G Model MAT-6389; No of Pages 13 ARTICLE IN PRESS Maturitas xxx (2015) xxx–xxx Contents lists available at ScienceDirect Maturitas journal homepage: www.elsevier.com/locate/maturitas Review Individualized home-based exercise programs for older people to reduce falls and improve physical performance: A systematic review and meta-analysis Keith D Hill a,c,∗ , Susan W Hunter b , Frances Batchelor c , Vinicius Cavalheri a , Elissa Burton a a School of Physiotherapy and Exercise Science, Faculty of Health Sciences, Curtin University, Perth, Western Australia 6845, Australia School of Physical Therapy, The University of Western Ontario, Canada c National Ageing Research Institute, Royal Melbourne Hospital, PO Box 2027, Parkville, Victoria 3050, Australia b a r t i c l e Article history: Available online xxx Keywords: Exercise Falls prevention Community Elderly Personalized i n f o a b s t r a c t There is considerable diversity in the types of exercise programs investigated to reduce falls in older people The purpose of this paper was to review the effectiveness of individualized (tailored) home-based exercise programs in reducing falls and improving physical performance among older people living in the community A systematic review and meta-analysis was conducted of randomized or quasi-randomized trials that utilized an individualized home-based exercise program with at least one falls outcome measure reported Single intervention exercise studies, and multifactorial interventions where results for an exercise intervention were reported independently were included Two researchers independently rated the quality of each included study Of 16,871 papers identified from six databases, 12 met all inclusion criteria (11 randomized trials and a pragmatic trial) Study quality overall was high Sample sizes ranged from 40 to 981, participants had an average age 80.1 years, and although the majority of studies targeted the general older population, several studies included clinical groups as their target (Parkinson’s disease, Alzheimer’s disease, and hip fracture) The meta-analysis results for the five studies reporting number of fallers found no significant effect of the intervention (RR [95% CI] = 0.93 [0.72–1.21]), although when a sensitivity analysis was performed with one study of participants recently discharged from hospital removed, this result was significant (RR [95% CI] = 0.84 [0.72–0.99]) The meta-analysis also found that intervention led to significant improvements in physical activity, balance, mobility and muscle strength There were no significant differences for measures of injurious falls or fractures © 2015 Elsevier Ireland Ltd All rights reserved Contents Introduction Methods 2.1 Objectives 2.2 Eligibility criteria 2.3 Information sources and search strategy 2.4 Study selection 2.5 Data collection process 2.6 Study quality 2.7 Data analysis 00 00 00 00 00 00 00 00 00 ∗ Corresponding author at: School of Physiotherapy and Exercise Science, Faculty of Health Sciences, Curtin University, Perth, Western Australia 6845, Australia Tel.: +61 92663618; fax: +61 92663699 E-mail addresses: Keith.Hill@Curtin.edu.au (K.D Hill), smuir4@uwo.ca (S.W Hunter), f.batchelor@nari.unimelb.edu.au (F Batchelor), Vinicius.Cavalher@curtin.edu.au (V Cavalheri), E.Burton@curtin.edu.au (E Burton) http://dx.doi.org/10.1016/j.maturitas.2015.04.005 0378-5122/© 2015 Elsevier Ireland Ltd All rights reserved Please cite this article in press as: Hill KD, et al Individualized home-based exercise programs for older people to reduce falls and improve physical performance: A systematic review and meta-analysis Maturitas (2015), http://dx.doi.org/10.1016/j.maturitas.2015.04.005 G Model MAT-6389; No of Pages 13 ARTICLE IN PRESS K.D Hill et al / Maturitas xxx (2015) xxx–xxx Results 3.1 Study selection 3.2 Interventions 3.3 Outcome measures 3.4 Dropout and adherence to home exercises 3.5 Quality of studies 3.6 Effectiveness of intervention programs 3.6.1 Falls 3.6.2 Physical activity 3.6.3 Balance 3.6.4 Muscle strength 3.6.5 Mobility Discussion Conclusion Ethical approval Contributors Competing interests Funding Provenance and peer review References Introduction Falls are the leading cause of injury related hospitalizations in Australia [1] and many other countries, and the greatest percentage of those injured from falls are people aged over 65 years Falls are also the 19th leading cause of disability-adjusted life years lost globally across all health conditions, and trends indicate they will become an even stronger contributor to disability-adjusted life years in the future [2] Hospitalization figures only reflect a minority of the impact of falls among older people