bias in included studies

We summarised the assessment of each domain of risk of bias for the included trials in the Characteristics of included studies. We also graphically presented the results of our risk of bias assessment in Figure 2 and Figure 3.

Risk of bias graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

Risk of bias summary: review authors' judgements about each risk of bias item for each included study.

Allocation

Four trials reported adequate random sequence generation and allocation concealment, and we deemed these to have low risk of selection bias (Binder 2002; Bonnefoy 2003; Fairhall 2012; Gine‐Garriga 2010). One trial had inadequate random sequence generation and allocation concealment, so we deemed it to have high risk of bias (Rydwik 2008). We judged one trial to have unclear risk of selection bias because of failure to report their method of randomisation (Chin A Paw 2002), and we judged two trials to have unclear risk of selection bias because of failure to report their method of allocation concealment (Gill 2004; Kim 2015). We judged four trials to have an unclear risk of selection bias because of failure to report their method of randomisation and allocation concealment (Langlois 2013; Mugueta‐Aguinaga 2017; Tarazona‐Santabalbina 2016; Worm 2001).

Blinding

Blinding of participants and personnel (performance bias)

Because of the nature of the intervention, it was not possible to blind the personnel and participants to group allocation; hence, all studies are at high risk of performance bias except for one (Gill 2004), where it was unclear whether participants were blinded to group allocations or not.

Blinding of outcome assessment for assessor‐reported outcomes (detection bias)

Four trials reported blinding of outcome assessors for all assessor‐reported outcomes (mobility and function), and we judged these to have low risk of detection bias (Fairhall 2012; Gill 2004; Kim 2015; Tarazona‐Santabalbina 2016). Four trials reported that they did not blind their outcome assessors to group allocation, so we considered these trials to have high risk of detection bias for mobility and function outcomes (Binder 2002; Chin A Paw 2002; Gine‐Garriga 2010; Rydwik 2008). We judged four trials to have unclear risk of detection bias as they did not report on the blinding of assessors for mobility or function (Bonnefoy 2003; Langlois 2013; Mugueta‐Aguinaga 2017; Worm 2001). There is a low risk of bias in all studies for both 'nursing home admissions' and 'death' as knowledge of intervention should not impact on the outcome.

Blinding of outcome assessment for self‐reported outcomes (detection bias)

Almost all trials had high risk of detection bias for the self‐reported outcomes of number of fallers and adverse events, as none of the included trials blinded participants. It was unclear if one trial, Gill 2004, blinded participants to group allocations or not; hence, there is unclear risk of detection bias for falls and adverse events.

Incomplete outcome data

Risk of attrition was low in eight trials (Binder 2002; Fairhall 2012; Gill 2004; Kim 2015; Langlois 2013; Mugueta‐Aguinaga 2017; Tarazona‐Santabalbina 2016; Worm 2001) because there was either no withdrawals or numbers and reasons for withdrawals were similar across groups. Risk of attrition was high in four trials as they all reported an unequal dropout rate between groups (Bonnefoy 2003; Chin A Paw 2002; Gine‐Garriga 2010; Rydwik 2008).

Selective reporting

Risk of reporting bias was low in one trial as the paper reported all proposed outcomes in the results (Fairhall 2012). Risk of reporting bias was high in two trials as the papers did not report all prespecified outcomes in the results (Kim 2015; Tarazona‐Santabalbina 2016). Risk of reporting bias was unclear in nine trials as these trials had not been registered in a trial registry or a protocol was unavailable (Binder 2002; Bonnefoy 2003; Chin A Paw 2002; Gill 2004; Gine‐Garriga 2010; Langlois 2013; Mugueta‐Aguinaga 2017; Rydwik 2008; Worm 2001).

Other potential sources of bias

None of the included studies were subject to other bias.

Effects of interventions

See: Table 1; Table 2; Table 3

Mobility training versus control

Twelve studies compared mobility training to control (Binder 2002; Bonnefoy 2003; Chin A Paw 2002; Fairhall 2012; Gill 2004; Gine‐Garriga 2010; Kim 2015; Langlois 2013; Mugueta‐Aguinaga 2017; Rydwik 2008; Tarazona‐Santabalbina 2016; Worm 2001). Table 1 gives the results for the comparison mobility training versus control (this included general health education, social visits, cognitive therapy, no intervention, and very gentle exercise not expected to impact on mobility).

Five studies reported outcomes up to six months post‐intervention for mobility training versus control (Bonnefoy 2003; Gill 2004; Gine‐Garriga 2010; Kim 2015; Rydwik 2008). We present the results in Table 2.

Benefits

Mobility

High‐certainty evidence indicates that mobility training is effective for improving mobility levels compared to those not receiving mobility training. Mobility training improves the level of mobility by 0.47 standardised mean difference (SMD) at the end of the intervention period compared to those not receiving mobility training (SMD 0.47, 95% confidence interval (CI) 0.24 to 0.71; participants = 1151; studies = 12; I² = 72%; high‐certainty evidence; Analysis 1.1). When we back‐translated the SMD to the Short Physical Performance Battery (SPPB) (scale of 0 to 12, higher scores indicate better mobility levels), this resulted in a clinically important difference of 1.00 point (95% CI 0.51 to 1.51). The mean mobility score was 4.69 in the control group and improved by 1.00 point (95% CI 0.51 to 1.51) with mobility training, an absolute improvement of 8% (4% higher to 13% higher) and a relative improvement of 17% (9% higher to 26% higher) on the 12‐point scale. This better performance in the mobility training group translated to a number needed to treat for an additional beneficial outcome (NNTB) of 5 (95% CI 3.00 to 9.00). A difference of 0.5 points on the SPPB is considered a clinically significant difference (Perera 2006). When we undertook the GRADE assessment, we did not downgrade for risk of bias as results were essentially unchanged when we removed the trials with a high risk of bias for one or more items or when we removed the studies with high or unclear risk for selection bias or detection bias.

1.1 — Comparison 1: Mobility training vs control (mobility), Outcome 1: Mobility ‐ overall analysis

There is moderate‐certainty evidence that the improvement is maintained six months post the intervention period. Mobility training improves the level of mobility by 0.32 SMD compared to those not receiving mobility training (SMD 0.32, 95% CI 0.10 to 0.54; participants = 451; studies = 5; I² = 22%; moderate‐certainty evidence; Analysis 1.2). When we back‐translated the SMD to the SPPB, this resulted in a clinically important difference of 0.72 points (95% CI 0.21 to 1.14), an absolute improvement of 6% (2% higher to 10% higher) and a relative improvement of 13% (4% higher to 20% higher), which translated to a NNTB of 7 (95% CI 4.00 to 22).

1.2 — Comparison 1: Mobility training vs control (mobility), Outcome 2: Mobility ‐ 4‐ to 6‐month follow‐up

Function

Mobility training likely improves the level of functioning by 0.60 SMD at the end of the intervention period compared to those not receiving mobility training (SMD 0.60, 95% CI 0.21 to 1.00; participants = 916; studies = 9; I² = 87%; moderate‐certainty evidence downgraded for inconsistency; Analysis 2.1). When we back‐translated the SMD to the Barthel Index (scale of 0 to 100, higher scores indicate better mobility levels), this resulted in a difference of 8.58 points (95% CI 3.00 to 14.30). The mean function score was 86.1 in the control group and improved by 8.58 points (95% CI 3.00 to 14.30) on the Barthel Index (on a scale of 0 to 100, higher scores indicate better functioning levels) with mobility training, an absolute improvement of 9% (3% higher to 14% higher) and a relative improvement of 9% (3% higher to 15% higher). A difference of 9.8 point on the Barthel Index is considered a clinically significant difference (Unnanuntana 2018).

2.1 — Comparison 2: Mobility training vs control (function), Outcome 1: Function ‐ overall analysis

Mobility training may result in little to no difference to the function level six months post‐intervention as the 95% confidence interval includes the possibility of both reduced and increased function levels (SMD 1.29, 95% CI ‐0.38 to 2.96; participants = 278; studies = 3; I² = 96%; low‐certainty evidence downgraded for risk of bias and inconsistency; Analysis 2.2). When we back‐translated the SMD to the Barthel Index, this resulted in an improvement of 18.45 points (95% CI 5.43 worse to 42.33 higher), an absolute improvement of 18% (5% lower to 42% higher) and a relative improvement of 20% (6% lower to 46% higher).

2.2 — Comparison 2: Mobility training vs control (function), Outcome 2: Function (4‐ to 6‐month follow‐up)

Harms

Adverse effects

Five of the 12 trials reported adverse events (Appendix 6). Participants in both the intervention and control groups reported adverse events in two trials, and participants in the intervention group only reported adverse events in three trials. The majority were non‐serious and of a musculoskeletal nature (rotator cuff injury, worsening of an existing shoulder problem, back pain, fall‐related fractures, chest pain, physician‐diagnosed angina, and musculoskeletal complaints leading to restriction in usual activities). We are uncertain of the effect of mobility training on adverse events as we assessed the certainty of the evidence as very low. The number of events per group was 771 per 1000 in the control group and 562 per 1000 in the intervention group; risk ratio (RR) 0.74, 95% CI 0.63 to 0.88; an absolute difference of 19% fewer (9% fewer to 26% fewer) and a relative difference of 26% fewer (12% fewer to 37% fewer) (participants = 225; studies = 2; downgraded two levels for risk of bias and imprecision; Analysis 3.1).

3.1 — Comparison 3: Mobility training vs control (adverse effects), Outcome 1: Adverse events

Admission to a nursing care facility

One study reported admission to nursing care facilities (Fairhall 2012). Mobility training may result in little to no difference in the number of people who are admitted to nursing care facilities as the 95% confidence interval includes the possibility of both reduced and decreased admissions to nursing care facilities. The number of events per group was 248 per 1000 in the control group and 208 per 1000 in the intervention group; RR 0.84, 95% CI 0.53 to 1.34; an absolute difference of 4% fewer (8% more to 12% fewer) and a relative difference of 16% fewer (34% more to 47% fewer) (participants = 241; studies = 1; low‐certainty evidence downgraded for risk of bias and imprecision; Analysis 4.1).

4.1 — Comparison 4: Mobility training vs control (admission to a nursing care facility), Outcome 1: Admission to a nursing care facility

Falls

Two studies reported on this outcome (Fairhall 2012; Gill 2004). Mobility training may result in little to no difference in the number of people who fall as the 95% confidence interval includes the possibility of both reduced and decreased number of fallers. The number of events per group was 573 per 1000 in the control group and 584 per 1000 in the intervention group; RR 1.02, 95% CI 0.87 to 1.20; an absolute improvement of 1% (12% higher to 7% less) and a relative improvement of 2% (20% more to 13% fewer) (participants = 425; studies = 2; I² = 0%; low‐certainty evidence downgraded for imprecision (wide CI) and bias; Analysis 5.1).

5.1 — Comparison 5: Mobility training vs control (falls), Outcome 1: Number of fallers

Death

Mobility training probably results in little to no difference in the death rate as the 95% confidence interval includes the possibility of both reduced and increased death rates. The number of events per group was 51 per 1000 in the control group and 59 per 1000 in the intervention group; RR 1.16, 95% CI 0.64 to 2.10; an absolute difference of 1% more deaths (6% more deaths to 2% less deaths) and a relative difference of 16% more (110% more to 36% fewer) (participants = 747; studies = 6; I² = 0%; moderate‐certainty evidence downgraded for bias; Analysis 6.1).

6.1 — Comparison 6: Mobility training vs control (death), Outcome 1: Number of people who died

Costs

Only one study reported the economic costs (Fairhall 2012). The cost for 1 extra person to transition out of frailty was $15,955 (Australian dollars).

Subgroup analysis

Four studies included participants with cognitive impairment (Bonnefoy 2003; Gine‐Garriga 2010; Mugueta‐Aguinaga 2017; Worm 2001), and eight studies included participants without cognitive impairment (Binder 2002; Chin A Paw 2002; Fairhall 2012; Gill 2004; Kim 2015; Langlois 2013; Rydwik 2008; Tarazona‐Santabalbina 2016). Table 3 gives the results for the comparison mobility training versus control.

Mobility

Subgroup analysis of trials that included participants with cognitive impairment showed there was little or no difference in the effect of mobility training on mobility levels where all participants were without cognitive impairment (SMD 0.37, 95% CI 0.14 to 0.60; participants = 980; studies = 8; I² = 66%) compared to trials that included participants with cognitive impairments (SMD 0.91, 95% CI 0.19 to 1.64; participants = 171; studies = 4; I² = 80%; test for subgroup differences: Chi² = 1.97; degrees of freedom (df) = 1; P = 0.16; I² = 49.3%; Analysis 1.3).

1.3 — Comparison 1: Mobility training vs control (mobility), Outcome 3: Mobility ‐ subgrouped by inclusion of participants with cognitive impairment

Function

Subgroup analysis of trials that included or excluded participants with cognitive impairment showed there was little or no difference in the effect of mobility training on function where all participants had no cognitive impairment (SMD 0.31, 95% CI 0.07 to 0.56; participants = 839; studies = 7; I² = 63%) compared to trials that included participants with cognitive impairments (SMD 2.92, 95% CI ‐0.22 to 6.07; participants = 80; studies = 2; I² = 95; test for subgroup differences: Chi² = 2.62; df = 1; P = 0.11; I² = 61.9%; Analysis 2.3).

2.3 — Comparison 2: Mobility training vs control (function), Outcome 3: Function ‐ subgrouped by inclusion of participants with cognitive impairment

We did not complete subgroup analysis on 'Interventions specifically prescribed to challenge the participant versus interventions that are not prescribed with the intent to challenge the participant' as all studies reported having interventions specifically prescribed to challenge the participant.

We did not complete subgroup analysis based on adherence because of the low number of trials that reported the same methods for adherence.

Sensitivity analyses

As shown in Analysis 7.1 (SMD 0.44, 95% CI 0.32 to 0.56; participants = 1151; studies = 12; I² = 72%), Analysis 7.2.2 (SMD 0.63, 95% CI 0.10 to 1.16; participants = 419; studies = 3; I² = 82%), and Analysis 7.2.3 (SMD 0.54, 95% CI 0.20 to 0.89; participants = 595; studies = 4; I² = 75%), the sensitivity analyses for mobility outcomes (removing studies with a high risk of bias; removal of studies with high or unclear risk for selection or detection bias; or undertaking fixed‐effect analyses) made little difference to the results of the primary pooled analysis. The following made little difference to the results of the primary pooled analysis for mobility: removal of studies that only provided home‐based programmes (Fairhall 2012; Gill 2004) (SMD 0.50, 95% CI 0.20 to 0.81; participants = 766; studies = 10; I² = 76%; Analysis 7.3); removal of studies with the longest and shortest intervention period (Fairhall 2012; Gill 2004; Mugueta‐Aguinaga 2017) (SMD 0.46, 95% CI 0.15 to 0.78; participants = 727; studies = 9; I² = 76%; Analysis 7.4); and removal of studies with the largest and smallest dose of intervention per week (Tarazona‐Santabalbina 2016; Mugueta‐Aguinaga 2017) (SMD 0.41, 95% CI 0.08 to 0.74; participants = 645; studies = 8; I² = 75%; Analysis 7.5). This indicates the robustness of these findings.

7.1 — Comparison 7: Sensitivity analysis, Outcome 1: Mobility ‐ analysis using fixed‐effects

7.2 — Comparison 7: Sensitivity analysis, Outcome 2: Mobility ‐ overall vs risk of bias studies removed

7.3 — Comparison 7: Sensitivity analysis, Outcome 3: Mobility ‐ removal of studies that only provided home‐based programmes

7.4 — Comparison 7: Sensitivity analysis, Outcome 4: Mobility ‐ removal of studies with the longest and shortest intervention periods

7.5 — Comparison 7: Sensitivity analysis, Outcome 5: Mobility ‐ removal of studies with the longest and shortest dose per week

For functional outcomes, as shown in Analysis 7.6 (SMD 0.42, 95% CI 0.28 to 0.55; participants = 916; studies = 9; I² = 87%), Analysis 7.7.2 (SMD 1.32, 95% CI 0.05 to 2.60; participants = 465; studies = 3; I² = 79%), and Analysis 7.7.3 (SMD 0.43, 95% CI 0.00 to 0.85; participants = 465; studies = 3; I² = 79%), removing studies with high or unclear risk of selection bias and detection bias had an impact on results of the primary pooled analysis. The following made little difference to the results of the primary pooled analysis to function: removal of studies that only provided home‐based programmes (SMD 0.75, 95% CI 0.20 to 1.29; participants = 533; studies = 7; I² = 88%; Analysis 7.8); removal of studies with the longest and shortest intervention period (SMD 0.67, 95% CI 0.09 to 1.25; participants = 494; studies = 9; I² = 89%; Analysis 7.9); and removal of studies with the largest and smallest dose of intervention per week (Mugueta‐Aguinaga 2017; Tarazona‐Santabalbina 2016) (SMD 0.64, 95% CI ‐0.03 to 1.30; participants = 422; studies = 5; I² = 89%; Analysis 7.10).

7.6 — Comparison 7: Sensitivity analysis, Outcome 6: Function ‐ analysis using fixed‐effects

7.7 — Comparison 7: Sensitivity analysis, Outcome 7: Function ‐ overall vs risk of bias studies removed

7.8 — Comparison 7: Sensitivity analysis, Outcome 8: Function ‐ removal of studies that only provided home‐based programmes

7.9 — Comparison 7: Sensitivity analysis, Outcome 9: Function ‐ removal of studies with the longest and shortest intervention period

7.10 — Comparison 7: Sensitivity analysis, Outcome 10: Function ‐ removal of studies with the longest and shortest dose per week

This review's primary analyses display moderate to substantial heterogeneity with P < 0.05 for the Chi² test and I² values up to 87%. Our subgroup analyses did not explain this variability. We consider this likely to represent between‐study differences in the nature of programmes and populations, which requires ongoing investigation. Given the overall positive impact of the programmes and the stability of results, we do not consider this to preclude the meta‐analyses we have undertaken.

Funnel plots

The funnel plots in Figure 4 (1.1. Mobility ‐ overall analysis), Figure 5 (2.1 Function ‐ overall analysis), and Figure 6 (6.1 Number of people who died) do show some asymmetry; however, we did not consider the asymmetry sufficient to downgrade the level of evidence for these outcomes.

Figure 4. Funnel plot of comparison: 1 Mobility training vs control (mobility), outcome: 1.1 Mobility ‐ overall analysis.

Figure 5. Funnel plot of comparison: 2 Mobility training vs control (Function), outcome: 2.1 Function ‐ overall analysis.

Figure 6. Funnel plot of comparison: 6 Mobility training vs control (number of people who died), outcome: 6.1 Number of people who died ‐ overall analysis.

Discussion

Summary of main results

This Cochrane Review included 12 trials, with 1317 participants. All trials compared mobility training with a control intervention (defined as one that is not thought to improve mobility, such as general health education, social visits, very gentle exercise, or "sham" exercise not expected to impact on mobility). We found high‐certainty evidence (12 studies, 1151 participants) that mobility training provides a clinically important improvement to mobility in a frail community‐dwelling population compared to control (general health education, social visits, cognitive therapy, no intervention, very gentle exercise not expected to impact on mobility). There is moderate‐certainty evidence that a clinically important improvement in mobility is maintained at six months post‐intervention. Moderate‐certainty evidence from 9 trials, with 916 participants, showed that mobility training probably provides an improvement to functioning. Mobility training may improve function levels at six months post‐intervention. Trials were subject to performance and selection biases and inconsistency. All interventions involved a component of gait, balance, or functional training (based on the Prevention of Falls Network Europe (ProFaNE) taxonomy).

The number needed to treat for an additional beneficial outcome (NNTB), to make a minimally clinical important difference in mobility (Short Physical Performance Battery (SPPB)), at the end of the intervention period was five, and at six months after the intervention period, the NNTB was slightly higher (seven).

Overall, subgroup analyses revealed that the effects of mobility training on both mobility and functional outcomes were similar for studies that excluded people with cognitive impairments compared to studies that included people with cognitive impairments.

Sensitivity analyses revealed the results for mobility were stable (see Appendix 8), which suggests that the results are robust to key risks of bias; hence, we did not downgrade the evidence. When we undertook the GRADE assessment, we downgraded the certainty of evidence for function based on sensitivity analysis (removal of trials with high or unclear risk of selection bias).

The majority of the studies provided the intervention in a group setting. Two of the three studies that provided the intervention one‐to‐one provided face‐to‐face sessions combined with a home exercise programme. These are both efficient ways of delivering exercise, and although not measured, the economic cost of the intervention is likely to be low.

Only two studies reported adverse events in both exercise and control groups over the whole trial period (225 participants). The adverse events included rotator cuff injury, worsening of an existing shoulder problem, back pain, fall‐related fractures, chest pain, physician‐diagnosed angina, and musculoskeletal complaints leading to restriction in usual activities. There was no consistent method of monitoring adverse events across all studies. Due to the small number of studies and inconsistent method of monitoring for adverse events, we are unsure of the harmful effects of the intervention.

Due to low‐certainty evidence, we are unsure of the effect of mobility training on admissions to nursing care facilities, on the number of people who fall, and on mortality rates. Only one study reported on admissions to nursing care facilities (241 participants). Two studies reported on number of people who had a fall (425 participants). Six studies, with 747 participants, provided data on death rates. For all of these outcomes, we downgraded the evidence because of bias and imprecision. Considering the short follow‐up duration, the intervention was unlikely to show a difference in the mortality rate.

Given that our inclusion criteria meant that all included trials needed to include mobility training, we did not undertake any meta‐analyses of comparisons between different types of exercise as it is likely that we excluded trials that included other exercise types from this review.

As there was only a short follow‐up period, we are unable to conclude the effect of the intervention on long‐term mobility outcomes. We are also unlikely to see a difference in outcomes, such as admissions to nursing care facilities or mortality, with a short follow‐up period.

Overall completeness and applicability of evidence

The 12 studies in this Cochrane Review included 1317 community‐dwelling older people with frailty, predominantly women (73%). Nine of the trials reported outcomes for function. All 12 trials reported outcomes for mobility. The trials included participants from nine different countries.

We found similar results for those studies that excluded individuals with cognitive impairment as we did for those studies that included individuals with cognitive impairment. Therefore, the results of this review are applicable to individuals with cognitive impairment.

The median sample size of the trials was 97 participants, with a mean age of 82. Seven studies reported follow‐up lengths of seven months or longer. Two studies reported follow‐up length of 12 months. The primary time period of outcome assessment was the completion of the intervention programme.

Very few studies (three) reported on falls (either falls rates or number of fallers). The small number of studies makes it difficult to draw conclusions for this outcome.

The sensitivity analyses for mobility outcomes (removing studies with a high risk of bias and undertaking fixed‐effect analyses) made little difference to the results of the primary pooled analysis. This indicates the robustness of these findings. Measurement of mobility varied across the 12 studies. The included studies used seven different mobility outcomes (Short Physical Performance Battery, Timed Up and Go test, walking distance, walking speed, sit to stand, Berg Balance Scale, single leg standing time) (Appendix 9).

Measurement of function varied across the nine studies. The included studies used eight different functional outcomes (Barthel Index, Lawton scale, Physical Performance test, Performance Sum Score, Activities of Daily Living (ADL) score, Instrumental Activities of Daily Living (Instrumental Activities Measure), Activity Measure for Post Acute Care (AMPAC), Functional Independence Measure (FIM)) (Appendix 9).

Measurement of adverse events varied across the five studies that reported this outcome. Examples of the sorts of adverse events reported in the studies are rotator cuff injury, worsening of an existing shoulder problem, back pain, fall‐related fractures, chest pain, physician‐diagnosed angina, and musculoskeletal complaints leading to restriction in usual activities. There was no consistent method of monitoring adverse events across all studies.

Five studies reported measurement of adherence. All five studies reported the proportion of scheduled sessions attended or completed.

Many of the studies did not provide sufficient details of the intervention provided, which makes it difficult to replicate these interventions in clinical practice.

The included trials did not sufficiently report the degree of participant frailty to identify if the level of frailty would have an impact on the results.

Quality of the evidence

This review, which included 12 trials (1317 participants), provides low‐ to high‐certainty evidence of the effectiveness of mobility training interventions for improving functioning and mobility in older community‐dwelling people.

For mobility, there was high‐certainty evidence from 12 studies. We did not downgrade the certainty of the evidence. Our included studies reported seven different mobility outcome measures. For function, there was moderate‐certainty evidence from nine studies, which we downgraded by one level for inconsistency (as there was considerable heterogeneity amongst the studies). Our included studies reported eight different functional outcome measures. As none of the 12 studies used the same mobility outcome measure, we used the standardised mean difference to enable the comparison of results.

For adverse events, there was very low‐certainty evidence from two studies, which we downgraded by three levels. We downgraded one level for imprecision, with only two studies reporting adverse events in both groups, and two levels for limitations in the design of studies, which suggests a high likelihood of bias. This is because throughout the trial period, both groups did not report the number of participants experiencing adverse events in the same manner.

For the number of people admitted to nursing care facilities, there was low‐certainty evidence from one study, which we downgraded by two levels. We downgraded once because of imprecision, as there was only one study, and once because of the probability of study design limitations (selective reporting bias), as it is reasonable to assume that other studies measured the outcome but did not report it.

For the number of people who fell, there was low‐certainty evidence from two studies, which we downgraded by two levels. We downgraded one level because of imprecision, as there were only two studies, and one level because of the probability of study design limitations (selective reporting bias), as it is reasonable to assume that other studies measured falls but did not report them.

For the death rate, there was moderate‐certainty evidence from six studies, which we downgraded one level because of the probability of study design limitations (selective reporting bias), as it is reasonable to assume that other studies measured deaths but did not report them.

Overall, the certainty of the evidence for different outcomes ranged from very low to high.

Potential biases in the review process

We conducted a comprehensive search of the published literature using multiple databases and also searched clinical trial registries for completed trials for which we had not identified full reports. Two authors independently performed screening and data extraction in duplicate to minimise bias. Despite this thorough search strategy, we acknowledge the possibility that we may have missed some relevant trials, especially if they were published in languages other than English.

Agreements and disagreements with other studies or reviews

To date, this appears to be the first systematic review that specifically looks at the effect of mobility training interventions in older people with frailty. A previous systematic review by Gine‐Garriga 2014 looked at the effect of all types of physical exercise on a community‐dwelling frail population. This review found improvements in some mobility outcomes (walking speed and the SPPB); however, there were inconsistent results for other mobility and functional outcomes. Similar to our systematic review, Gine‐Garriga 2014 found no effect on the rate of falls in their review. A systematic review by Chou 2012 similarly looked at the effect of all types of physical exercise in a community‐dwelling frail population. They also found improvements in some mobility and functional outcomes measures (walking speed; Berg Balance Scale; and a mean weighted score of three Activity of Daily Living scales: FIM and Functional Status Questionnaire). While both of these reviews included all types of physical exercise, the majority of studies within these two systematic reviews included mobility training interventions.

Authors' conclusions

Implications for practice.

The synthesis of the data in this review supports the use of mobility training for increasing mobility in older people with frailty living in the community. This review provides high‐certainty evidence that mobility training in older people with frailty provides clinically important benefits to mobility at the end of the intervention period. The improvement in mobility continued up to six months after completion of the intervention. This indicates there is probably some short‐ to medium‐term carryover benefit upon completion of mobility training. This review provides moderate‐certainty evidence that mobility training may improve function at the end of the intervention period, but this effect did not sustain at six months after completion of the intervention. This would indicate that with mobility training, improvement in mobility may not necessarily be accompanied by an improvement in function at six months.

Mobility training did not appear to make a difference to the number of people admitted to a nursing care facility, and there did not appear to be an increase in falls or deaths. We are unsure of the effects of mobility training on adverse events.

Implications for research.

The findings of the subgroup analyses reveal similar effects of mobility training on mobility and function in studies that included individuals without cognitive impairment as well as studies that included participants with cognitive impairment. Within these studies, we were unable to separate results for those individuals with cognitive impairments from those without cognitive impairment. Further work is indicated in identifying the effect of mobility training on frail individuals with cognitive impairment.

Further work is indicated, especially given the relatively small total number of trial participants (< 1500) and the importance of mobility and functioning for older adults. There is also a need for studies that compare the effect of different types of exercise interventions on mobility and functional outcomes in the growing global older frail population. The findings of this review identified improvement in mobility outcomes six months post‐intervention. Additional studies with longer term follow‐up will assist to determine if the benefit can be long term.

Further well‐designed trials investigating mobility training for frail individuals are required to provide high‐quality evidence on the relative impact of mobility training interventions on adverse events, admission to nursing care facilities, and the number of people who fall.

History

Protocol first published: Issue 5, 2013

Acknowledgements

The authors would like to thank Cochrane Musculoskeletal for their editorial support, as well as Louise Falzon for assistance in developing the search strategy. We thank the following peer reviewers for their valuable comments on this review: Professor Walter R Frontera, MD, PhD, FRCP, University of Puerto Rico School of Medicine; Professor Allyson Jones, PhD, Department of Physical Therapy, University of Alberta, Edmonton, Alberta, Canada; Associate Network Editor Ms Jennifer Hilgart; and consumer reviewer Ms Catherine Hofstetter.

Appendices

Appendix 1. Categories of exercise (ProFaNE) in interventions in the included trials

Study ID	Gait or balance functional training	Strength or resistance training	Flexibility	3D (Tai Chi, dance, etc)	General physical activity	Endurance
Binder 2002	Primary	Secondary	Secondary	‐	‐	Secondary
Bonnefoy 2003	Primary	Primary	Primary	‐	‐	‐
Chin A Paw 2002	Primary	‐	‐	‐	‐	‐
Fairhall 2012	Primary	‐	‐	‐	‐	‐
Gill 2004	Primary	Secondary	‐	‐	‐	‐
Gine‐Garriga 2010	Primary	Secondary	‐	‐	‐	‐
Kim 2015	Secondary	Primary	‐	‐	‐	‐
Langlois 2013	Secondary	Secondary	Secondary	‐	‐	Primary
Mugueta‐Aguinaga 2017	Primary	‐	‐	‐	‐	‐
Rydwik 2008	Primary	Secondary	‐	‐	‐	Secondary
Tarazona‐Santabalbina 2016	Secondary	Primary	‐	‐	‐	Primary
Worm 2001	Primary	Secondary	Secondary	‐	‐	Secondary

Open in a new tab

Appendix 2. Categories of exercise (ProFaNE): definitions and application

Exercise category	ProFaNE description	How the category criteria were applied in this review*
Gait, balance, and functional training	Gait training involves specific correction of walking technique (e.g., posture, stride length and cadence) and changes of pace, level and direction. Balance Training involves the efficient transfer of bodyweight from one part of the body to another or challenges specific aspects of the balance systems (eg vestibular systems). Balance retraining activities range from the re‐education of basic functional movement patterns to a wide variety of dynamic activities that target more sophisticated aspects of balance. Functional training utilises functional activities as the training stimulus, and is based on the theoretical concept of task specificity. All gait, balance and functional training should be based on an assessment of the participant’s abilities prior to starting the program; tailoring of the intervention to the individuals abilities; and progression of the exercise program as ability improves.	Selected as exercise category if the intervention met the baseline assessment, tailoring and progression criteria. Selected as primary category for interventions where the majority of exercises were conducted in standing and where the intervention focus and majority of time spent was on exercise in this category
Strength/resistance (including power)	The term Resistance Training covers all types of weight training ie contracting the muscles against a resistance to ‘overload’ and bring about a training effect in the muscular system. The resistance is an external force, which can be one’s own body placed in an unusual relationship to gravity (e.g. prone back extension) or an external resistance (e.g. free weight). All strength or resistance training should be based on an assessment of the participant’s abilities prior to starting the program; tailoring of the intervention to the individuals abilities; and progression of the exercise program as ability improves.	Selected as exercise category if the intervention met the baseline assessment, tailoring and progression criteria. Selected as primary category for interventions where additional resistance was used or where it was clear that overload was sufficient without external resistance and where the intervention focus and majority of time spent was on exercise in this category
Flexibility	Flexibility Training is the planned process by which stretching exercises are practised and progressed to restore or maintain the optimal Range Of Movement (ROM) available to a joint or joints. The ranges of motion used by flexibility programs may vary from restoration/maintenance of the entire physiological range of motion, or alternatively, maintenance of range that is essential to mobility or other functions.	Selected as exercise category if the intervention met the progression of stretching criterion. Selected as primary category for interventions where flexibility training was a stated aim of the intervention and where the intervention focus and majority of time spent was on exercise in this category
3D	3D training involves constant movement in a controlled, fluid, repetitive way through all 3 spatial planes or dimensions (forward and back, side to side, and up and down). Tai Chi and Qi Gong incorporate specific weight transferences and require upright posture and subtle changes of head position and gaze direction. Dance involves a wide range of dynamic movement qualities, speeds and patterns.	Selected as exercise category if the intervention involved Tai Chi or dance. Selected as primary category for interventions where the intervention focus and majority of time spent was on exercise in this category
General physical activity	Physical Activity is any bodily movement produced by skeletal muscle contraction resulting in a substantial increase in energy expenditure. Physical activity has both occupational, transportation and recreational components and includes pursuits like golf, tennis, and swimming. It also includes other active pastimes like gardening, cutting wood, and carpentry. Physical activity can provide progressive health benefits and is a catalyst for improving health attitudes, health habits, and lifestyle. Increasing habitual physical activity should be with specific recommendations as to duration, frequency and intensity if a physical or mental health improvement is indicated.	Selected as exercise category if the intervention included unstructured physical activity. Programmes that included unstructured walking received this category. Selected as primary category for interventions where the intervention focus and majority of time spent was on exercise in this category
Endurance	Endurance training is aimed at cardiovascular conditioning and is aerobic in nature and simultaneously increases the heart rate and the return of blood to the heart.	Selected as exercise category if the intervention focused on structured aerobic training. Programmes that included treadmill walking received this category. Selected as primary category for interventions where the intervention focus and majority of time spent was on exercise in this category
Other	Other kind of exercises not described	Selected as exercise category if the intervention did not meet the other categories listed and where the intervention focus and majority of time spent was on exercise in this category
*Interventions were allocated a secondary category if some but not all criteria were met by the intervention or where the category was not the primary focus of the intervention, or both.

Open in a new tab

Appendix 3. Search strategy

Cochrane Central Register of Controlled Trial (Ovid SP)

1. MeSH descriptor Frail Elderly explode all trees

2. frail*.ti,ab.

3. (1 OR 2)

* indicates truncation

ti,ab denotes word in the title or abstract

Ovid MEDLINE (1946 to present)

1. exp Frail Elderly/

2. frail$.tw

3. or/1‐2

4. randomized controlled trial.pt.

5. controlled clinical trial.pt.

6. randomized.ab.

7. placebo.ab.

8. randomly.ab.

9. trial.ab.

10. groups.ab.

11. or/4‐10

12. humans.sh.

13. and/11,12

14. and/3,13

Ovid Embase (1947 to present)

1. Frail Elderly/

2. frail$.tw

3. or/1‐2

4. (random$ or placebo$).ti,ab.

5. ((single$ or double$ or triple$ or treble$) and (blind$ or mask$)).ti,ab.

6. controlled clinical trial$.ti,ab.

7. RETRACTED ARTICLE/

8. or/4‐8

9. (animal$ not human$).sh,hw.

10. 8 not 9

11. and/3,10

AMED (1985 to present)

1. Frail Elderly/

2. frail$.tw

3. or/1‐2

4. (random$ or placebo$).ti,ab.

5. ((single$ or double$ or triple$ or treble$) and (blind$ or mask$)).ti,ab.

6. controlled clinical trial$.ti,ab.

7. RETRACTED ARTICLE/

8. or/4‐8

9. (animal$ not human$).sh,hw.