In a recent epidemiological study in Scotland, only 20% of the 294,000 people aged over 65 living in the community who fell in a 12 month period presented to medical services, with only 16% presenting to the Emergency Department, and 6% of these admitted to hospital [3] Even in the absence of injury, other common sequelae of falls that can impact substantially on quality of life and ability to live independently include loss of confidence in mobility, reduced activity level, depression, and impaired balance and function [4,5] Exercise is a well-researched area of falls prevention A systematic review and meta-analysis published in 2011 included 54 randomized trials covering community and residential care settings (85% community) [6] This review and meta-analysis combined all studies irrespective of setting and identified three main characteristics of exercise programs that increased the likelihood of the program being effective in reducing falls: (1) moderate or high challenge to balance; (2) at least 50 h of exercise equating to h a week intensity; and (3) the exercise must be ongoing, once stopped the effect is lost quickly It is important to note that there are substantial differences between settings in terms of population, environment, and health and care staff support that necessitate considering settings separately An exploratory sub-analysis of studies undertaken only in residential care settings in Sherrington and colleagues’ meta-analysis identified a non-significant reduction in falls following the exercise intervention, highlighting the need for separate analyses by setting The focus of the current paper is limited to exercise programs within the community setting (that is excluding hospital and residential care facilities, both low and high care) Even in the community setting, there is considerable diversity in the types of exercise programs available, where and how they are implemented, and their associated outcomes One of the important distinctions about exercise programs for older people is whether they are centre-based (i.e the older person needs to travel to an external venue to participate in the exercise program, and 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 the program is often group-based) or home-based (i.e the exercise program is able to be undertaken individually in their own home) Differentiating home-based from group-based exercise programs is important, as there is some evidence of differing outcomes [7,8], different adherence rates [7] and different factors influencing preference and participation in these two types of program by older people [9,10] A further important distinction is whether the exercise program is individualized (i.e tailored to meet the specific needs of an individual, in terms of balance, mobility, function and co-morbidities) or a generic program where the same exercise program is provided to all exercisers Individualized programs are more likely to be at a suitable level to safely stress balance and function in a manner likely to achieve health benefits, and to be monitored and progressed if performance changes over time The focus of this review is individualized home-based exercise programs for older people in the community setting, aiming to reduce falls Methods 2.1 Objectives The purpose of this systematic review and meta-analysis was to determine the effectiveness of individualized home-based exercise programs for older people in the community setting in reducing falls, and improving secondary outcomes of physical performance including physical activity, balance, mobility and strength 2.2 Eligibility criteria The review was limited to studies meeting the following eligibility criteria: • Study participants: ◦ aged 60 years and over (at least 50% of the sample), ◦ living in the community • A home-based exercise program that is personalized or individualized to the older person’s capabilities (different exercises selected for each participant based on assessment, exercises modified based on individual progress or needs) and targets a reduction in falls (and/or) risk of falls Home-based exercise programs were included if they were a single intervention; or if a home-based exercise program was part of a multifactorial Please cite this article in press as: Hill KD, et al Individualized home-based exercise programs for older people to reduce falls and improve physical performance: A systematic review and meta-analysis Maturitas (2015), http://dx.doi.org/10.1016/j.maturitas.2015.04.005 G Model ARTICLE IN PRESS MAT-6389; No of Pages 13 K.D Hill et al / Maturitas xxx (2015) xxx–xxx Table Search strategy (according to CINAHL terminology) 10 11 12 13 14 person* exercise