10. 8 not 9

11. and/3,10

Footnote for OVID

.pt. denotes a Publication Type term

.ab. denotes a word in the abstract

.sh. or / denotes a Medical Subject Heading (MeSH) term

.ti. denotes a word in the title

PEDro (1999 to present)

1. Frail

2. Clinical trial

Appendix 4. Description of included studies: reference links

Study description	Links to references
Setting (country)	Australia: Fairhall 2012 Canada: Langlois 2013 Denmark: Worm 2001 France: Bonnefoy 2003 Japan: Kim 2015 The Netherlands: Chin A Paw 2002 Spain: Gine‐Garriga 2010; Mugueta‐Aguinaga 2017; Tarazona‐Santabalbina 2016 Sweden: Rydwik 2008 USA: Binder 2002; Gill 2004
Participants
Trials in which all participants were women	Kim 2015
Trials excluding people with cognitive impairment	Binder 2002; Chin A Paw 2002; Fairhall 2012; Gill 2004; Kim 2015; Langlois 2013; Rydwik 2008; Tarazona‐Santabalbina 2016
Study including people with cognitive impairment	Bonnefoy 2003; Gine‐Garriga 2010; Mugueta‐Aguinaga 2017; Worm 2001

Open in a new tab

Appendix 5. Adherence

Study ID	Adherence was measured	Adherence data were reported	Measurement of adherence	Reported adherence results
Binder 2002	No	No	‐	‐
Bonnefoy 2003	Yes	Yes	‐	‐
Chin A Paw 2002	Yes	Yes	Percentage of training sessions attended	Median: 90%; range: 47% to 100%
Fairhall 2012	Yes	Yes	Percentage of training sessions completed	13% did not complete any of the intervention. 29% completed 1% to 25% of the intervention. 16% completed 26% to 50% of the intervention. 21% completed 51% to 75% of the intervention. 21% completed 76% to 100%.
Gill 2004	No	No	‐	‐
Gine‐Garriga 2010	No	No	‐	‐
Kim 2015	No	No	‐	‐
Langlois 2013	No	No	‐	‐
Mugueta‐Aguinaga 2017	No	No	‐	‐
Rydwik 2008	Yes	Yes	Compliance was regarded as either a continuous variable or dichotomised, in which participants were regarded as compliers if they had attended > 65% of the training sessions.	Mean: 65%
Tarazona‐Santabalbina 2016	Yes	Yes	Percentage of training sessions attended	Mean: 47.3%
Worm 2001	Yes	Yes	Percentage of training sessions attended	Mean: 91.9%

Open in a new tab

Appendix 6. Adverse events

Study ID	Group in which adverse events were reported	Adverse events reported in intervention group	Adverse events reported in control group
Binder 2002	Intervention only	Rotator cuff injury (n = 1); worsening of an existing shoulder problem (n = 1)	Not applicable
Bonnefoy 2003	Neither	Not applicable	Not applicable
Chin A Paw 2002	Intervention only	Nil	Not applicable
Fairhall 2012	Intervention only	Participants with pre‐existing musculoskeletal conditions experienced back pain (n = 2)	Not applicable
Gill 2004	Intervention and control	Fall‐related fractures (n = 1); chest pain (23); physician‐diagnosed angina (6); musculoskeletal complaints leading to restriction in usual activities (30)	Fall‐related fractures (n = 6); chest pain (32); physician‐diagnosed angina (16); musculoskeletal complaints leading to restriction in usual activities (28)
Gine‐Garriga 2010	Intervention and control	Nil	Nil
Kim 2015	Neither	Not applicable	Not applicable
Langlois 2013	Neither	Not applicable	Not applicable
Mugueta‐Aguinaga 2017	Neither	Not applicable	Not applicable
Rydwik 2008	Neither	Not applicable	Not applicable
Tarazona‐Santabalbina 2016	Neither	Not applicable	Not applicable
Worm 2001	Neither	Not applicable	Not applicable

Open in a new tab

Appendix 7. Frailty definitions

Study ID	Definition	Degree of frailty
Binder 2002	A score less than 32 on the Modified Physical Performance Test	Not reported
Bonnefoy 2003	Elderly, with multiple diagnoses, on several medications, and were living in a retirement village	Not reported
Chin A Paw 2002	Inactivity and involuntary weight loss or a body mass index of less than 25 kg/m²	Not reported
Fairhall 2012	Frail according to the Cardiovascular Health Study Frailty Phenotype (Fried 2001)	Number of frailty criteria present (%): 3 (65), 4 (27), 5 (9)
Gill 2004	Physical frailty was defined on the basis of 2 tests of physical capability that are highly predictive of functional decline and disability. Persons were considered physically frail if they scored greater than 10 seconds on the Rapid Gait Test or could not stand from the chair with their arms folded. Persons meeting one of these criteria were considered moderately frail, and those meeting both criteria were considered severely frail.	Moderate frailty (meeting one of these criteria) 62%; severe frailty (meeting both of these criteria) 38%
Gine‐Garriga 2010	Physical frailty was defined according to the results of 2 tests of physical abilities (Gill 2002; Tinetti 1988) and according to 2 questions on the Center for Epidemiological Studies Depression Scale (Fried 2001). Individuals were considered physically frail and invited to participate in the study if they required more than 10 seconds to perform a Rapid Gait Test (i.e. to walk along a 3‐metre course and back at a quick comfortable pace), if they could not stand up 5 times from a seated position in a hardback chair with their arms folded (Gill 2002), or if they were categorised as frail by the exhaustion criterion (Fried 2001). Participants were asked to self‐report exhaustion using the following statements from the depression scale: “I felt that everything I did was an effort” and “I could not get going.” The question was asked as “How often in the last week did you feel this way?” and their answers graded as 0 = rarely or never (< 1 day), 1 = some or a little of the time (1 to 2 days), 2 = a moderate amount of time (3 to 4 days), or 3 = most of the time. Participants graded 2 or 3 were categorised as frail by the exhaustion criterion (Fried 2001).	Not reported
Kim 2015	Frail according to the Cardiovascular Health Study Frailty Phenotype (Fried 2001)	Number of frailty criteria present (%): 3 (46), 4 (41), 5 (13)
Langlois 2013	Participants were categorised as frail if they met at least 2 of the 3 following diagnostic criteria: (a) 3 of the 5 symptoms of frailty, according to the Cardiovascular Health Study Frailty Phenotype (Fried 2001); (b) a score of ≤ 28/36 on the Modified Physical Performance Test (Binder 2004); and (c) identified as frail according to the geriatrician’s judgment (mildly frail or worse on the clinical frailty scale) after assessing the 70 possible deficits of the frailty index (Rockwood 2005).	Not reported
Mugueta‐Aguinaga 2017	Short Physical Performance Battery score less than 10	Not reported
Rydwik 2008	Unintentional weight loss ≥ 5% or body mass index (BMI) ≤ 20 kg/m2, or both; and low physical activity level (≤ grade 3 on a 6‐graded scale of physical activity)	Not reported
Tarazona‐Santabalbina 2016	Frail according to the Cardiovascular Health Study Frailty Phenotype (Fried 2001)	Not reported
Worm 2001	Above 74 years and not able to leave their home unaided or unattended or without mobility aids	Not reported

Open in a new tab

Appendix 8. Sensitivity analyses exploring impact of risk of bias on effect sizes

Outcome	Number of comparisons in the analysis	Standard mean difference, 95% confidence interval (CI)	Number of comparisons in the sensitivity analysis^a	Standard mean difference, 95% CI	Impact
Mobility	12	0.47, 95% CI 0.24 to 0.71	4	0.63, 95% CI 0.10 to 1.16	Minimal
Function	9	0.60, 95% CI 0.21 to 1.00	3	1.32, 95% CI 0.05 to 2.60	Downgraded for bias

Open in a new tab

^aAfter removing trials assessed as high or unclear risk of bias in 1 or more key domains: random sequence generation (selection bias), allocation concealment (selection bias), blinding of outcome assessors (detection bias), and incomplete outcome data (attrition bias)

Appendix 9. Outcomes

Outcome	Assessment tool	Studies
Mobility	Short Physical Performance Battery	4 (Binder 2002; Fairhall 2012; Mugueta‐Aguinaga 2017; Tarazona‐Santabalbina 2016)
	Timed Up and Go	3 (Kim 2015; Langlois 2013; Rydwik 2008)
	Walking distance	1 (Langlois 2013)
	Walking speed	9 (Bonnefoy 2003; Chin A Paw 2002; Fairhall 2012; Gill 2004; Gine‐Garriga 2010; Kim 2015; Langlois 2013; Rydwik 2008; Worm 2001)
	Sit to stand	5 (Bonnefoy 2003; Chin A Paw 2002; Gill 2004; Gine‐Garriga 2010; Rydwik 2008)
	Berg Balance Scale	2 (Binder 2002; Worm 2001)
	Single leg standing time	3 (Binder 2002; Gine‐Garriga 2010; Rydwik 2008)
Function	Barthel Index	4 (Fairhall 2012; Gine‐Garriga 2010; Mugueta‐Aguinaga 2017; Tarazona‐Santabalbina 2016)
	Lawton scale	1 (Tarazona‐Santabalbina 2016)
	Physical Performance Test	4 (Binder 2002; Gill 2004; Langlois 2013; Tarazona‐Santabalbina 2016)
	Performance Sum Score	1 (Chin A Paw 2002)
	Activity of Daily Living (ADL) score	1 (Chin A Paw 2002)
	Instrumental Activities of Daily Living	1 (Rydwik 2008)
	Activity Measure for Post Acute Care (AMPAC)	1 (Fairhall 2012)
	Functional Independence Measure	1 (Rydwik 2008)

Open in a new tab

Appendix 10. Sensitivity analyses exploring impact of fixed‐effect analysis on effect sizes

Outcome	Number of comparisons in the analysis	Standard mean difference, 95% confidence interval (CI)	Number of comparisons in the sensitivity analysis (fixed‐effect meta‐analysis)	Standard mean difference, 95% CI	Impact
Mobility	12	0.47, 95% CI 0.24 to 0.71	12	0.44, 0.32 to 0.56	Minimal
Function	9	0.60, 95% CI 0.21 to 1.00	9	0.42, 0.28 to 0.55	Minimal

Open in a new tab

Data and analyses

Comparison 1. Mobility training vs control (mobility).

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Mobility ‐ overall analysis	12	1151	Std. Mean Difference (IV, Random, 95% CI)	0.47 [0.24, 0.71]
1.2 Mobility ‐ 4‐ to 6‐month follow‐up	5	451	Std. Mean Difference (IV, Random, 95% CI)	0.32 [0.10, 0.54]
1.3 Mobility ‐ subgrouped by inclusion of participants with cognitive impairment	12		Std. Mean Difference (IV, Random, 95% CI)	Subtotals only
1.3.1 Study included participants with cognitive impairments	4	171	Std. Mean Difference (IV, Random, 95% CI)	0.91 [0.19, 1.64]
1.3.2 Study excluded participants with cognitive impairments	8	980	Std. Mean Difference (IV, Random, 95% CI)	0.37 [0.14, 0.60]

Open in a new tab

Comparison 2. Mobility training vs control (function).

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 Function ‐ overall analysis	9	916	Std. Mean Difference (IV, Random, 95% CI)	0.60 [0.21, 1.00]
2.2 Function (4‐ to 6‐month follow‐up)	3	278	Std. Mean Difference (IV, Random, 95% CI)	1.29 [‐0.38, 2.96]
2.3 Function ‐ subgrouped by inclusion of participants with cognitive impairment	9		Std. Mean Difference (IV, Random, 95% CI)	Subtotals only
2.3.1 Study included participants with cognitive impairments	2	80	Std. Mean Difference (IV, Random, 95% CI)	2.92 [‐0.22, 6.07]
2.3.2 Study excluded participants with cognitive impairments	7	839	Std. Mean Difference (IV, Random, 95% CI)	0.31 [0.07, 0.56]

Open in a new tab

Comparison 3. Mobility training vs control (adverse effects).

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
3.1 Adverse events	2	225	Risk Ratio (IV, Fixed, 95% CI)	0.74 [0.63, 0.88]

Open in a new tab

Comparison 4. Mobility training vs control (admission to a nursing care facility).

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
4.1 Admission to a nursing care facility	1	241	Risk Ratio (IV, Random, 95% CI)	0.84 [0.53, 1.34]

Open in a new tab

Comparison 5. Mobility training vs control (falls).

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
5.1 Number of fallers	2	425	Risk Ratio (IV, Random, 95% CI)	1.02 [0.87, 1.20]

Open in a new tab

Comparison 6. Mobility training vs control (death).

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
6.1 Number of people who died	6	747	Risk Ratio (IV, Random, 95% CI)	1.16 [0.64, 2.10]

Open in a new tab

Comparison 7. Sensitivity analysis.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
7.1 Mobility ‐ analysis using fixed‐effects	12	1151	Std. Mean Difference (IV, Fixed, 95% CI)	0.44 [0.32, 0.56]
7.2 Mobility ‐ overall vs risk of bias studies removed	12		Std. Mean Difference (IV, Random, 95% CI)	Subtotals only
7.2.1 All studies included	12	1151	Std. Mean Difference (IV, Random, 95% CI)	0.47 [0.24, 0.71]
7.2.2 Trials at high or unclear risk of selection bias removed	4	419	Std. Mean Difference (IV, Random, 95% CI)	0.63 [0.10, 1.16]
7.2.3 Trials at high or unclear risk of detection bias removed	4	595	Std. Mean Difference (IV, Random, 95% CI)	0.54 [0.20, 0.89]
7.3 Mobility ‐ removal of studies that only provided home‐based programmes	12		Std. Mean Difference (IV, Random, 95% CI)	Subtotals only
7.3.1 Overall	12	1151	Std. Mean Difference (IV, Random, 95% CI)	0.47 [0.24, 0.71]
7.3.2 Removal of studies that only provided home‐based programmes	10	766	Std. Mean Difference (IV, Random, 95% CI)	0.50 [0.20, 0.81]
7.4 Mobility ‐ removal of studies with the longest and shortest intervention periods	12		Std. Mean Difference (IV, Random, 95% CI)	Subtotals only
7.4.1 Overall	12	1151	Std. Mean Difference (IV, Random, 95% CI)	0.47 [0.24, 0.71]
7.4.2 Removal of studies with the longest and shortest intervention	9	727	Std. Mean Difference (IV, Random, 95% CI)	0.46 [0.15, 0.78]
7.5 Mobility ‐ removal of studies with the longest and shortest dose per week	12		Std. Mean Difference (IV, Random, 95% CI)	Subtotals only
7.5.1 Overall	12	1151	Std. Mean Difference (IV, Random, 95% CI)	0.47 [0.24, 0.71]
7.5.2 Removal of studies with the longest and shortest dose per week	8	645	Std. Mean Difference (IV, Random, 95% CI)	0.41 [0.08, 0.74]
7.6 Function ‐ analysis using fixed‐effects	9	916	Std. Mean Difference (IV, Fixed, 95% CI)	0.42 [0.28, 0.55]
7.7 Function ‐ overall vs risk of bias studies removed	9		Std. Mean Difference (IV, Random, 95% CI)	Subtotals only
7.7.1 All studies included	9	916	Std. Mean Difference (IV, Random, 95% CI)	0.60 [0.21, 1.00]
7.7.2 Trials at high or unclear risk of selection bias removed	3	370	Std. Mean Difference (IV, Random, 95% CI)	1.32 [0.05, 2.60]
7.7.3 Trials at high or unclear risk of detection bias removed	3	465	Std. Mean Difference (IV, Random, 95% CI)	0.43 [0.00, 0.85]
7.7.4 All studies included	9	916	Std. Mean Difference (IV, Random, 95% CI)	0.60 [0.21, 1.00]
7.8 Function ‐ removal of studies that only provided home‐based programmes	9		Std. Mean Difference (IV, Random, 95% CI)	Subtotals only
7.8.1 Overall	9	916	Std. Mean Difference (IV, Random, 95% CI)	0.60 [0.21, 1.00]
7.8.2 Removal of studies that only provided home‐based programmes	7	533	Std. Mean Difference (IV, Random, 95% CI)	0.75 [0.20, 1.29]
7.9 Function ‐ removal of studies with the longest and shortest intervention period	9		Std. Mean Difference (IV, Random, 95% CI)	Subtotals only
7.9.1 Overall	9	916	Std. Mean Difference (IV, Random, 95% CI)	0.60 [0.21, 1.00]
7.9.2 Removal of studies with the longest and shortest intervention period	6	494	Std. Mean Difference (IV, Random, 95% CI)	0.67 [0.09, 1.25]
7.10 Function ‐ removal of studies with the longest and shortest dose per week	9		Std. Mean Difference (IV, Random, 95% CI)	Subtotals only
7.10.1 Overall	9	916	Std. Mean Difference (IV, Random, 95% CI)	0.60 [0.21, 1.00]
7.10.2 Removal of studies with the longest and shortest dose per week	5	412	Std. Mean Difference (IV, Random, 95% CI)	0.64 [‐0.03, 1.30]

Open in a new tab

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Binder 2002.

*Study characteristics*
Methods	Study design: randomised controlled trial Number of study arms: 2 Length of follow‐up: 9 months
Participants	Setting: St Louis, USA Number of participants randomised: 119 Number analysed: 115 Number lost to follow‐up: 4 Sample: community‐dwelling males and females aged 78 years and older Age (years): mean 83 (SD 4) Sex: 52% female Inclusion criteria Individuals had to meet at least 2 of the following 3 criteria. Score between 18 and 32 on the Modified Physical Performance Test Report of difficulty or need for assistance with up to 2 IADLs or 1 ADL Achievement of a VO₂ peak between 10 and 18 mL/kg^‐1/m^‐1 Exclusion criteria Did not meet 2 of our 3 frailty criteria Medical conditions that would contraindicate vigorous exercise Neuromuscular disorders unlikely to improve with exercise Chronic use of corticosteroids, immunosuppressive drugs, or androgen‐, oestrogen‐, or progestin‐containing compounds Cigarette use within the previous year Diagnosis of cancer within the previous 5 years Sensory impairments that would interfere with following instructions for testing or exercise training Significant cognitive impairment
Interventions	Intervention Exercise training began with 3 months of flexibility, light‐resistance, and balance training. During the next 3 months, resistance training was added, and during the next 3 months, endurance training was added. The programme was for 60 minutes, 3 times a week, for 9 months delivered in a group environment. Control 9‐month low‐intensity home programmes (stretching), 2 to 3 times per week for 60 minutes
Outcomes	Modified Physical Performance Test Functional Status Questionnaire Single limb stance time (seconds) Berg Balance Score
Duration	9 months
Adherence	Intervention: 23 (33%) dropouts. Participants who remained in the study attended 100% of the sessions. Control: 9 (18%) dropouts
Notes	This work was supported by NIH Claude Pepper Older Americans Independence Center Award Grant P01‐AG13629 and NIH General Clinical Research Center Grant 5‐M01 RR00036.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Study participants were randomly assigned to the intervention or control group in a 3:2 ratio, using a computer‐generated random permutation procedure and a block design.
Allocation concealment (selection bias)	Low risk	The study personnel who maintained the randomisation log were not involved in screening, testing, or training procedures.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants and personnel were not blind to the allocated group.
Blinding of outcome assessment ‐ subjective outcomes (detection bias)	High risk	As participants were not blinded, there is high risk of detection bias for self‐reported outcomes.
Blinding of outcome assessment for assessor‐reported outcomes (detection bias) (Mobility and Function)	High risk	Outcome assessors were not blind to group allocation.
Incomplete outcome data (attrition bias) All outcomes	Low risk	There were a minor amount of missing data (4% in the intervention group; 2% in the control group) (similar numbers across intervention groups).
Selective reporting (reporting bias)	Unclear risk	We did not identify a prespecified outcome document.
Other bias	Low risk	None

Open in a new tab

Bonnefoy 2003.

*Study characteristics*
Methods	Study design: randomised controlled trial Number of study arms: 4 (the 2 groups receiving exercise were compared to the 2 groups not receiving exercise) Length of follow‐up: 9 months
Participants	Setting: Lyon, France Number of participants randomised: 57 Number analysed: 42 Number lost to follow‐up: 15 Sample: participants living in retirement homes Age (years): mean 83 (SD 1.05) Sex: 88% female Inclusion criteria Not described Exclusion criteria Uncontrolled or rapidly evolving diseases Dementia Type 1 diabetes Severe renal insufficiency Functional handicap preventing exercising Long‐term corticosteroid therapy with receipt of vitamin supplements before the study
Interventions	Nutritional supplement (nutritional supplements were given twice daily at 10.00 and 16.00 hours, and consisted of 2 nutritional energy drinks of 200 ml) plus memory (3 sessions per week) Supplement (nutritional supplements were given twice daily at 10.00 and 16.00 hours, and consisted of 2 nutritional energy drinks of 200 ml) plus exercise (exercise of moderate, gradually increasing intensity given 3 times weekly for 60 minutes, with strengthening, balance, and flexibility exercises) Placebo (placebo nutritional supplements) plus exercise (exercise of moderate, gradually increasing intensity given 3 times weekly for 60 minutes, with strengthening, balance, and flexibility exercises) Placebo (placebo nutritional supplements) plus memory (3 sessions per week)
Outcomes	6‐metre walk (seconds) 5 time chair rise 6 stair climb time
Duration	9 months
Adherence	63% compliance with the exercise sessions
Notes	Grants received from INSERM; from le Ministere de l’ Agriculture et de la Recherche (Programme Aliments Demain) and from Laboratoire Diepal 2 x 2 factorial design The paper did not supply standard deviations: for the 6‐metre walk test, we used the standard deviation (2.77) from Langlois 2013; for the 5 time chair rise, we used the standard deviation (6.83) from Gill 2004.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The randomisation was centralised using a remote data entry system.
Allocation concealment (selection bias)	Low risk	Randomisation was centralised.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants and personnel were not blind to all allocated groups.
Blinding of outcome assessment ‐ subjective outcomes (detection bias)	High risk	As participants were not blinded, there is high risk of detection bias for self‐reported outcomes.
Blinding of outcome assessment for assessor‐reported outcomes (detection bias) (Mobility and Function)	Unclear risk	It is unclear if assessors were blinded; there is insufficient information to permit judgement.
Incomplete outcome data (attrition bias) All outcomes	High risk	Greater than 20% of participant data were not available (26%), and there was an unequal dropout rate between groups.
Selective reporting (reporting bias)	Unclear risk	Outcomes identified in the paper were reported (adverse events not reported). We did not identify a prespecified outcome document.
Other bias	Low risk	None

Open in a new tab

Chin A Paw 2002.

*Study characteristics*
Methods	Study design: randomised controlled trial Number of study arms: 4 Length of follow‐up: 17 weeks
Participants	Setting: the Netherlands Number of participants randomised: 217 Number analysed: 157 Number lost to follow‐up: 60 Sample: community dwellers 70 years and older Age (years): mean 78.7 (SD 5.6) Sex: 71% female Inclusion criteria Age 70 or older Use of care services Not participating regularly in physical activity of moderate to high intensity Self‐reported body mass index (BMI) < 25 kg/m2 or involuntary weight loss The ability to understand the study procedures Exclusion criteria Institutionalised Terminal disease or rapidly deteriorating health status Taking multivitamins for the preceding month
Interventions	Supervised exercise: gradually increasing intensity exercises aimed at improving mobility or performance of daily activities, twice a week for 45 minutes for 17 weeks. Exercise was provided in a group setting. Enriched foods: fruit and dairy products enriched with vitamins and minerals usually low in older people's diet No exercise: social programme every 2 weeks for 90 minutes, i.e. lectures, social activities, crafts Regular foods: provided fruit and diary products not enriched with vitamins and minerals Groups (2 x 2 factorial design): Supervised exercise and enriched foods (intervention group 1) Enriched foods and no exercise (control group 1) Supervised exercise and regular foods (intervention group 2) Control ‐ no exercise and regular foods (control group 2)
Outcomes	Performance Score ADL Score Usual gait speed over 6 metres (m/s) Stand up from a chair 5 times (seconds)
Duration	17 weeks
Adherence	Attendance at the exercise programme: median 90% (range 47% to 100%)
Notes	Supported by the Dutch Dairy Federation on Nutrition and Health, Maarsen, and the Dutch Prevention Fund, the Hague, the Netherlands 2 x 2 factorial design
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The method of randomisation was not specified.
Allocation concealment (selection bias)	Low risk	Assignment envelopes were sealed; confirmed envelopes were sequentially numbered and opaque.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants and personnel were not blind to all allocated groups.
Blinding of outcome assessment ‐ subjective outcomes (detection bias)	High risk	As participants were not blinded, there is high risk of detection bias for self‐reported outcomes.
Blinding of outcome assessment for assessor‐reported outcomes (detection bias) (Mobility and Function)	High risk	Quote: "Not all observers could be blinded to exercise or social group assignment."
Incomplete outcome data (attrition bias) All outcomes	High risk	Greater than 20% of participant data were not available (28%): 25 of the 56 dropouts quit during or immediately after the baseline measurements.
Selective reporting (reporting bias)	Unclear risk	Outcomes identified in the paper were reported (adverse events not reported). We did not identify a prespecified outcome document.
Other bias	Low risk	None

Open in a new tab

Fairhall 2012.

*Study characteristics*
Methods	Study design: randomised controlled trial Number of study arms: 2 Length of follow‐up: 12 months
Participants	Setting: Sydney, Australia Number of participants randomised: 241 Number analysed: 216 Number lost to follow‐up: 25 Sample: community‐dwelling people aged 70 or greater Age (years): mean 83.5 (SD 5.9) Sex: 83.5% female Inclusion criteria Adults aged 70 years or older with 3 or more of the Cardiovascular Health Study frailty criteria Not usually living in a residential aged care facility Residing in the Hornsby or Ku‐ring‐gai local government areas Without moderate or severe cognitive impairment (defined as a Mini‐Mental State Examination score of ≤ 18) Not an ongoing client of the Division of Rehabilitation and Aged Care Service Without an illness likely to be associated with a life expectancy of < 12 months, estimated by a score of ≤ 3 on a modified version of the Implicit Illness Severity Scale Not participating in another physical intervention research project Exclusion criteria Residents of nursing care facilities
Interventions	Intervention 12‐month individualised programme targeting identified characteristics of frailty. Those requiring an exercise component received up to 10 home‐based physiotherapy sessions and performed a home exercise programme. The exercise component was based on the Weight Bearing Exercise for Better Balance program. Control The control group received usual care.
Outcomes	Short Physical Performance Battery Barthel Index Walking speed Activity Measure for Post Acute Care Nottingham Extended Activities of daily living (ADL) Scale
Duration	12 months
Adherence	Median adherence was in the category of 26% to 50%.
Notes	The Frailty Intervention Trial was funded by an Australian National Health and Medical Research Council Health Services Research Grant (reference number NHMRC 402791) in 2006.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The study used a computer‐generated random number list.
Allocation concealment (selection bias)	Low risk	The study used central allocation (randomisation list stored off‐site).
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants and personnel were not blind to all allocated groups.
Blinding of outcome assessment ‐ subjective outcomes (detection bias)	High risk	As participants were not blinded, there is high risk of detection bias for self‐reported outcomes.
Blinding of outcome assessment for assessor‐reported outcomes (detection bias) (Mobility and Function)	Low risk	Outcome assessors were blind to group allocation.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Less than 20% of outcome data were missing (10%), and losses were balanced across groups.
Selective reporting (reporting bias)	Low risk	Outcomes were noted in trial registration.
Other bias	Low risk	None

Open in a new tab

Gill 2004.

*Study characteristics*
Methods	Study design: randomised controlled trial Number of study arms: 2 Length of follow‐up: 12 months
Participants	Setting: Connecticut, USA Number of participants randomised: 188 Number analysed: 178 Number lost to follow‐up: 10 Sample: community‐dwelling people aged 75 years or older Age (years): mean 83 (SD 5.1) Sex: 80% female Inclusion criteria Aged 75 years or older Physically frail (determined by 2 tests of physical capability) Exclusion criteria Non‐ambulatory without personal assistance Non‐English‐speaking Nursing home resident Lives outside of greater Bridgeport area or planning to move Enrolled in Wellness Program or participated in pilot‐testing (member of household already enrolled) Diagnosis of dementia or Mini‐Mental State Examination (MMSE) score < 20 Severe visual impairment or hearing loss Progressive, degenerative neurologic disease Terminal illness with life expectancy < 12 months Exercises or too physically active Receiving physical therapy Stroke, hip fracture, or hip or knee replacement within 6 months Myocardial infarction within 6 months
Interventions	Intervention 6‐month home‐based physical therapy programme (including on average 16 home visits) that focused primarily on improving underlying impairments. Detailed algorithms and decision rules were developed to link the results of the assessment with the recommended interventions. Participants were instructed to do their balance exercises once each day and their conditioning exercises 3 days a week. Control The control group received 6 monthly home visits from a trained health educator of 45 to 60 minutes in length.
Outcomes	Modified Physical Performance Test Instrumental Activities of Daily Living Rapid Gait Test (seconds) 3 time chair stands (seconds) Modified Performance‐Oriented Mobility Assessment
Duration	12 months
Adherence	Completion of 73.4%, 78.4%, and 78.7% of the assigned exercises for balance, lower‐extremity conditioning, and upper‐extremity conditioning, respectively
Notes	Supported by the Claude D. Pepper Older Americans Independence Center (grant no. P60AG10469), the National Institute on Aging (award nos. K23AG00759, K24AG021507), and the Gaylord Rehabilitation Research Institute
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation was implemented by the data manager by using a computer‐generated algorithm.
Allocation concealment (selection bias)	Unclear risk	There was insufficient information to permit judgement.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	"Participants unaware of study hypothesis", but it is unclear whether participants were blinded to group allocation.
Blinding of outcome assessment ‐ subjective outcomes (detection bias)	Unclear risk	As it is unclear whether participants were blinded, there is an unclear risk of detection bias for self‐reported outcomes.
Blinding of outcome assessment for assessor‐reported outcomes (detection bias) (Mobility and Function)	Low risk	Quote: "Blinding of the research nurses with respect to the group assignments"
Incomplete outcome data (attrition bias) All outcomes	Low risk	Less than 20% of outcome data were missing (5%), and losses were balanced between groups.
Selective reporting (reporting bias)	Unclear risk	Outcomes identified in the paper were reported (adverse events not reported). We did not identify a prespecified outcome document.
Other bias	Low risk	None

Open in a new tab

Gine‐Garriga 2010.

*Study characteristics*
Methods	Study design: randomised controlled trial Number of study arms: 2 Length of follow‐up: 36 weeks
Participants	Setting: Barcelona, Spain Number of participants randomised: 51 Number analysed: 41 Number lost to follow‐up: 10 Sample: frail community‐dwelling adults Age (years): mean 84 (SD 2.9) Sex: 61% female Inclusion criteria Individuals aged between 80 to 90 who are considered frail Exclusion criteria Not reported
Interventions	Intervention 12‐week functional circuit training programme that focused on functional balance and strength exercises provided in a group setting. Sessions lasted 45 minutes and were twice a week. Control The control group received weekly health education sessions.
Outcomes	Barthel Index Rapid Gait Test (seconds) Stand‐up Test (seconds)
Duration	36 weeks
Adherence	Participants completed 90% of sessions.
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Participants were randomly assigned into 2 groups with the use of a computer‐generated algorithm.
Allocation concealment (selection bias)	Low risk	The study personnel who maintained the randomisation log were not involved in screening, testing, or training procedures.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants and personnel were not blind to all allocated groups.
Blinding of outcome assessment ‐ subjective outcomes (detection bias)	High risk	There is potential risk of bias for adverse events.
Blinding of outcome assessment for assessor‐reported outcomes (detection bias) (Mobility and Function)	High risk	Quote: "Two of three assessors were blinded."
Incomplete outcome data (attrition bias) All outcomes	High risk	Greater than 20% of outcome data were missing, and losses were unbalanced between groups.
Selective reporting (reporting bias)	Unclear risk	Outcomes identified in the paper were reported. We did not identify a prespecified outcome document.
Other bias	Low risk	None

Open in a new tab

Kim 2015.

*Study characteristics*
Methods	Study design: randomised controlled trial Number of study arms: 4 Length of follow‐up: 7 months
Participants	Setting: Tokyo, Japan Number of participants randomised: 131 Number analysed: 126 Number lost to follow‐up: 5 Sample: community‐dwelling frail elderly women over 75 years Age (years): mean 80.9 Sex: 100% female Inclusion criteria Frail Female Aged 74 years and over Exclusion criteria Severe knee or back pain Severely impaired mobility Impaired cognition (Mini‐Mental State Examination (MMSE) score < 24) Missing baseline data Unstable cardiac conditions, such as ventricular dysrhythmias, pulmonary oedema, or other musculoskeletal conditions
Interventions	Exercise: 12 weeks of moderate‐intensity physical comprehensive training programme consisting of strength and gait and balance exercise delivered in a group environment. The classes were twice a week for 60 minutes. Milk fat globule membrane (MFGM): 6 supplement pills ingested in the mornings, prior to activity for 12 weeks Placebo: 6 placebo pills ingested in the mornings, prior to activity for 12 weeks Groups (2 x 2 factorial design): Exercise and MFGM (intervention group 1) MFGM (control group 1) Exercise and placebo (intervention group 2) Placebo (control group 2)
Outcomes	Timed Up and Go (seconds) Usual walking speed (seconds)
Duration	7 months
Adherence	Not reported
Notes	This study was supported by a research grant from the Ministry of Health and Welfare of Japan and a Grant‐in‐Aid for Scientific Research B from the Japan Society for the Promotion of Science (22300243). The authors have the following interests: co‐authors Noriyasu Ota, Akira Shimotoyodome, and Tadashi Hase are employed by Kao Corporation. Kao Corporation provided the MFGM supplementation, and the authors included in this publication affiliated with the Biological Science Laboratories of Kao Corporation analysed the blood samples. 2 x 2 factorial design
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The study used computer‐generated random numbers.
Allocation concealment (selection bias)	Unclear risk	There was insufficient information to permit judgement.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants and personnel were not blind to all allocated groups.
Blinding of outcome assessment ‐ subjective outcomes (detection bias)	High risk	As participants were not blinded, there is high risk of detection bias for self‐reported outcomes.
Blinding of outcome assessment for assessor‐reported outcomes (detection bias) (Mobility and Function)	Low risk	Data collection was conducted by separate physical therapy staff members who were also blind to the allocation of treatments.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Less than 20% of outcome data were missing (4%), and losses were balanced across groups.
Selective reporting (reporting bias)	High risk	ADL was included as a secondary outcome in trial registration, but the measure was not included in the results paper.
Other bias	Low risk	None

Open in a new tab

Langlois 2013.

*Study characteristics*
Methods	Study design: matched control Number of study arms: 2 Length of follow‐up: 12 weeks
Participants	Setting: Montreal, Canada Number of participants randomised: 34 Number analysed: 29 Number lost to follow‐up: 5 Sample: community dwellers aged 61 to 89 Age (years): mean 74.9 Sex: 74% female Inclusion criteria Not described Exclusion criteria Limitations to undertake a physical exercise programme Signs of dementia (< 25 at the Mini‐Mental State Examination) Depression (> 10 on the Geriatric Depression Scale)
Interventions	Intervention 12 weeks of a combination of aerobic, gait, and balance exercises. The classes were 3 times a week for 60 minutes. Control The control group received no intervention.
Outcomes	Modified Physical Performance Test Timed Up and Go (seconds) 6‐minute walk test (metres) 10‐metre walk test (seconds)
Duration	12 weeks
Adherence	Not reported
Notes	Study included both frail and non‐frail participants. Only included the cohort in the data that were considered frail
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The method of randomisation was not specified.
Allocation concealment (selection bias)	Unclear risk	There was insufficient information to permit judgement.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants and personnel were not blind to all allocated groups.
Blinding of outcome assessment ‐ subjective outcomes (detection bias)	High risk	As participants were not blinded, there is high risk of detection bias for self‐reported outcomes.
Blinding of outcome assessment for assessor‐reported outcomes (detection bias) (Mobility and Function)	Unclear risk	There was insufficient information to permit judgement.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Less than 20% of outcome data were missing (13%), and losses were balanced between groups.
Selective reporting (reporting bias)	Unclear risk	Outcomes identified in the paper were reported (adverse events not reported). We did not identify a prespecified outcome document.
Other bias	Low risk	None

Open in a new tab

Mugueta‐Aguinaga 2017.