ti,ab person* activity ti,ab individ* exercise ti,ab individ* active* ti,ab or or or fall* ti,ab accidental fall ti,ab fall* prevent* ti,ab or or older ti,ab age* ti.ab elder ti,ab 10 or 11 or 12 and and 13 percentages, average age, withdrawal rate, outcome measures, number of falls, effect of the intervention and length of follow-up 2.6 Study quality The Cochrane’s Collaboration’s “risk of bias tool” was used by three independent researchers (EB, SWH, FB) to assess the methodological quality of each paper [12] Categories that were assessed included sequence generation, allocation concealment, participant and staff blinding, blinding of outcome assessment, incomplete outcome data, selective outcome reporting, and other sources of bias [12] Risk of bias included three different levels of assessment: “low risk”, “unclear risk”, or “high risk” of bias [12] 2.7 Data analysis intervention, using a factorial design, with results for the exercise intervention reported separately • Outcome measures included one or more of: number of falls, rate of falls, number of fallers, or time to first fall Other outcomes may include fear of falling, function, physical performance (e.g balance or strength), or adherence to the exercise intervention • Study design: randomized controlled trials (RCT) and quasiexperimental trials • Studies written in English Where two or more studies report data from the same sample, only one of these studies was included in the meta-analysis 2.3 Information sources and search strategy Databases searched included Medline (ProQuest), CINAHL, PubMed, PsycInfo, EMBASE and Scopus, from January 1974 to December 2014 Reference lists of the identified papers were scanned and only papers in English were included, no unpublished data, books, conference proceedings, theses or poster abstracts were included The search strategy was conducted using a number of keywords that were to be identified in the title or abstract of the paper Table outlines the search strategy undertaken in CINAHL Language and syntax were adapted dependent on the database: for example, PubMed allowed title/abstract to be searched simultaneously however not all databases allowed this and in these cases the abstract was searched 2.4 Study selection The study selection was a three stage process: stage one involved one author (EB) initially scanning the titles and abstracts to exclude articles not meeting the criteria Stage two was a full screening of the abstracts by EB Full articles were screened by two authors (EB and FB) to identify which articles met the inclusion criteria, where disagreement occurred, EB and FB referred to the inclusion criteria and study protocol and communicated until consensus was achieved Reference lists of included papers and recent reviews (in particular Sherrington et al.’s [6] review and meta-analysis of exercise RCTs with falls related outcomes) were screened for additional articles, and three additional studies not found during the search were also included as they met the criteria We used the PRISMA checklist to ensure that the results were reported systematically [11] 2.5 Data collection process Each of the included studies had the following data extracted: design, purpose, details of the intervention, sample size, gender Each study was described outlining their characteristics, the intervention and outcome measures used, adherence to the exercise interventions, quality of the studies and effectiveness of the intervention programs (Tables and 3) The Review Manager (RevMan) version 5.3 was used to conduct statistical analyses and create forest plots [13] Both the I2 statistic and visual inspection of the forest plots were used to assess heterogeneity Initially, a random-effects model was used to calculate summary estimates When studies were found to be homogeneous, a fixed-effect model was applied When I2 was > 50% a random-effect model was applied For dichotomous outcomes, results of studies were meta-analyzed using the Mantel-Haenszel’s fixed effects model [13], and risk ratios (RR) with their respective 95% confidence intervals (CI) were calculated For continuous outcomes, the results of studies were meta-analyzed using the inverse variance DerSimonian and Laird method [14], and either the mean differences (MD) or standardized mean differences (SMD), with their respective 95% CI, were calculated For continuous outcomes, if means and standard deviations for differences on outcome measures collected at baseline and at follow up were not available, the meta-analyses of continuous outcomes were performed using follow-up data only, however, studies with baseline differences between the control group and intervention group were omitted