*Study characteristics*
Methods	Study design: randomised controlled trial Number of study arms: 2 Length of follow‐up: 3 weeks
Participants	Setting: Barakaldo, Spain Number of participants randomised: 40 Number analysed: 39 Number lost to follow‐up: 1 Sample: frail participants who attend an elderly day care centre Age (years): mean 84.3 Sex: 60% female Inclusion criteria Persons aged 65 years of age with a Barthel Index score equal to or above 90 points who carry out no scheduled physical activity Exclusion criteria Persons over 65 years of age with a Barthel Index score less than 90 points or with a Barthel Index score equal to or above 90 points who carry out scheduled physical activity
Interventions	Intervention 3 weeks of exercise delivered via an interactive videogame (exergame) focusing on functional, gait, and balance tasks supervised directly by a therapist. The intervention was 3 times a week for 20 minutes. Control The control group received no intervention.
Outcomes	Short Physical Performance Battery
Duration	3 weeks
Adherence	Participants attended 100% of the intervention sessions.
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The method of randomisation was not specified.
Allocation concealment (selection bias)	Unclear risk	There was insufficient information to permit judgement.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants and personnel were not blind to all allocated groups.
Blinding of outcome assessment ‐ subjective outcomes (detection bias)	High risk	As participants were not blinded, there is high risk of detection bias for self‐reported outcomes.
Blinding of outcome assessment for assessor‐reported outcomes (detection bias) (Mobility and Function)	Unclear risk	There was insufficient information to permit judgement.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Less than 20% of outcome data were missing (3%), and losses were balanced between groups.
Selective reporting (reporting bias)	Unclear risk	We did not identify a prespecified outcome document.
Other bias	Low risk	None

Open in a new tab

Rydwik 2008.

*Study characteristics*
Methods	Study design: randomised controlled trial Number of study arms: 4 Length of follow‐up: 9 months
Participants	Setting: Stockholm, Sweden Number of participants randomised: 96 Number analysed: 65 Number lost to follow‐up: 31 Sample: frail community‐dwelling elderly people over the age of 75 Age (years): mean 83 Sex: 60% female Inclusion criteria Uninintentional weight loss ≥ 5% or body mass index (BMI) ≤ 20 kg/m2, or both Low physical activity level (≤ grade 3 on a 6‐graded scale of physical activity) Exclusion criteria Age under 75 BMI > 30 kg/m2 Non‐walkers People with recent cardiac problems requiring hospital care Recent hip fracture or surgery during the last 6 months Present cancer treatment Stroke within the last 2 years Less than 7 points of a total 9‐point score on the short form of the Mini‐Mental State Examination Institutionalisation
Interventions	Physical training: 12 weeks of physical training programme consisting of aerobic, muscle strength, balance exercises delivered in a group environment. The classes were twice a week for 60 minutes. This was followed by 6 months of unsupervised home‐based exercises (participants were encouraged to perform balance and functional muscle strength training and to take regular walks several times per week). Nutrition: specific individualised diet counselling and group session education Control: general physical training advice and general diet advice Groups (2 x 2 factorial design): Physical Training (intervention group 1) Control (control group 1) Physical Training and Nutrition (intervention group 2) Nutrition (control group 2)
Outcomes	Timed Up and Go (seconds) Maximal walking speed (m/s) Single leg stance (seconds) Functional Independence Measure Instrumental Activities of Daily Living (Instrumental Activities Measure)
Duration	9 months
Adherence	Mean compliance was 65%.
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Participants were randomised consecutively in batches to 4 different groups. For each new batch, randomisation started with the oldest individual to avoid age differences between groups.
Allocation concealment (selection bias)	High risk	Study personnel conducted the randomisation procedure in an open manner.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants and personnel were not blind to all allocated groups.
Blinding of outcome assessment ‐ subjective outcomes (detection bias)	High risk	As participants were not blinded, there is high risk of detection bias for self‐reported outcomes.
Blinding of outcome assessment for assessor‐reported outcomes (detection bias) (Mobility and Function)	High risk	Neither the test leaders nor the participants were blinded.
Incomplete outcome data (attrition bias) All outcomes	High risk	Greater than 20% of outcome data were missing (33%), and losses were unbalanced between groups.
Selective reporting (reporting bias)	Unclear risk	Outcomes identified in the paper were reported (adverse events not reported). We did not identify a prespecified outcome document.
Other bias	Low risk	None

Open in a new tab

Tarazona‐Santabalbina 2016.

*Study characteristics*
Methods	Study design: randomised controlled trial Number of study arms: 2 Length of follow‐up: 24 weeks
Participants	Setting: Alencia, Spain Number of participants randomised: 100 Number analysed: 82 Number lost to follow‐up: 18 Sample: community‐dwelling frail older adults Age (years): mean 80 Sex: 54% female Inclusion criteria Men and women aged 70 years or older who were: sedentary (less than 3 hours of weekly physical activity) frail according to the frailty phenotype with a gait speed slower than 0.8 m/s, and were community dwellers Exclusion criteria Life expectancy less than 6 months by any cause A value of 7c‐7d in the Global Disability Score (GGS‐FAST) in cognitive‐impaired people (score < 24 in the Mini‐Mental State Examination) Severe disability (score < 15 points on the Barthel Index or an E score or higher in the Katz Scale) Ejection fraction left ventricle 20% or less Hospital admission in the past 3 months for any reason Oncologic patient with active treatment with chemotherapy or radiotherapy Major non‐ambulatory surgery in the past 6 months before the beginning of the study Family member centenarian in the previous 2 generations Person with a coronary event in the past 12 months Institutionalised person Impossibility of going to the primary care centre when using their own means of transport
Interventions	Intervention A multicomponent exercise programme consisting of endurance, strength, co‐ordination, balance, and flexibility exercises that have the potential to impact a variety of functional performance measures. Exercises were performed for 65 minutes, 5 days per week for 24 weeks in a group setting. Control The control group received no training.
Outcomes	Short Physical Performance Battery Physical Performance Test Barthel Index Lawton Scale
Duration	24 weeks
Adherence	Not reported
Notes
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The method of randomisation was not specified.
Allocation concealment (selection bias)	Unclear risk	There was insufficient information to permit judgement.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants and personnel were not blind to all allocated groups.
Blinding of outcome assessment ‐ subjective outcomes (detection bias)	High risk	As participants were not blinded, there is high risk of detection bias for self‐reported outcomes.
Blinding of outcome assessment for assessor‐reported outcomes (detection bias) (Mobility and Function)	Low risk	Researchers responsible for data gathering were blinded for this study.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Less than 20% of outcome data were missing (18%), and losses were balanced across groups.
Selective reporting (reporting bias)	High risk	Did not report all outcomes (Timed Up and Go) due to incorrect recording
Other bias	Low risk	None

Open in a new tab

Worm 2001.

*Study characteristics*
Methods	Study design: randomised controlled trial Number of study arms: 2 Length of follow‐up: 12 weeks
Participants	Setting: Odense, Denmark Number of participants randomised: 46 Number analysed: 44 Number lost to follow‐up: 2 Sample: community‐dwelling frail people Age (years): mean 81.2 Sex: 57% female Inclusion criteria Above 74 years Living in their own home and not able to leave their home unaided or unattended or without mobility aids Exclusion criteria People with a life‐threatening illness Symptom‐giving somatic disease Confined to bed
Interventions	Intervention Flexibility training, aerobics, rhythm, balance, and reaction exercises and muscle training (strength and endurance). Exercises and games involving complex functional movements were primarily used. The training group also performed a daily short (8 to 10 minutes) home‐based programme including muscle and flexibility training. Control The control group received no training.
Outcomes	10‐metre walk test (seconds) Berg Balance Scale
Duration	12 weeks
Adherence	91.9%
Notes	Standard deviations were not supplied in the text: for the 10‐metre walk test, the standard deviation (2.77) from Langlois 2013 was used; for the 5‐time chair rise, the standard deviation (6.83) from Gill 2004 was used; for the Berg Balance Test, the standard deviation (4.1) from Binder 2002 was used.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	The method of randomisation was not specified.
Allocation concealment (selection bias)	Unclear risk	There was insufficient information to permit judgement.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants and personnel were not blind to all allocated groups.
Blinding of outcome assessment ‐ subjective outcomes (detection bias)	High risk	As participants were not blinded, there is high risk of detection bias for self‐reported outcomes.
Blinding of outcome assessment for assessor‐reported outcomes (detection bias) (Mobility and Function)	Unclear risk	There was insufficient information to permit judgement.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Less than 20% of outcome data were missing (4%), and losses were balanced between groups.
Selective reporting (reporting bias)	Unclear risk	Outcomes identified in the paper were reported (adverse events not reported). We did not identify a prespecified outcome document.
Other bias	Low risk	None

Open in a new tab

SD: standard deviation

IADL: Instrumental Activities of Daily Living

ADL: Activities of Daily Living

NIH: National Institutes of Health

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Beling 2009	The study did not meet our criteria for participants.
Binder 2005	The only outcome measured was body fat‐free mass, which is not an outcome of interest in this review.
Clegg 2014	The study did not meet our criteria for participants.
De Vries 2015	The study did not meet our criteria for the control group.
Faber 2006	The study did not meet our criteria for participants.
Halvarsson 2013	The study did not meet our criteria for participants.
Hiroyuki 2003	The study did not meet our criteria for participants.
Ikezoe 2005	The study did not meet our criteria for setting.
Latham 2003	The study did not meet our criteria for the intervention.
Law 2018	The study did not meet our criteria for participants.
Lord 2003	The study did not meet our criteria for participants.
Ng 2014	The study did not meet our criteria for participants.
Pahor 2014	The study did not meet our criteria for participants.
Park 2012	The study did not meet our criteria for participants.
Salem 2017	The study did not meet our criteria for the intervention.
Seino 2017	The study did not meet our criteria for participants.
Trombetti 2018	The study did not meet our criteria for participants.

Open in a new tab

Differences between protocol and review

Compared with the protocol (Fairhall 2013), there was a reordering in the priority of mobility and function scores.

Mobility: Short Physical Performance Battery, Timed Up and Go test, walking distance, walking speed, timed sit to stand, Berg Balance Scale, single leg stand time.
Functioning: Barthel Index, Physical Performance Test, Lawton Independent Activities of Daily Living Index, Functional Independence Measure, Performance Sum Score, Activity of Daily Living (ADL) score, Instrumental Activities of Daily Living.

DT and NF decided the priority order of the outcomes. The order of priority for mobility outcomes was determined on the basis of the mobility outcome measure that reflects the most components of everyday mobility. The order of priority for functional outcomes was determined on the basis of the functional outcome measure that encompasses the most activities and participation domains of the International Classification of Functioning, Disability and Health (ICF). The SPPB was chosen as it is widely used in practice and has been shown to be predictive of an increased risk of falling, loss of independence in activities of daily living, decreased mobility, disability, decline in health, re‐hospitalisation, increased hospital length of stay, nursing home admission, and death (Treacy 2018).

We did not perform an analysis on costs as only one study reported these data.

We chose a new subgroup analysis (inclusion and exclusion of participants with cognitive impairment) after the protocol was written. People with cognitive impairment, such as dementia, are not routinely offered rehabilitation services and are frequently excluded from studies (Cations 2017; Laver 2020). We felt it was important to understand whether similar improvements are seen when studies include individuals with cognitive impairment.

We did not complete subgroup analysis based on adherence because of the low number of trials that reported the same methods for adherence.

Contributions of authors

Nicola Fairhall, Catherine Sherrington, and Ian Cameron developed the Cochrane Review protocol. Daniel Treacy and Nicola Fairhall were responsible for co‐ordinating the review, carrying out the searches, and locating studies. They decided independently, and then by consensus, which studies met the inclusion criteria. Daniel Treacy, Nicola Fairhall, and Karl Schurr assessed quality and extracted data from the included studies. Daniel Treacy drafted the review and performed primary data entry and analysis into RevMan 5.4. Nicola Fairhall, Catherine Sherrington, and Ian Cameron provided guidance with this process. All authors were involved in revisions at all stages.

Sources of support

Internal sources

No sources of support provided

External sources

National Health and Medical Research Council, Australia

Salary (NF, CS, IC)

Declarations of interest

DT: has declared that they have no conflict of interest. LH: has declared that they have no conflict of interest. KS: I teach two‐ and three‐day workshops for physiotherapists and occupational therapists for which I am paid a fee by various health services and individuals. NF: has declared that they have no conflict of interest. IC: has declared that they have no conflict of interest. CS: has declared that they have no conflict of interest. NF, IC, and CS were authors of the included study Fairhall 2012. These authors did not review this study.

New

References

References to studies included in this review

Binder 2002 {published data only}

Binder EF, Schechtman KB, Ehsani AA, Steger-May K, Brown M, Sinacore DR, et al. Effects of exercise training on frailty in community-dwelling older adults: results of a randomized, controlled trial. Journal of the American Geriatrics Society 2002;50(12):1921-8. [DOI] [PubMed] [Google Scholar]
Brown M, Sinacore DR, Ehsani AA, Binder EF, Holloszy JO, Kohrt WM. Low-intensity exercise as a modifier of physical frailty in older adults. Archives of Physical Medicine and Rehabilitation 2000;81(7):960-5. [DOI] [PubMed] [Google Scholar]

Bonnefoy 2003 {published data only}

Bonnefoy M, Cornu C, Normand S, Boutitie F, Bugnard F, Rahmani A, et al. The effects of exercise and protein-energy supplements on body composition and muscle function in frail elderly individuals: a long-term controlled randomised study. British Journal of Nutrition 2003;89(5):731-9. [DOI] [PubMed] [Google Scholar]
Normand S. Clarification of study design [personal communication]. Email to: Daniel Treacy 17 April 2018.

Chin A Paw 2002 {published and unpublished data}

Chin A Paw MJ, De Jong N, Schouten EG, Hiddink GJ, Kok FJ. Physical exercise and/or enriched foods for functional improvement in frail, independently living elderly: a randomized controlled trial. Archives of Physical Medicine and Rehabilitation 2001;82(6):811-7. [DOI] [PubMed] [Google Scholar]
Chin A Paw MJ, De Jong N, Schouten EG, Van Staveren WA, Kok FJ. Physical exercise or micronutrient supplementation for the wellbeing of the frail elderly? A randomised controlled trial. British Journal of Sports Medicine 2002;36(2):126-31. [DOI] [PMC free article] [PubMed] [Google Scholar]
Chin A Paw MJ. Additional outcome data [personal communication]. Email to: Daniel Treacy 8 January 2019.

Fairhall 2012 {published data only}

Cameron ID, Fairhall N, Langron C, Lockwood K, Monaghan N, Aggar C, et al. A multifactorial interdisciplinary intervention reduces frailty in older people: randomized trial. BMC Medicine 2013;11:65. [DOI] [PMC free article] [PubMed] [Google Scholar]
Fairhall N, Sherrington C, Kurrle SE, Lord SR, Lockwood K, Cameron ID. Effect of a multifactorial interdisciplinary intervention on mobility-related disability in frail older people: randomised controlled trial. BMC Medicine 2012;10:120. [DOI: 10.1186/1741-7015-10-120] [DOI] [PMC free article] [PubMed] [Google Scholar]
Fairhall N, Sherrington C, Kurrle SE, Lord SR, Lockwood K, Howard K, et al. Economic evaluation of a multifactorial, interdisciplinary intervention versus usual care to reduce frailty in frail older people. Journal of the American Medical Directors Association 2015;16(1):41-8. [DOI] [PubMed] [Google Scholar]
Fairhall N, Sherrington C, Lord SR, Kurrle SE, Langron C, Lockwood K, et al. Effect of a multifactorial, interdisciplinary intervention on risk factors for falls and fall rate in frail older people: a randomised controlled trial. Age and Ageing 2014;43(5):616-22. [DOI] [PubMed] [Google Scholar]

Gill 2004 {published data only}

Gill TM, Baker DI, Gottschalk M, Gahbauer EA, Charpentier PA, De Regt PT, et al. A prehabilitation program for physically frail community-living older persons. Archives of Physical Medicine and Rehabilitation 2003;84(3):394-404. [DOI] [PubMed] [Google Scholar]
Gill TM, Baker DI, Gottschalk M, Peduzzi PN, Allore H, Byers A. A program to prevent functional decline in physically frail, elderly persons who live at home. New England Journal of Medicine 2002;347(14):1068-74. [DOI] [PubMed] [Google Scholar]
Gill TM, Baker DI, Gottschalk M, Peduzzi PN, Allore H, Van Ness PH. A prehabilitation program for the prevention of functional decline: effect on higher-level physical function. Archives of Physical Medicine and Rehabilitation 2004;85(7):1043-9. [DOI] [PubMed] [Google Scholar]
Gill TM, McGloin JM, Gahbauer EA, Shepard DM, Bianco LM. Two recruitment strategies for a clinical trial of physically frail community-living older persons. Journal of the American Geriatrics Society 2001;49(8):1039-45. [DOI] [PubMed] [Google Scholar]

Gine‐Garriga 2010 {published data only}

Gine-Garriga M, Guerra M, Pages E, Manini TM, Jimenez R, Unnithan VB. The effect of functional circuit training on physical frailty in frail older adults: a randomized controlled trial. Journal of Aging and Physical Activity 2010;18(4):401-24. [DOI] [PubMed] [Google Scholar]
Gine-Garriga M, Guerra M, Unnithan VB. The effect of functional circuit training on self-reported fear of falling and health status in a group of physically frail older individuals: a randomized controlled trial. Aging Clinical and Experimental Research 2013;25(3):329-36. [DOI] [PubMed] [Google Scholar]

Kim 2015 {published data only}

Kim H, Suzuki T, Kim M, Kojima N, Ota N, Shimotoyodome A, et al. Effects of exercise and milk fat globule membrane (MFGM) supplementation on body composition, physical function, and hematological parameters in community-dwelling frail Japanese women: a randomized double blind, placebo-controlled, follow-up trial. PLoS One 2015;10(2):e0116256. [DOI] [PMC free article] [PubMed] [Google Scholar]

Langlois 2013 {published and unpublished data}

Desjardins L. Additional outcome data [personal communication]. Email to: Daniel Treacy 2 June 2018.
Langlois F, Vu TT, Chasse K, Dupuis G, Kergoat MJ, Bherer L. Benefits of physical exercise training on cognition and quality of life in frail older adults. Journals of Gerontology Series B: Psychological Sciences and Social Sciences 2013;68(3):400-4. [DOI] [PubMed] [Google Scholar]

Mugueta‐Aguinaga 2017 {published data only}

Mugueta-Aguinaga I, Garcia-Zapirain B. FRED: Exergame to prevent dependence and functional deterioration associated with ageing. A pilot three-week randomized controlled clinical trial. International Journal of Environmental Research and Public Health 2017;14:1439. [DOI] [PMC free article] [PubMed] [Google Scholar]

Rydwik 2008 {published and unpublished data}

Rydwik E, Frandin K, Akner G. Effects of a physical training and nutritional intervention program in frail elderly people regarding habitual physical activity level and activities of daily living - a randomized controlled pilot study. Archives of Gerontology and Geriatrics 2010;51(3):283-9. [DOI] [PubMed] [Google Scholar]
Rydwik E, Gustafsson T, Frandin K, Akner G. Effects of physical training on aerobic capacity in frail elderly people (75+ years). Influence of lung capacity, cardiovascular disease and medical drug treatment: a randomized controlled pilot trial. Aging Clinical and Experimental Research 2010;22(1):85-94. [DOI] [PubMed] [Google Scholar]
Rydwik E, Lammes E, Frandin K, Akner G. Effects of a physical and nutritional intervention program for frail elderly people over age 75. A randomized controlled pilot treatment trial. Aging Clinical and Experimental Research 2008;20(2):159-70. [DOI] [PubMed] [Google Scholar]
Rydwik E. Additional outcome data [personal communication]. Email to: Daniel Treacy 16 January 2019.