from these analyses Results 3.1 Study selection Fig shows the study selection flow chart The search generated 16,871 articles from the six databases After duplicate articles were removed, 3889 articles remained Abstracts and subsequently full text articles of those remaining at each latter stage were reviewed against review inclusion criteria, following which articles remained in the review Reference lists for included papers and recent exercise and falls prevention reviews were scanned, including Sherrington et al.’s [6] updated meta-analysis, and three further articles were included Eleven of the 12 articles included in the systematic review were RCTs [15,17–26] The other article was a pragmatic trial [27] The sample sizes ranged from 40 [25] to 981 [27], with an average sample size of 250 (Table 2) The average age across the 12 studies was 80.1 years, with an average age range between 72.2 and 84.1 years Two thousand, nine hundred and ninety nine participants completed baseline testing and 2570 completed post-testing across the 12 studies, an average retention rate of 82.24% The largest dropout rate was found for Campbell et al.’s (1999) study where only 67.76% of the study population were retained [18] However, given the study period was two years and the women participating were aged, on average over 80 years of age this retention rate seems Please cite this article in press as: Hill KD, et al Individualized home-based exercise programs for older people to reduce falls and improve physical performance: A systematic review and meta-analysis Maturitas (2015), http://dx.doi.org/10.1016/j.maturitas.2015.04.005 G Model ARTICLE IN PRESS MAT-6389; No of Pages 13 K.D Hill et al / Maturitas xxx (2015) xxx–xxx Table Study characteristics Study Study Design Study Purpose Intervention Sample Size; % female; age (years); specific population; drop out Number of falls/fallers Intervention effect Follow-up Ashburn et al., 2007 UK RCT Assess the effectiveness of a home exercise and strategy program for repeat fallers with Parkinson’s Disease (PD) E: Exercise performed 7×/week for muscle strengthening, ROM, balance training, and walking Falls prevention strategies were taught Contact was weekly visits for h over weeks, then participants contacted monthly by phone to provide encouragement n = 142 (E: 70, C: 72); female: 39.5%; age: 72.15 yrs (range 44–91); population Parkinson’s Disease; drop out (total): 10.56% (n = 15) Fallers at months: E = 37/65 (57%), C = 42/64 (66%), No significant difference between groups in falls Significant difference in near falls and repeat near falls rates at weeks and months for exercise group and months n = 233 (E: 116, C: 117); female: 100%; age: 84.1 yrs; older women women ≥ 75 years; drop out: 8.58% (n = 20) Falls at 12 months: E = 88 C = 152 Balance improved in exercise group: 0.42 (0.86) compared to controls: −0.01 (0.80) months and 12 months n = 152 (E: 71, C: 81); female: 100%; age: 83.9yrs; older women ≥ 80 years; drop out: 32.24% (n = 49) n = 317, (E1: 107; E2: 105; C: 105; female: 54.9%; age: 83.4 yrs; general ≥70 years; drop out at 12 months assessment, E1: 24.3% (n = 26); E2: 22.9% (n = 24); C: 23.8% (n = 25) 2nd year only, E: 50, control: 68 Exercise program showed a significant reduction in falls 24 months 31% reduction in falls for the LiFE group compared to the control group, no significant reduction in falls for structured exercise group compared to controls LiFE participant’s significantly improved strength and balance compared to control group The structured program showed small and significant effects for the five level balance hierarchy scale ADLs were significantly improved for the LiFE group compared to the controls months and 12 months Campbell et al., 1997 New Zealand Campbell et al., 1999 New Zealand Clemson et al., 2012 Australia RCT RCT Randomized Parallel Trial Evaluate the effectiveness of a home exercise program compared with usual care C: Usual care (contact with local PD nurse) E: Exercises for 30 3×/week, and walk outside 3×/week Four, h PT visits in first weeks, then regular phone contact to continue motivation Assess the effectiveness of a home-based exercise program for older women over two years C: Research nurse made social visit times in weeks and phoned regularly See above intervention and sample same as Campbell et al., 1997 Follow-up results are over the second year and two years combined Determine the effectiveness of a lifestyle balance and strength program in reducing falls in older, high risk people living in the community E1 The LiFE exercise program included movements to improve balance, increase strength and are embedded into everyday activity and are therefore completed multiple times throughout each day Taught by either PTs or OTs over five sessions with two booster sessions and two follow-up phone calls over months E2: Structured program involved exercise for balance and for strength using ankle cuffs Taught by either PTs or OTs over five sessions with two booster sessions and two follow-up phone calls over months C: two sessions, one booster and follow-up phone calls comprised 12 gentle exercises no change or increase in intensity was provided Fallers at months: E = 46/63 (73%) C = 49/63 (78%) Difference between groups in FR and quality of life was significant at months Total falls for two years, E: 138, C: 220 Falls at 12 months: E1: 172; E2: 193; C: 224 Please cite this article in press as: Hill KD, et al Individualized home-based exercise programs for older people to reduce falls and improve physical performance: A systematic review and meta-analysis Maturitas (2015), http://dx.doi.org/10.1016/j.maturitas.2015.04.005 G Model ARTICLE IN PRESS MAT-6389; No of Pages 13 K.D Hill et al / Maturitas xxx (2015) xxx–xxx Table (Continued) Study Study Design Study Purpose Intervention Sample Size; % female; age (years); specific population; drop out Number of falls/fallers Intervention effect Follow-up Gardner et al., 2002 New Zealand Pragmatic trial in three exercise centres and four control centres Investigate the program reach, uptake and compliance and also test the effectiveness of the exercise program in people older than 80 years E: The Otago Exercise program: leg strengthening and balance retraining exercise (3×/week) and an individually prescribed walking plan (2×/week) Nurse made five home visits and phoned participant monthly Postcard calendars where used to monitor compliance n = 981 (E: 700, C 281; female: 66.5%; age: 83.6 yrs; general ≥ 80 years; drop out: E: 20% (n = 65), C: 12% (n = 14) Fallers at 12 months: E: 103 (44%); C:51 (52%) 12 months n = 150, each group n = 50; female: 51%; age: 76.8 yrs; general ≥65 years; drop-out: E: 22% (n = 11); C1: 8% (n = 5); C2: 20% (n = 10) Fall incidence rate (per 1000 person years) E: 2.4; C1: 1.1; C2: 1.6 The exercise program reduced the number of falls by 30% and the number of falls resulting in injury by 28% in a general practice community setting Overall balance improved for the exercise centre participants compared to the control centres as did the time taken to complete the chair stand test No significant differences in rate of falls between the groups Significant difference between exercise and education groups for balance, functional reach and fear of falling and for the physical, psychological and environmental domains of the WHOQOL-BREF n = 59, (E: 31, C: 28; female: 69.4%; age: 82.25 yrs; general ≥70 years; drop-out: E: 9.7% (n = 3), C: 14.3% (n = 4) When two outliers excluded, adjusted incident rate ratio was 0.47 (95% CI 0.24–0.96) No significant difference between groups at months for fall risk or functional mobility There was a significant difference between groups for the response inhibition (part of Stroop Test) and 12 months n = 180, (E: 91; C: 89; female: 100%; age: 82.4 yrs; older women with a hip fracture recently discharged from hospital; drop out: E: 23.1% (n = 21), C: 31.5% (n = 28) Falls at 12 months: E: 31, C: 31 The intervention group bone mineral density showed small effect sizes between and 0.2 SDs 2, 6, 12 months C: usual care Lin et al., 2007 Taiwan RCT Liu-Ambrose et al., 2008 Canada RCT Orwig et al., 2011 United States RCT Assess the effects of three fall-prevention programs on quality of life, function, activities of daily living, fear of falling and depression in adults aged 65 and over Determine the effects of the Otago Exercise program on falls risk, mobility and executive functioning after months in older adults with a history of falling Assess whether a 12 month home-based exercise program could improve outcomes for people with hip fracture Three groups included exercise (E), home safety assessment and modification (C1) and education (C2) Interventions conducted every weeks for months E: Exercise program consisted of stretching, strengthening and balance training Performed at least 3×/week C1: Home safety group received modification after each visit C2: Education group received social visits every weeks and were provided with falls prevention pamphlets E: The Otago Exercise program First four visits every weeks and a final (fifth visit at months) Exercises performed 3×/week, and walk for 30 2×/week C: Care as per American Geriatrics Society Fall Prevention Guidelines E: Exercise Plus program consisted of: exercise and self-efficacy based motivational components run by exercise trainers They