Tarazona‐Santabalbina 2016 {published data only}

Tarazona-Santabalbina FJ, Gomez-Cabrera MC, Perez-Ros P, Martinez-Arnau FM, Cabo H, Tsaparas K, et al. A multicomponent exercise intervention that reverses frailty and improves cognition, emotion, and social networking in the community-dwelling frail elderly: a randomized clinical trial. Journal of the American Medical Directors Association 2016;17(5):426-33. [DOI] [PubMed] [Google Scholar]

Worm 2001 {published data only}

Worm CH, Vad E, Puggaard L, Stovring H, Lauritsen J, Kragstrup J. Effects of a multicomponent exercise program on functional ability in community-dwelling, frail older adults. Journal of Aging and Physical Activity 2001;9(4):414-24. [Google Scholar]

References to studies excluded from this review

Beling 2009 {published data only}

Beling J, Roller M. Multifactorial intervention with balance training as a core component among fall-prone older adults. Journal of Geriatric Physical Therapy 2009;32(3):125-33. [DOI] [PubMed] [Google Scholar]

Binder 2005 {published data only}

Binder E F, Yarasheski K E, Steger-May K, Sinacore D R, Brown M, Schechtman K B, et al. Effects of progressive resistance training on body composition in frail older adults: results of a randomized, controlled trial. Journals of Gerontology Series A: Biological Sciences and Medical Sciences 2005;60(11):1425-31. [DOI] [PubMed] [Google Scholar]

Clegg 2014 {published data only}

Clegg A, Barber S, Young J, Iliffe S, Forster A. The Home-based Older People's Exercise (HOPE) trial: a pilot randomised controlled trial of a home-based exercise intervention for older people with frailty. Age and Ageing 2014;43(5):687-95. [DOI] [PMC free article] [PubMed] [Google Scholar]

De Vries 2015 {published data only}

De Vries NM, Staal JB, Van Der Wees PJ, Akkermans R, Adang E, Olde Rikkert MGM, et al. Personalized physiotherapy in frail older adults with mobility problems is (cost)-effective in improving physical activity and frailty: a RCT. Physiotherapy (United Kingdom) 2015;101:eS1089-90. [Google Scholar]

Faber 2006 {published data only}

Faber MJ, Bosscher RJ, Chin A Paw MJ, Van Wieringen PC. Effects of exercise programs on falls and mobility in frail and pre-frail older adults: a multicenter randomized controlled trial. Archives of Physical Medicine and Rehabilitation 2006;87(7):885-96. [DOI] [PubMed] [Google Scholar]

Halvarsson 2013 {published data only}

Halvarsson A, Franzen E, Faren E, Olsson E, Oddsson L, Stahle A. Long-term effects of new progressive group balance training for elderly people with increased risk of falling - a randomized controlled trial. Clinical Rehabilitation 2013;27(5):450-8. [DOI] [PubMed] [Google Scholar]

Hiroyuki 2003 {published data only}

Hiroyuki S, Uchiyama Y, Kakurai S. Specific effects of balance and gait exercises on physical function among the frail elderly. Clinical Rehabilitation 2003;17(5):472-9. [DOI] [PubMed] [Google Scholar]

Ikezoe 2005 {published data only}

Ikezoe T, Tsutou A, Asakawa Y, Tsuboyama T. Low intensity training for frail elderly women: long-term effects on motor function and mobility. Journal of Physical Therapy Science 2005;17(1):43-9. [Google Scholar]

Latham 2003 {published data only}

Latham NK, Anderson CS, Lee A, Bennett DA, Moseley A, Cameron ID, et al. A randomized, controlled trial of quadriceps resistance exercise and vitamin D in frail older people: the Frailty Interventions Trial in Elderly Subjects (FITNESS). Journal of the American Geriatrics Society 2003;51(3):291-9. [DOI] [PubMed] [Google Scholar]

Law 2018 {published data only}

Law H, Bte Ismail N, Wong W, Ramason R, Kua J, Natesan S. Targeting falls through frailty intervention by a combined nutritional intervention and physical exercise programme in community-dwelling older fallers: a pilot randomized clinical trial. Journal of the American Geriatrics Society 2018;66(S2):S128. [Google Scholar]

Lord 2003 {published data only}

Lord S R, Castell S, Corcoran J, Dayhew J, Matters B, Shan A, et al. The effect of group exercise on physical functioning and falls in frail older people living in retirement villages: a randomized, controlled trial. Journal of the American Geriatrics Society 2003;51(12):1685-92. [DOI] [PubMed] [Google Scholar]

Ng 2014 {published data only}

Ng TP, Lim LM, Niti M, Nyunt MSZ, Feng L, Gwee X, et al. Effects of nutritional, physical, cognitive interventions on cognitive outcomes in the Singapore frailty intervention trial (S-FIT). Annals of the Academy of Medicine Singapore 2014;1:S68-9. [Google Scholar]

Pahor 2014 {published data only}

Pahor M, Guralnik JM, Ambrosius WT, Blair S, Bonds DE, Church TS, et al. Effect of structured physical activity on prevention of major mobility disability in older adults: the LIFE study randomized clinical trial. Journal of the American Medical Association 2014;311(23):2387-96. [DOI] [PMC free article] [PubMed] [Google Scholar]

Park 2012 {published data only}

Park YI, Lee KY, Kim TI, Jeon MH, Kim DO, Kim JH. The effects of exercise in the frail elderly. Journal of Korean Academy of Community Health Nursing 2012;23(1):91-101. [Google Scholar]

Salem 2017 {published data only}

Salem BE, Ma-Pham J, Chen S, Brecht ML, Antonio AL, Ames M. Impact of a community-based frailty intervention among middle-aged and older prefrail and frail homeless women: a pilot randomized controlled trial. Community Mental Health Journal 2017;53(6):688-94. [DOI] [PMC free article] [PubMed] [Google Scholar]

Seino 2017 {published data only}

Seino S, Nishi M, Murayama H, Narita M, Yokoyama Y, Nofuji Y, et al. Effects of a multifactorial intervention comprising resistance exercise, nutritional and psychosocial programs on frailty and functional health in community-dwelling older adults: a randomized, controlled, cross-over trial. Geriatrics & Gerontology International 2017;17(11):2034-45. [DOI] [PubMed] [Google Scholar]

Trombetti 2018 {published data only}

Trombetti A, Hars M, Hsu FC, Reid KF, Church TS, Gill TM, et al. Effect of physical activity on frailty: secondary analysis of a randomized controlled trial. Annals of Internal Medicine 2018;168(5):309-16. [DOI] [PMC free article] [PubMed] [Google Scholar]

Additional references

Allen 2011

Allen NE, Sherrington C, Paul SS, Canning CG. Balance and falls in Parkinson's disease: a meta-analysis of the effect of exercise and motor training. Movement Disorders 2011;26(9):1605-15. [PMID: ] [DOI] [PubMed] [Google Scholar]

Ashworth 2005

Ashworth NL, Chad KE, Harrison EL, Reeder BA, Marshall SC. Home versus center based physical activity programs in older adults. Cochrane Database of Systematic Reviews 2005, Issue 1. Art. No: CD004017. [DOI: 10.1002/14651858.CD004017.pub2] [DOI] [PMC free article] [PubMed] [Google Scholar]

Binder 2002

Binder EF, Schechtman KB, Ehsani AA, Steger-May K, Brown M, Sinacore DR, et al. Effects of exercise training on frailty in community-dwelling older adults: results of a randomized, controlled trial. Journal of the American Geriatrics Society 2002;50(12):1921-8. [DOI] [PubMed] [Google Scholar]

Binder 2004

Binder EF, Brown M, Sinacore DR, Steger-May K, Yarasheski KE, Schechtman KB. Effects of extended outpatient rehabilitation after hip fracture: a randomized controlled trial. Journal of the American Medical Association 2004;292(7):837-46. [DOI] [PubMed] [Google Scholar]

Cameron 2018

Cameron ID, Dyer SM, Panagoda CE, Murray GR, Hill KD, Cumming RG, et al. Interventions for preventing falls in older people in care facilities and hospitals. Cochrane Database of Systematic Reviews 2018, Issue 9. Art. No: CD005465. [DOI: 10.1002/14651858.CD005465.pub4] [DOI] [PMC free article] [PubMed] [Google Scholar]

Cates 2008 [Computer program]

Visual Rx. Cates C, Version 3. Dr. Christopher Cates EBM website. Available at www.nntonline.net, 2008.

Cations 2017

Cations M, Laver KE, Crotty M, Cameron ID. Rehabilitation in dementia care. Age and Ageing 2017;47(2):171-74. [DOI] [PubMed] [Google Scholar]

Chin 2001

Chin A Paw MJ, De Jong N, Schouten EG, Hiddink GJ, Kok FJ. Physical exercise and/or enriched foods for functional improvement in frail, independently living elderly: a randomized controlled trial. Archives of Physical Medicine and Rehabilitation 2001;82(6):811-7. [DOI] [PubMed] [Google Scholar]

Chou 2012

Chou CH, Hwang CL, Wu YT. Effect of exercise on physical function, daily living activities, and quality of life in the frail older adults: a meta-analysis. Archives of Physical Medicine and Rehabilitation 2012;93:237-44. [DOI] [PubMed] [Google Scholar]

Clegg 2013

Clegg A, Young J, Iliffe S, Rikkert MO, Rockwood K. Frailty in elderly people. Lancet 2013;381(9868):752-62. [DOI] [PMC free article] [PubMed] [Google Scholar]

Clegg 2014

Clegg A, Barber S, Young J, Ilffe S, Forster A. The Homebased Older People's Exercise (HOPE) trial: a pilot randomised controlled trial of a home-based exercise intervention for older people with frailty. Age and Ageing 2014;43(5):687-95. [DOI] [PMC free article] [PubMed] [Google Scholar]

Daniels 2008

Daniels R, Van Rossum E, De Witte L, Kempen GI, Van den Heuvel W. Interventions to prevent disability in frail community-dwelling elderly: a systematic review. BMC Health Services Research 2008;8:278. [DOI] [PMC free article] [PubMed] [Google Scholar]

Deeks 2020

Deeks JJ, Higgins JP, Altman DG. Chapter 10: Analysing data and undertaking meta-analyses. In: Higgins JP, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA, editor(s). Cochrane Handbook for Systematic Reviews of Interventions Version 6.1 (updated September 2020). Cochrane, 2020. Available from www.training.cochrane.org/handbook.

De Morton 2007

De Morton N, Keating JL, Jeffs K. Exercise for acutely hospitalised older medical patients. Cochrane Database of Systematic Reviews 2007, Issue 1. Art. No: CD005955. [DOI: 10.1002/14651858.CD005955.pub2] [DOI] [PMC free article] [PubMed] [Google Scholar]

Dent 2017

Dent E, Kowal P, Hoogendijk EO. Frailty measurement in research and clinical practice: a review. European Journal of Internal Medicine 2016;31:3-10. [DOI] [PubMed] [Google Scholar]

Dent 2019

Dent E, Morley JE, Cruz-Jentoft AJ, Woodhouse L, Rodriguez-Manas L, Fried LP, et al. Physical frailty: ICFSR international clinical practice guidelines for identification and management. Journal of Nutrition, Health and Aging 2019;23(9):771-87. [DOI] [PMC free article] [PubMed] [Google Scholar]

Fairhall 2013

Fairhall N, Sherrington C, Cameron I. Mobility training for increasing mobility and functioning in older people with frailty. Cochrane Database of Systematic Reviews 2013, Issue 5. Art. No: CD010494. [DOI: 10.1002/14651858.CD010494] [DOI] [PMC free article] [PubMed] [Google Scholar]

Fiatarone 1994

Fiatarone MA, O'Neill EF, Ryan ND, Clements KM, Solares GR, Nelson ME, et al. Exercise training and nutritional supplementation for physical frailty in very elderly people. New England Journal of Medicine 1994;330(25):1769-75. [DOI] [PubMed] [Google Scholar]

Fried 2001

Fried LP, Tangen CM, Walston J, Newman AB, Hirsch C, Gottdiener J, et al. Frailty in older adults: evidence for a phenotype. Journals of Gerontology Series A: Biological Sciences and Medical Sciences 2001;56(3):M146-56. [DOI] [PubMed] [Google Scholar]

Fried 2004

Fried LP, Ferrucci L, Darer J, Williamson JD, Anderson G. Untangling the concepts of disability, frailty, and comorbidity: implications for improved targeting and care. Journals of Gerontology Series A: Biological Sciences and Medical Sciences 2004;59(3):255-63. [DOI] [PubMed] [Google Scholar]

Gill 1995

Gill TM, Williams CS, Tinetti ME. Assessing risk for the onset of functional dependence among older adults: the role of physical performance. Journal of the American Geriatrics Society 1995;43(6):603-9. [DOI] [PubMed] [Google Scholar]

Gill 2002

Gill TM, Baker DI, Gottschalk M, Peduzzi PN, Allore H, Byers A. A program to prevent functional decline in physically frail, elderly persons who live at home. New England Journal of Medicine 2002;347(14):1068-74. [DOI] [PubMed] [Google Scholar]

Gill 2004

Gill TM, Allore HG, Holford TR, Guo Z. Hospitalization, restricted activity, and the development of disability among older persons. JAMA 2004;292(17):2115-24. [DOI] [PubMed] [Google Scholar]

Gill 2006

Gill TM, Gahbauer EA, Allore HG, Han L. Transitions between frailty states among community-living older persons. Archives of Internal Medicine 2006;166(4):418-23. [DOI] [PubMed] [Google Scholar]

Gine‐Garriga 2014

Gine-Garriga M, Roque-Fıguls M, Coll-Planas L, Sitja-Rabert M, Salva A. Physical exercise interventions for improving performance-based measures of physical function in community-dwelling, frail older adults: a systematic review and meta-analysis. Archives of Physical Medicine and Rehabilitation 2014;95:753-69. [DOI] [PubMed] [Google Scholar]

GRADEpro GDT [Computer program]

GRADEpro GDT. Version accessed 6 August 2021. Hamilton (ON): McMaster University (developed by Evidence Prime). Available at gradepro.org.

Higgins 2020

Higgins JP, Savović J, Page MJ, Elbers RG, Sterne JAC. Chapter 8: Assessing risk of bias in a randomized trial. In: Higgins JP, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA, editor(s). Cochrane Handbook for Systematic Reviews of Interventions Version 6.1 (updated September 2020). Cochrane, 2020. Available from training.cochrane.org/handbook.

Higgins 2020a

Higgins JP, Li T, Deeks JJ. Chapter 6: Choosing effect measures and computing estimates of effect. In: Higgins JP, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA, editor(s). Cochrane Handbook for Systematic Reviews of Interventions Version 6.1 (updated September 2020). Cochrane, 2020. Available from training.cochrane.org/handbook .