received trainer-supervised exercise sessions per week for the first months, and then per week for the next months It then dropped to once a week, then once a fortnight for a maximum of 56 supervised sessions Phone calls were made to keep motivation when supervised sessions were decreased Exercise combined aerobic exercise, strength and stretching exercises Participants undertook aerobic activity 3×/week and strength 2×/week and months No significant differences for any other outcome measures including falls Please cite this article in press as: Hill KD, et al Individualized home-based exercise programs for older people to reduce falls and improve physical performance: A systematic review and meta-analysis Maturitas (2015), http://dx.doi.org/10.1016/j.maturitas.2015.04.005 G Model ARTICLE IN PRESS MAT-6389; No of Pages 13 K.D Hill et al / Maturitas xxx (2015) xxx–xxx Table (Continued) Study Study Design Study Purpose Intervention Sample Size; % female; age (years); specific population; drop out Number of falls/fallers Intervention effect Follow-up Robertson et al., 2001 New Zealand RCT Determine the effectiveness of a trained nurse individually prescribing a home exercise program to reduce falls and injuries E: program was based on the Otago Exercise program and included individually prescribed balance and strength exercises during home visits, including a booster visit at month Participants completed strength exercises 3×/week and walked 2×/week for 12 months Compliance was monitored by postcard calendars, nurses telephoned participants in the months when they did not visit n = 240 (E: 121, C: 119); female: 67.5%; age: 80.95 yrs; general ≥75years; drop out: E: 6.6% (n = 8), C: 17.6% (n = 21) Falls at 12 months E: 80, C: 109 Falls reduction of 46% was found for exercise group 12 months n = 340 (E: 171; C: 169); female: 74%; age: 81.2 yrs; recently discharged from hospital; drop out: E: 7.0% (n = 12), C: 7.6% (n = 13) Falls at 12 months: E: 177 falls, C: 123 falls n = 40, (E: 19; C: 21); female: 62.5%; age: 81.9 yrs; older people living with Alzheimer’s Disease; drop out: E: 42.1% (n = 8), C: 14.3% (n = 3) Fallers at months E: (47%), C: (33%) Sherrington et al., 2014 Australia Suttanon et al., 2012 Australia RCT RCT Determine the effects of a home-based exercise program on mobility and falls among people recently discharged from hospital Determine the effectiveness of a home-based exercise program for people with Alzheimer’s Disease to improve balance, mobility and reduce risk of falls C: usual care E: Three PT delivered exercise program in participant’s homes 10 visits over 12 months, more visits at start of intervention Participants completed 20–30 balance and strength exercise of lower limb 6×/week for a year Exercises were based on WEBB exercise program, the PT described the level of intensity and repetitions Physical Activity Stage of Change model was used by the PTs to encourage on-going exercise, where appropriate weight belts or weighted vests were worn Participants used a log book to record exercises completed and any soreness from them Participants in both groups received a booklet on falls prevention C: Usual care E: program based on the Otago Exercise program – included standing balance and strength exercises and a walking programs Intervention of PT visits, and encouraged to exercise 5×/week Caregivers were also instructed how to the exercises and were asked to encourage regular exercise Between visits PTs followed-up with phone calls, compliance data were collected using a monthly exercise sheet which the PT reviewed at each home visit C: The control group were given the same number of home visits and phone calls as the intervention group, consisting of education and information sessions on dementia and ageing These were delivered by an OT No hospital admission from injurious falls for the exercise group, five for the control group Exercise group fell significantly more than the control group at 12 months 3, 12 months Using the Short Physical Performance Battery mobility was significantly better for the intervention group compared to the control group at 12 months Functional Reach improved significantly in the exercise group compared to the control group as did the Falls Risk for Older People – Community Score months Falls rate/1000 person days reduced by 33% for the exercise group, whereas the control group increased by around 89% Similar pattern was also seen for the change in proportion of fallers in the two groups Please cite this article in press as: Hill KD, et al Individualized home-based exercise programs for older people to reduce falls and improve physical performance: A systematic review and meta-analysis