Hopewell 2020

Hopewell S, Copsey B, Nicolson P, Adedire B, Boniface G, Lamb S. Multifactorial interventions for preventing falls in older people living in the community: a systematic review and meta-analysis of 41 trials and almost 20 000 participants. British Journal of Sports Medicine 2020;54(22):1340-50. [DOI] [PMC free article] [PubMed] [Google Scholar]

Howe 1997

Howe A. Health care costs of an aging population: the case of Australia. Reviews in Clinical Gerontology 1997;7:359-65. [Google Scholar]

Howe 2011

Howe TE, Rochester L, Neil F, Skelton DA, Ballinger C. Exercise for improving balance in older people. Cochrane Database of Systematic Reviews 2011, Issue 11. Art. No: CD004963. [DOI: 10.1002/14651858.CD004963.pub3] [DOI] [PMC free article] [PubMed] [Google Scholar]

Hozo 2005

Hozo SP, Djulbegovic B, Hozo I. Estimating the mean and variance from the median, range, and the size of a sample. BMC Medical Research Methodology 2005;5(1):13. [DOI] [PMC free article] [PubMed] [Google Scholar]

Hubbard 2009

Hubbard RE, Fallah N, Searle SD, Mitnitski A, Rockwood K. Impact of exercise in community-dwelling older adults. PLoS One 2009;4(7):e6174. [DOI] [PMC free article] [PubMed] [Google Scholar]

Keeble 2019

Keeble E, Roberts HC, Williams CD, Van Oppen J, Conroy SP. Outcomes of hospital admissions among frail older people: a 2-year cohort study. British Journal of General Practice 2019;69(685):e555-60. [DOI] [PMC free article] [PubMed] [Google Scholar]

Lamb 2007

Lamb SE, Hauer K, Becker C. Manual for the fall prevention classification system. www.profane.eu.org/documents/Falls_Taxonomy.pdf (accessed 18 July 2021).

Lamb 2011

Lamb SE, Becker C, Gillespie LD, Smith JL, Finnegan S, Potter R, et al, Taxonomy Investigators. Reporting of complex interventions in clinical trials: development of a taxonomy to classify and describe fall‐prevention interventions. Trials 2011;12:125. [DOI] [PMC free article] [PubMed] [Google Scholar]

Laver 2020

Laver K, Crotty M, Low LF, Clemson L, Whitehead C, McLoughlin J, et al. Rehabilitation for people with dementia: a multi-method study examining knowledge and attitudes. BMC Geriatrics 2020;20:531. [DOI] [PMC free article] [PubMed] [Google Scholar]

Liu 2009

Liu CJ, Latham NK. Progressive resistance strength training for improving physical function in older adults. Cochrane Database of Systematic Reviews 2009, Issue 3. Art. No: CD002759. [DOI: 10.1002/14651858.CD002759.pub2] [DOI] [PMC free article] [PubMed] [Google Scholar]

Mitnitski 2001

Mitnitski AB, Mogilner AJ, Rockwood K. Accumulation of deficits as a proxy measure of aging. Scientific World Journal 2001;1:323-36. [DOI] [PMC free article] [PubMed] [Google Scholar]

Mitnitski 2004

Mitnitski AB, Song X, Rockwood K. The estimation of relative fitness and frailty in community-dwelling older adults using self-report data. Journals of Gerontology Series A: Biological Sciences and Medical Sciences 2004;59(6):M627-32. [DOI] [PubMed] [Google Scholar]

Ofori‐Asenso 2019

Ofori-Asenso R, Chin KL, Mazidi M, Zomer E, Ilomaki J, Zullo AR, et al. Global incidence of frailty and prefrailty among community-dwelling older adults: a systematic review and meta-analysis. JAMA Network Open 2019;2(8):e198398. [DOI] [PMC free article] [PubMed] [Google Scholar]

Page 2020

Page MJ, Higgins JP, Sterne JAC. Chapter 13: Assessing risk of bias due to missing results in a synthesis. In: Higgins JP, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA, editor(s). Cochrane Handbook for Systematic Reviews of Interventions Version 6.1 (updated September 2020). Cochrane, 2020. Available from training.cochrane.org/handbook.

Pel‐Littel 2009

Pel-Littel RE, Schuurmans MJ, Emmelot-Vonk MH, Verhaar HJ. Frailty: defining and measuring of a concept. Journal of Nutrition, Health and Aging 2009;13(4):390-4. [DOI] [PubMed] [Google Scholar]

Perera 2006

Perera S, Moody SH, Woodman RC, Studenski A. Meaningful change and responsiveness in common physical performance measures in older adults. Journal of the American Geriatrics Society 2006;54(4):743-9. [DOI] [PubMed] [Google Scholar]

Ridda 2008

Ridda I, Lindley R, MacIntyre RC. The challenges of clinical trials in the exclusion zone: the case of the frail elderly. Australasian Journal on Ageing 2008;27(2):61-6. [DOI] [PubMed] [Google Scholar]

Rockwood 2005

Rockwood K, Song X, MacKnight C, Bergman H, Hogan DB, McDowell I, et al. A global clinical measure of fitness and frailty in elderly people. Canadian Medical Association Journal 2005;173(5):489-95. [DOI] [PMC free article] [PubMed] [Google Scholar]

Rolfson 2006

Rolfson DB, Majumdar SR, Tsuyuki RT, Tahir A, Rockwood K. Validity and reliability of the Edmonton frail scale. Age and Ageing 2006;35(5):526-9. [DOI] [PMC free article] [PubMed] [Google Scholar]

Rydwik 2004

Rydwik E, Frandin K, Akner G. Effects of physical training on physical performance in institutionalised elderly patients (70+) with multiple diagnoses. Age and Ageing 2004;33(1):13-23. [DOI] [PubMed] [Google Scholar]

Schünemann 2020

Schünemann HJ, Higgins JP, Vist GE, Glasziou P, Akl EA, Skoetz N, et al. Chapter 14: Completing ‘Summary of findings’ tables and grading the certainty of the evidence. In: Higgins JP, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA, editor(s). Cochrane Handbook for Systematic Reviews of Interventions Version 6.1 (updated September 2020). Cochrane, 2020. Available from training.cochrane.org/handbook.

Schünemann 2020a

Schünemann HJ, Vist GE, Higgins JP, Santesso N, Deeks JJ, Glasziou P, et al. Chapter 15: Interpreting results and drawing conclusions. In: Higgins JP, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA, editor(s). Cochrane Handbook for Systematic Reviews of Interventions Version 6.1 (updated September 2020). Cochrane, 2020. Available from training.cochrane.org/handbook.

Sherrington 2019

Sherrington C, Fairhall NJ, Wallbank GK, Tiedemann A, Michaleff ZA, Howard K, et al. Exercise for preventing falls in older people living in the community. Cochrane Database of Systematic Reviews 2019, Issue 1. Art. No: CD012424. [DOI: 10.1002/14651858.CD012424] [DOI] [PMC free article] [PubMed] [Google Scholar]

Thompson 2018

Thompson MQ, Theou O, Karnon J, Adams RJ, Visvanathan R. Frailty prevalence in Australia: findings from four pooled Australian cohort studies. Australasian Journal on Ageing 2018;37(2):155-8. [DOI] [PubMed] [Google Scholar]

Tinetti 1988

Tinetti ME, Speechley M, Ginter SF. Risk factors for falls among elderly persons living in the community. New England Journal of Medicine 1988;319(26):1701-7. [DOI] [PubMed] [Google Scholar]

Treacy 2018

Treacy D, Hassett L. The Short Physical Performance Battery. Journal of Physiotherapy 2018;64(1):61. [DOI] [PubMed] [Google Scholar]

Tricco 2017

Tricco AC, Thomas SM, Veroniki AA, Hamid JS, Cogo E, Strifler L, et al. Comparisons of interventions for preventing falls in older adults: a systematic review and meta-analysis. JAMA 2017;318(17):1687-99. [DOI] [PMC free article] [PubMed] [Google Scholar]

Unnanuntana 2018

Unnanuntana A, Jarusriwanna A, Nepal S. Validity and responsiveness of Barthel Index for measuring functional recovery after hemiarthroplasty for femoral neck fracture. Archives of Orthopaedic and Trauma Surgery 2018;138(12):1671-7. [DOI] [PubMed] [Google Scholar]

WHO 2001

World Health Organization. International Classification of Functioning, Disability and Health. apps.who.int/iris/bitstream/handle/10665/42407/9241545429.pdf?sequence=1 (accessed 30 August 2021).

WHO 2010

World Health Organization. International Statistical Classification of Diseases and Related Health Problems 10th Revision. apps.who.int/classifications/apps/icd/icd10online/ (accessed 30 August 2021).

Woo 2012

Woo J, Leung J, Morley JE. Comparison of frailty indicators based on clinical phenotype and the multiple deficit approach in predicting mortality and physical limitation. Journal of the American Geriatrics Society 2012;60(8):1478-86. [DOI] [PubMed] [Google Scholar]

Woolford 2020

Woolford SJ, Sohan O, Dennison EM, Cooper C, Patel HP. Approaches to the diagnosis and prevention of frailty. Aging Clinical and Experimental Research 2020;32(9):1629-37. [DOI] [PMC free article] [PubMed] [Google Scholar]

Worm 2001

Worm CH, Vad E, Puggaard L, Stovring H, Lauritsen J, Kragstrup J. Effects of a multicomponent exercise program on functional ability in community-dwelling, frail older adults. Journal of Aging and Physical Activity 2001;9(4):414-24. [Google Scholar]

[CD010494-bib-0001] Binder EF, Schechtman KB, Ehsani AA, Steger-May K, Brown M, Sinacore DR, et al. Effects of exercise training on frailty in community-dwelling older adults: results of a randomized, controlled trial. Journal of the American Geriatrics Society 2002;50(12):1921-8. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0002] Brown M, Sinacore DR, Ehsani AA, Binder EF, Holloszy JO, Kohrt WM. Low-intensity exercise as a modifier of physical frailty in older adults. Archives of Physical Medicine and Rehabilitation 2000;81(7):960-5. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0003] Bonnefoy M, Cornu C, Normand S, Boutitie F, Bugnard F, Rahmani A, et al. The effects of exercise and protein-energy supplements on body composition and muscle function in frail elderly individuals: a long-term controlled randomised study. British Journal of Nutrition 2003;89(5):731-9. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0004] Normand S. Clarification of study design [personal communication]. Email to: Daniel Treacy 17 April 2018.

[CD010494-bib-0005] Chin A Paw MJ, De Jong N, Schouten EG, Hiddink GJ, Kok FJ. Physical exercise and/or enriched foods for functional improvement in frail, independently living elderly: a randomized controlled trial. Archives of Physical Medicine and Rehabilitation 2001;82(6):811-7. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0006] Chin A Paw MJ, De Jong N, Schouten EG, Van Staveren WA, Kok FJ. Physical exercise or micronutrient supplementation for the wellbeing of the frail elderly? A randomised controlled trial. British Journal of Sports Medicine 2002;36(2):126-31. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0007] Chin A Paw MJ. Additional outcome data [personal communication]. Email to: Daniel Treacy 8 January 2019.

[CD010494-bib-0008] Cameron ID, Fairhall N, Langron C, Lockwood K, Monaghan N, Aggar C, et al. A multifactorial interdisciplinary intervention reduces frailty in older people: randomized trial. BMC Medicine 2013;11:65. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0009] Fairhall N, Sherrington C, Kurrle SE, Lord SR, Lockwood K, Cameron ID. Effect of a multifactorial interdisciplinary intervention on mobility-related disability in frail older people: randomised controlled trial. BMC Medicine 2012;10:120. [DOI: 10.1186/1741-7015-10-120] [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0010] Fairhall N, Sherrington C, Kurrle SE, Lord SR, Lockwood K, Howard K, et al. Economic evaluation of a multifactorial, interdisciplinary intervention versus usual care to reduce frailty in frail older people. Journal of the American Medical Directors Association 2015;16(1):41-8. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0011] Fairhall N, Sherrington C, Lord SR, Kurrle SE, Langron C, Lockwood K, et al. Effect of a multifactorial, interdisciplinary intervention on risk factors for falls and fall rate in frail older people: a randomised controlled trial. Age and Ageing 2014;43(5):616-22. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0012] Gill TM, Baker DI, Gottschalk M, Gahbauer EA, Charpentier PA, De Regt PT, et al. A prehabilitation program for physically frail community-living older persons. Archives of Physical Medicine and Rehabilitation 2003;84(3):394-404. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0013] Gill TM, Baker DI, Gottschalk M, Peduzzi PN, Allore H, Byers A. A program to prevent functional decline in physically frail, elderly persons who live at home. New England Journal of Medicine 2002;347(14):1068-74. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0014] Gill TM, Baker DI, Gottschalk M, Peduzzi PN, Allore H, Van Ness PH. A prehabilitation program for the prevention of functional decline: effect on higher-level physical function. Archives of Physical Medicine and Rehabilitation 2004;85(7):1043-9. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0015] Gill TM, McGloin JM, Gahbauer EA, Shepard DM, Bianco LM. Two recruitment strategies for a clinical trial of physically frail community-living older persons. Journal of the American Geriatrics Society 2001;49(8):1039-45. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0016] Gine-Garriga M, Guerra M, Pages E, Manini TM, Jimenez R, Unnithan VB. The effect of functional circuit training on physical frailty in frail older adults: a randomized controlled trial. Journal of Aging and Physical Activity 2010;18(4):401-24. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0017] Gine-Garriga M, Guerra M, Unnithan VB. The effect of functional circuit training on self-reported fear of falling and health status in a group of physically frail older individuals: a randomized controlled trial. Aging Clinical and Experimental Research 2013;25(3):329-36. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0018] Kim H, Suzuki T, Kim M, Kojima N, Ota N, Shimotoyodome A, et al. Effects of exercise and milk fat globule membrane (MFGM) supplementation on body composition, physical function, and hematological parameters in community-dwelling frail Japanese women: a randomized double blind, placebo-controlled, follow-up trial. PLoS One 2015;10(2):e0116256. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0019] Desjardins L. Additional outcome data [personal communication]. Email to: Daniel Treacy 2 June 2018.

[CD010494-bib-0020] Langlois F, Vu TT, Chasse K, Dupuis G, Kergoat MJ, Bherer L. Benefits of physical exercise training on cognition and quality of life in frail older adults. Journals of Gerontology Series B: Psychological Sciences and Social Sciences 2013;68(3):400-4. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0021] Mugueta-Aguinaga I, Garcia-Zapirain B. FRED: Exergame to prevent dependence and functional deterioration associated with ageing. A pilot three-week randomized controlled clinical trial. International Journal of Environmental Research and Public Health 2017;14:1439. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0022] Rydwik E, Frandin K, Akner G. Effects of a physical training and nutritional intervention program in frail elderly people regarding habitual physical activity level and activities of daily living - a randomized controlled pilot study. Archives of Gerontology and Geriatrics 2010;51(3):283-9. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0023] Rydwik E, Gustafsson T, Frandin K, Akner G. Effects of physical training on aerobic capacity in frail elderly people (75+ years). Influence of lung capacity, cardiovascular disease and medical drug treatment: a randomized controlled pilot trial. Aging Clinical and Experimental Research 2010;22(1):85-94. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0024] Rydwik E, Lammes E, Frandin K, Akner G. Effects of a physical and nutritional intervention program for frail elderly people over age 75. A randomized controlled pilot treatment trial. Aging Clinical and Experimental Research 2008;20(2):159-70. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0025] Rydwik E. Additional outcome data [personal communication]. Email to: Daniel Treacy 16 January 2019.

[CD010494-bib-0026] Tarazona-Santabalbina FJ, Gomez-Cabrera MC, Perez-Ros P, Martinez-Arnau FM, Cabo H, Tsaparas K, et al. A multicomponent exercise intervention that reverses frailty and improves cognition, emotion, and social networking in the community-dwelling frail elderly: a randomized clinical trial. Journal of the American Medical Directors Association 2016;17(5):426-33. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0027] Worm CH, Vad E, Puggaard L, Stovring H, Lauritsen J, Kragstrup J. Effects of a multicomponent exercise program on functional ability in community-dwelling, frail older adults. Journal of Aging and Physical Activity 2001;9(4):414-24. [Google Scholar]

[CD010494-bib-0028] Beling J, Roller M. Multifactorial intervention with balance training as a core component among fall-prone older adults. Journal of Geriatric Physical Therapy 2009;32(3):125-33. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0029] Binder E F, Yarasheski K E, Steger-May K, Sinacore D R, Brown M, Schechtman K B, et al. Effects of progressive resistance training on body composition in frail older adults: results of a randomized, controlled trial. Journals of Gerontology Series A: Biological Sciences and Medical Sciences 2005;60(11):1425-31. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0030] Clegg A, Barber S, Young J, Iliffe S, Forster A. The Home-based Older People's Exercise (HOPE) trial: a pilot randomised controlled trial of a home-based exercise intervention for older people with frailty. Age and Ageing 2014;43(5):687-95. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0031] De Vries NM, Staal JB, Van Der Wees PJ, Akkermans R, Adang E, Olde Rikkert MGM, et al. Personalized physiotherapy in frail older adults with mobility problems is (cost)-effective in improving physical activity and frailty: a RCT. Physiotherapy (United Kingdom) 2015;101:eS1089-90. [Google Scholar]

[CD010494-bib-0032] Faber MJ, Bosscher RJ, Chin A Paw MJ, Van Wieringen PC. Effects of exercise programs on falls and mobility in frail and pre-frail older adults: a multicenter randomized controlled trial. Archives of Physical Medicine and Rehabilitation 2006;87(7):885-96. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0033] Halvarsson A, Franzen E, Faren E, Olsson E, Oddsson L, Stahle A. Long-term effects of new progressive group balance training for elderly people with increased risk of falling - a randomized controlled trial. Clinical Rehabilitation 2013;27(5):450-8. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0034] Hiroyuki S, Uchiyama Y, Kakurai S. Specific effects of balance and gait exercises on physical function among the frail elderly. Clinical Rehabilitation 2003;17(5):472-9. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0035] Ikezoe T, Tsutou A, Asakawa Y, Tsuboyama T. Low intensity training for frail elderly women: long-term effects on motor function and mobility. Journal of Physical Therapy Science 2005;17(1):43-9. [Google Scholar]

[CD010494-bib-0036] Latham NK, Anderson CS, Lee A, Bennett DA, Moseley A, Cameron ID, et al. A randomized, controlled trial of quadriceps resistance exercise and vitamin D in frail older people: the Frailty Interventions Trial in Elderly Subjects (FITNESS). Journal of the American Geriatrics Society 2003;51(3):291-9. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0037] Law H, Bte Ismail N, Wong W, Ramason R, Kua J, Natesan S. Targeting falls through frailty intervention by a combined nutritional intervention and physical exercise programme in community-dwelling older fallers: a pilot randomized clinical trial. Journal of the American Geriatrics Society 2018;66(S2):S128. [Google Scholar]

[CD010494-bib-0038] Lord S R, Castell S, Corcoran J, Dayhew J, Matters B, Shan A, et al. The effect of group exercise on physical functioning and falls in frail older people living in retirement villages: a randomized, controlled trial. Journal of the American Geriatrics Society 2003;51(12):1685-92. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0039] Ng TP, Lim LM, Niti M, Nyunt MSZ, Feng L, Gwee X, et al. Effects of nutritional, physical, cognitive interventions on cognitive outcomes in the Singapore frailty intervention trial (S-FIT). Annals of the Academy of Medicine Singapore 2014;1:S68-9. [Google Scholar]

[CD010494-bib-0040] Pahor M, Guralnik JM, Ambrosius WT, Blair S, Bonds DE, Church TS, et al. Effect of structured physical activity on prevention of major mobility disability in older adults: the LIFE study randomized clinical trial. Journal of the American Medical Association 2014;311(23):2387-96. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0041] Park YI, Lee KY, Kim TI, Jeon MH, Kim DO, Kim JH. The effects of exercise in the frail elderly. Journal of Korean Academy of Community Health Nursing 2012;23(1):91-101. [Google Scholar]

[CD010494-bib-0042] Salem BE, Ma-Pham J, Chen S, Brecht ML, Antonio AL, Ames M. Impact of a community-based frailty intervention among middle-aged and older prefrail and frail homeless women: a pilot randomized controlled trial. Community Mental Health Journal 2017;53(6):688-94. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0043] Seino S, Nishi M, Murayama H, Narita M, Yokoyama Y, Nofuji Y, et al. Effects of a multifactorial intervention comprising resistance exercise, nutritional and psychosocial programs on frailty and functional health in community-dwelling older adults: a randomized, controlled, cross-over trial. Geriatrics & Gerontology International 2017;17(11):2034-45. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0044] Trombetti A, Hars M, Hsu FC, Reid KF, Church TS, Gill TM, et al. Effect of physical activity on frailty: secondary analysis of a randomized controlled trial. Annals of Internal Medicine 2018;168(5):309-16. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0045] Allen NE, Sherrington C, Paul SS, Canning CG. Balance and falls in Parkinson's disease: a meta-analysis of the effect of exercise and motor training. Movement Disorders 2011;26(9):1605-15. [PMID: ] [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0046] Ashworth NL, Chad KE, Harrison EL, Reeder BA, Marshall SC. Home versus center based physical activity programs in older adults. Cochrane Database of Systematic Reviews 2005, Issue 1. Art. No: CD004017. [DOI: 10.1002/14651858.CD004017.pub2] [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0047] Binder EF, Schechtman KB, Ehsani AA, Steger-May K, Brown M, Sinacore DR, et al. Effects of exercise training on frailty in community-dwelling older adults: results of a randomized, controlled trial. Journal of the American Geriatrics Society 2002;50(12):1921-8. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0048] Binder EF, Brown M, Sinacore DR, Steger-May K, Yarasheski KE, Schechtman KB. Effects of extended outpatient rehabilitation after hip fracture: a randomized controlled trial. Journal of the American Medical Association 2004;292(7):837-46. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0049] Cameron ID, Dyer SM, Panagoda CE, Murray GR, Hill KD, Cumming RG, et al. Interventions for preventing falls in older people in care facilities and hospitals. Cochrane Database of Systematic Reviews 2018, Issue 9. Art. No: CD005465. [DOI: 10.1002/14651858.CD005465.pub4] [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0050] Visual Rx. Cates C, Version 3. Dr. Christopher Cates EBM website. Available at www.nntonline.net, 2008.

[CD010494-bib-0051] Cations M, Laver KE, Crotty M, Cameron ID. Rehabilitation in dementia care. Age and Ageing 2017;47(2):171-74. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0052] Chin A Paw MJ, De Jong N, Schouten EG, Hiddink GJ, Kok FJ. Physical exercise and/or enriched foods for functional improvement in frail, independently living elderly: a randomized controlled trial. Archives of Physical Medicine and Rehabilitation 2001;82(6):811-7. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0053] Chou CH, Hwang CL, Wu YT. Effect of exercise on physical function, daily living activities, and quality of life in the frail older adults: a meta-analysis. Archives of Physical Medicine and Rehabilitation 2012;93:237-44. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0054] Clegg A, Young J, Iliffe S, Rikkert MO, Rockwood K. Frailty in elderly people. Lancet 2013;381(9868):752-62. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0055] Clegg A, Barber S, Young J, Ilffe S, Forster A. The Homebased Older People's Exercise (HOPE) trial: a pilot randomised controlled trial of a home-based exercise intervention for older people with frailty. Age and Ageing 2014;43(5):687-95. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0056] Daniels R, Van Rossum E, De Witte L, Kempen GI, Van den Heuvel W. Interventions to prevent disability in frail community-dwelling elderly: a systematic review. BMC Health Services Research 2008;8:278. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0057] Deeks JJ, Higgins JP, Altman DG. Chapter 10: Analysing data and undertaking meta-analyses. In: Higgins JP, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA, editor(s). Cochrane Handbook for Systematic Reviews of Interventions Version 6.1 (updated September 2020). Cochrane, 2020. Available from www.training.cochrane.org/handbook.

[CD010494-bib-0058] De Morton N, Keating JL, Jeffs K. Exercise for acutely hospitalised older medical patients. Cochrane Database of Systematic Reviews 2007, Issue 1. Art. No: CD005955. [DOI: 10.1002/14651858.CD005955.pub2] [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0059] Dent E, Kowal P, Hoogendijk EO. Frailty measurement in research and clinical practice: a review. European Journal of Internal Medicine 2016;31:3-10. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0060] Dent E, Morley JE, Cruz-Jentoft AJ, Woodhouse L, Rodriguez-Manas L, Fried LP, et al. Physical frailty: ICFSR international clinical practice guidelines for identification and management. Journal of Nutrition, Health and Aging 2019;23(9):771-87. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0061] Fairhall N, Sherrington C, Cameron I. Mobility training for increasing mobility and functioning in older people with frailty. Cochrane Database of Systematic Reviews 2013, Issue 5. Art. No: CD010494. [DOI: 10.1002/14651858.CD010494] [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0062] Fiatarone MA, O'Neill EF, Ryan ND, Clements KM, Solares GR, Nelson ME, et al. Exercise training and nutritional supplementation for physical frailty in very elderly people. New England Journal of Medicine 1994;330(25):1769-75. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0063] Fried LP, Tangen CM, Walston J, Newman AB, Hirsch C, Gottdiener J, et al. Frailty in older adults: evidence for a phenotype. Journals of Gerontology Series A: Biological Sciences and Medical Sciences 2001;56(3):M146-56. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0064] Fried LP, Ferrucci L, Darer J, Williamson JD, Anderson G. Untangling the concepts of disability, frailty, and comorbidity: implications for improved targeting and care. Journals of Gerontology Series A: Biological Sciences and Medical Sciences 2004;59(3):255-63. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0065] Gill TM, Williams CS, Tinetti ME. Assessing risk for the onset of functional dependence among older adults: the role of physical performance. Journal of the American Geriatrics Society 1995;43(6):603-9. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0066] Gill TM, Baker DI, Gottschalk M, Peduzzi PN, Allore H, Byers A. A program to prevent functional decline in physically frail, elderly persons who live at home. New England Journal of Medicine 2002;347(14):1068-74. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0067] Gill TM, Allore HG, Holford TR, Guo Z. Hospitalization, restricted activity, and the development of disability among older persons. JAMA 2004;292(17):2115-24. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0068] Gill TM, Gahbauer EA, Allore HG, Han L. Transitions between frailty states among community-living older persons. Archives of Internal Medicine 2006;166(4):418-23. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0069] Gine-Garriga M, Roque-Fıguls M, Coll-Planas L, Sitja-Rabert M, Salva A. Physical exercise interventions for improving performance-based measures of physical function in community-dwelling, frail older adults: a systematic review and meta-analysis. Archives of Physical Medicine and Rehabilitation 2014;95:753-69. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0070] GRADEpro GDT. Version accessed 6 August 2021. Hamilton (ON): McMaster University (developed by Evidence Prime). Available at gradepro.org.

[CD010494-bib-0071] Higgins JP, Savović J, Page MJ, Elbers RG, Sterne JAC. Chapter 8: Assessing risk of bias in a randomized trial. In: Higgins JP, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA, editor(s). Cochrane Handbook for Systematic Reviews of Interventions Version 6.1 (updated September 2020). Cochrane, 2020. Available from training.cochrane.org/handbook.

[CD010494-bib-0072] Higgins JP, Li T, Deeks JJ. Chapter 6: Choosing effect measures and computing estimates of effect. In: Higgins JP, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA, editor(s). Cochrane Handbook for Systematic Reviews of Interventions Version 6.1 (updated September 2020). Cochrane, 2020. Available from training.cochrane.org/handbook .

[CD010494-bib-0073] Hopewell S, Copsey B, Nicolson P, Adedire B, Boniface G, Lamb S. Multifactorial interventions for preventing falls in older people living in the community: a systematic review and meta-analysis of 41 trials and almost 20 000 participants. British Journal of Sports Medicine 2020;54(22):1340-50. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0074] Howe A. Health care costs of an aging population: the case of Australia. Reviews in Clinical Gerontology 1997;7:359-65. [Google Scholar]

[CD010494-bib-0075] Howe TE, Rochester L, Neil F, Skelton DA, Ballinger C. Exercise for improving balance in older people. Cochrane Database of Systematic Reviews 2011, Issue 11. Art. No: CD004963. [DOI: 10.1002/14651858.CD004963.pub3] [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0076] Hozo SP, Djulbegovic B, Hozo I. Estimating the mean and variance from the median, range, and the size of a sample. BMC Medical Research Methodology 2005;5(1):13. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0077] Hubbard RE, Fallah N, Searle SD, Mitnitski A, Rockwood K. Impact of exercise in community-dwelling older adults. PLoS One 2009;4(7):e6174. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0078] Keeble E, Roberts HC, Williams CD, Van Oppen J, Conroy SP. Outcomes of hospital admissions among frail older people: a 2-year cohort study. British Journal of General Practice 2019;69(685):e555-60. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0079] Lamb SE, Hauer K, Becker C. Manual for the fall prevention classification system. www.profane.eu.org/documents/Falls_Taxonomy.pdf (accessed 18 July 2021).

[CD010494-bib-0080] Lamb SE, Becker C, Gillespie LD, Smith JL, Finnegan S, Potter R, et al, Taxonomy Investigators. Reporting of complex interventions in clinical trials: development of a taxonomy to classify and describe fall‐prevention interventions. Trials 2011;12:125. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0081] Laver K, Crotty M, Low LF, Clemson L, Whitehead C, McLoughlin J, et al. Rehabilitation for people with dementia: a multi-method study examining knowledge and attitudes. BMC Geriatrics 2020;20:531. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0082] Liu CJ, Latham NK. Progressive resistance strength training for improving physical function in older adults. Cochrane Database of Systematic Reviews 2009, Issue 3. Art. No: CD002759. [DOI: 10.1002/14651858.CD002759.pub2] [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0083] Mitnitski AB, Mogilner AJ, Rockwood K. Accumulation of deficits as a proxy measure of aging. Scientific World Journal 2001;1:323-36. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0084] Mitnitski AB, Song X, Rockwood K. The estimation of relative fitness and frailty in community-dwelling older adults using self-report data. Journals of Gerontology Series A: Biological Sciences and Medical Sciences 2004;59(6):M627-32. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0085] Ofori-Asenso R, Chin KL, Mazidi M, Zomer E, Ilomaki J, Zullo AR, et al. Global incidence of frailty and prefrailty among community-dwelling older adults: a systematic review and meta-analysis. JAMA Network Open 2019;2(8):e198398. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0086] Page MJ, Higgins JP, Sterne JAC. Chapter 13: Assessing risk of bias due to missing results in a synthesis. In: Higgins JP, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA, editor(s). Cochrane Handbook for Systematic Reviews of Interventions Version 6.1 (updated September 2020). Cochrane, 2020. Available from training.cochrane.org/handbook.

[CD010494-bib-0087] Pel-Littel RE, Schuurmans MJ, Emmelot-Vonk MH, Verhaar HJ. Frailty: defining and measuring of a concept. Journal of Nutrition, Health and Aging 2009;13(4):390-4. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0088] Perera S, Moody SH, Woodman RC, Studenski A. Meaningful change and responsiveness in common physical performance measures in older adults. Journal of the American Geriatrics Society 2006;54(4):743-9. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0089] Ridda I, Lindley R, MacIntyre RC. The challenges of clinical trials in the exclusion zone: the case of the frail elderly. Australasian Journal on Ageing 2008;27(2):61-6. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0090] Rockwood K, Song X, MacKnight C, Bergman H, Hogan DB, McDowell I, et al. A global clinical measure of fitness and frailty in elderly people. Canadian Medical Association Journal 2005;173(5):489-95. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0091] Rolfson DB, Majumdar SR, Tsuyuki RT, Tahir A, Rockwood K. Validity and reliability of the Edmonton frail scale. Age and Ageing 2006;35(5):526-9. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0092] Rydwik E, Frandin K, Akner G. Effects of physical training on physical performance in institutionalised elderly patients (70+) with multiple diagnoses. Age and Ageing 2004;33(1):13-23. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0093] Schünemann HJ, Higgins JP, Vist GE, Glasziou P, Akl EA, Skoetz N, et al. Chapter 14: Completing ‘Summary of findings’ tables and grading the certainty of the evidence. In: Higgins JP, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA, editor(s). Cochrane Handbook for Systematic Reviews of Interventions Version 6.1 (updated September 2020). Cochrane, 2020. Available from training.cochrane.org/handbook.

[CD010494-bib-0094] Schünemann HJ, Vist GE, Higgins JP, Santesso N, Deeks JJ, Glasziou P, et al. Chapter 15: Interpreting results and drawing conclusions. In: Higgins JP, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA, editor(s). Cochrane Handbook for Systematic Reviews of Interventions Version 6.1 (updated September 2020). Cochrane, 2020. Available from training.cochrane.org/handbook.

[CD010494-bib-0095] Sherrington C, Fairhall NJ, Wallbank GK, Tiedemann A, Michaleff ZA, Howard K, et al. Exercise for preventing falls in older people living in the community. Cochrane Database of Systematic Reviews 2019, Issue 1. Art. No: CD012424. [DOI: 10.1002/14651858.CD012424] [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0096] Thompson MQ, Theou O, Karnon J, Adams RJ, Visvanathan R. Frailty prevalence in Australia: findings from four pooled Australian cohort studies. Australasian Journal on Ageing 2018;37(2):155-8. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0097] Tinetti ME, Speechley M, Ginter SF. Risk factors for falls among elderly persons living in the community. New England Journal of Medicine 1988;319(26):1701-7. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0098] Treacy D, Hassett L. The Short Physical Performance Battery. Journal of Physiotherapy 2018;64(1):61. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0099] Tricco AC, Thomas SM, Veroniki AA, Hamid JS, Cogo E, Strifler L, et al. Comparisons of interventions for preventing falls in older adults: a systematic review and meta-analysis. JAMA 2017;318(17):1687-99. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0100] Unnanuntana A, Jarusriwanna A, Nepal S. Validity and responsiveness of Barthel Index for measuring functional recovery after hemiarthroplasty for femoral neck fracture. Archives of Orthopaedic and Trauma Surgery 2018;138(12):1671-7. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0101] World Health Organization. International Classification of Functioning, Disability and Health. apps.who.int/iris/bitstream/handle/10665/42407/9241545429.pdf?sequence=1 (accessed 30 August 2021).

[CD010494-bib-0102] World Health Organization. International Statistical Classification of Diseases and Related Health Problems 10th Revision. apps.who.int/classifications/apps/icd/icd10online/ (accessed 30 August 2021).

[CD010494-bib-0103] Woo J, Leung J, Morley JE. Comparison of frailty indicators based on clinical phenotype and the multiple deficit approach in predicting mortality and physical limitation. Journal of the American Geriatrics Society 2012;60(8):1478-86. [DOI] [PubMed] [Google Scholar]

[CD010494-bib-0104] Woolford SJ, Sohan O, Dennison EM, Cooper C, Patel HP. Approaches to the diagnosis and prevention of frailty. Aging Clinical and Experimental Research 2020;32(9):1629-37. [DOI] [PMC free article] [PubMed] [Google Scholar]

[CD010494-bib-0105] Worm CH, Vad E, Puggaard L, Stovring H, Lauritsen J, Kragstrup J. Effects of a multicomponent exercise program on functional ability in community-dwelling, frail older adults. Journal of Aging and Physical Activity 2001;9(4):414-24. [Google Scholar]

PERMALINK

Mobility training for increasing mobility and functioning in older people with frailty

Daniel Treacy

Leanne Hassett

Karl Schurr

Nicola J Fairhall

Ian D Cameron

Catherine Sherrington

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Summary of findings

Summary of findings 1. Mobility training compared to control.

Summary of findings 2. Mobility training compared to control: 6 months post‐intervention.

Summary of findings 3. Mobility training compared to control: subgroups with or without cognitive impairments.

Background

Description of the condition

Description of the intervention

How the intervention might work

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Major outcomes

Minor outcomes

Timing of outcome assessment

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Measures of treatment effect

Data relating to mobility

Data relating to functioning

Data relating to death, falls, adverse outcomes, and movement to nursing care facilities

Unit of analysis issues

Dealing with missing data

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

Summary of findings and assessment of the certainty of the evidence

Interpreting results and reaching conclusions

Results

Description of studies

Results of the search

1.

Included studies

Trial design

1. Study design, length of follow‐up, setting, and trial size.

Trial size and setting

Participants

Interventions

2. Key characteristics of participants and intervention approach.

Outcomes

Mobility

Functioning

Death

Admission to a nursing care facility

Falls

Adverse events

Frailty measures

Excluded studies

Risk of bias in included studies

2.

3.

Allocation