Maturitas (2015), http://dx.doi.org/10.1016/j.maturitas.2015.04.005 G Model ARTICLE IN PRESS MAT-6389; No of Pages 13 K.D Hill et al / Maturitas xxx (2015) xxx–xxx Table (Continued) Study Study Design Study Purpose Intervention Sample Size; % female; age (years); specific population; drop out Number of falls/fallers Intervention effect Follow-up Yang et al., 2012 Australia RCT Assessed the effectiveness of a home-based exercise program in older people with mild balance dysfunction E: based on the Otago Exercise program (see Campbell et al., 1997 above) with additional exercises from the Visual Health Information Balance and Vestibular Exercise kit if the therapist thought more challenging exercises were required Three visits: from PT: baseline, weeks and weeks Participants were asked to complete the exercises or walking 5×/week for months Exercise diaries were used to record performance n = 165; (E: 82, C: 83); female: 55.7%; age: 80.5 yrs; general ≥65 years; drop-out: E: 28% (n = 23), C: 25.3% (n = 21) Faller at months – E: 12 (20%); C: 18 (29%) Exercise group significantly more improved than the control group for: step width, functional reach, step test and activity levels months C: Control group were provided with a falls prevention booklet Note RCT = randomized controlled trial, PD = Parkinson’s Disease, E = exercise group; C = control group; ROM = range of movement, FR = functional Reach, OT = occupational therapist, PT = physiotherapist, ADLs = activities of daily living, SD = standard deviation, WHOQOL-BREF = World Health Organisation’s Quality of Life positive Sherrington and colleagues had the lowest dropout rate of 7.4% (25/340) over the one year study period Sample populations from the studies in the review included: Parkinson’s Disease [17], people living with dementia [25], hip fracture [23], older people recently discharged from hospital [15] and older people with no specific health problem [18–22,24,26,27] 3.2 Interventions The intervention period ranged from six weeks to two years, although longer term interventions were predominantly by phone call to continue motivation Participants randomized to the intervention (exercise) group were asked to complete the exercises daily [17,20], three to five times a week (includes strength days and aerobic days) [18,19,21–27], and six days a week [15] Seven of the studies were based on the Otago Exercise program, which includes strengthening exercises, balance exercises and a walking program [18,19,22,24–27] Ankle weights were used to progress the strength exercises over time and participants were given a booklet with illustrations of the exercises and instructions on how to complete them in case they had forgotten the explanation from the physiotherapist or nurse who delivered the program The Otago program required participants to perform approximately 30 of balance and strength activities three times a week and 30 walking twice a week (after the original study started with three times a week) [19] Strength exercises were predominantly lower body, balance was both static and dynamic, and stair climbing and range of movement exercises were also included One study combined the Otago program with another commercially available program to provide a greater range of balance challenging exercises (the Otago Plus program) [26] The Weight Bearing Exercise for Better Balance (WEBB) program was utilized in one study included in this review [15] A physiotherapist individually prescribed up to six exercises based on the participant’s physical performance assessment Exercises again were predominantly lower body specific including sit to stand, calf raises, step-ups, different stances that reduce base of support Table Assessment of risk of bias for included studies Study Selection bias Sequence generation Ashburn et al Campbell et al Campbell et al Clemson et al Orwig et al Lin et al Liu-Ambrose et al Robertson et al Sherrington et al Suttanon et al Yang et al Allocation concealment Performance bias Attrition bias Reporting bias Other bias Blinding of participants and personnel Incomplete outcome data Selective outcome reporting Free of other bias × × Note Bias was scored as low risk ( ), unclear (×), or high risk (᭹) Gardner et al was not included because it was not a RCT study design Please cite this article in press as: Hill KD, et al Individualized home-based exercise programs for older people to reduce falls and improve physical performance: A systematic review and meta-analysis Maturitas (2015), http://dx.doi.org/10.1016/j.maturitas.2015.04.005 G Model MAT-6389; No of Pages 13 ARTICLE IN PRESS K.D Hill et al / Maturitas xxx (2015) xxx–xxx Fig Study selection flow chart with eyes open and closed, stepping over objects, foot taps, lateral sidesteps and sideways walking [15] Other exercise program included in the review were: the LiFE program [20], the Exercise Plus program [23], and two others that were unnamed [17,21] Similar to the above programs they predominantly concentrated on lower body strength, balance and mobility exercises The philosophy behind Clemson et al’s LiFE program [20] was to include exercise that was not structured in nature and the philosophy behind completing the exercises was to incorporate them into usual daily activity such as standing on tip toes to reach for a cup in the kitchen, or bending knees to pick something up off the ground Because of this, participants were asked to perform these exercises daily in order for them to become a habit [20] The Exercise Plus program was intensively supervised by an exercise trainer, including up to 56 sessions in total (see Table for more details) The exercises were a combination of strength, aerobic and stretching and were completed three times a week for the aerobic and twice a week for the strength exercises which utilized thera-bands, ankle and wrist cuff weights [23] The exercise intervention used by Ashburn et al included six levels of progressive strength, range of movement, balance and walking exercises, again based on improving lower body performance [17] No equipment was described for this intervention Lin et al’s exercise program included stretching of all the major joints, and strength and balance exercises of the lower body Ankle weights were used to increase resistance and the exercises were completed three times a week [21] Please cite this article in press as: Hill KD, et al Individualized home-based exercise programs for older people to reduce falls and improve physical performance: A systematic review and meta-analysis Maturitas (2015), http://dx.doi.org/10.1016/j.maturitas.2015.04.005 G Model MAT-6389; No of Pages 13 ARTICLE IN PRESS K.D Hill et al / Maturitas xxx (2015) xxx–xxx Fig Forest plot of comparison: intervention vs control for number of fallers (studies with 12 month follow-up included) 3.3 Outcome measures 3.5 Quality of studies There were 17 outcome measures relating to falls, including number of falls or fallers (11 studies) [15,17–20,22–27], fall rate (per person year or week; four studies) [15,18,24,25], injurious falls [15,17–19,24], repeat falling [17,20], location of falls and falls by time period in study [15] Two studies measured physical activity using the Physical Activity Scale for the Elderly (PASE) [18–20,27] Balance, mobility and strength were all measured using many different tests Table shows the assessed potential bias in each study except Gardner et al [27] This study was not a RCT, therefore using the risk of bias tool was not appropriate Nine studies were assessed as having low risk of bias across all domains [15,17–20,23–26] However, the assessors deemed Lin et al [21] and Liu-Ambrose et al [22] as both unclear on allocation concealment The studies stated that randomization had occurred after baseline assessment, however no further detail was provided for how the group assignment was performed and by whom Overall, the 11 RCTs were regarded as high quality studies [15,17–26] 3.4 Dropout and adherence to home exercises 3.6 Effectiveness of intervention programs Study dropout rates ranged between 7.4% (n = 25)[15] and 32.2% (n = 49) [18] Eleven studies evaluated adherence to the exercise program, four methods to collect the data were used across the studies An average of 51.6% of the participants from 11 studies (who reported adherence) adhered to at least 50% of the exercises prescribed, this ranged between 25% completing three or more days of exercise [22] through to 81% of participants fully complying to exercising days a week [25] Not all studies could contribute data to the meta-analysis due to incomplete reporting in the published data For continuous outcomes, means and standard deviations for differences on outcome measures collected at baseline and at follow up were not available in most of the studies Therefore, the meta-analyses of continuous outcomes were performed using follow-up data for those studies in which no significant differences between the control group and intervention group were reported at baseline Ten studies had a Fig Forest plots of comparison: intervention vs control for (A) injuries requiring medical attention and (B) number of fractures Please cite this article in press as: Hill KD, et al Individualized home-based exercise programs for older people to reduce falls and improve physical performance: A systematic review and meta-analysis Maturitas (2015), http://dx.doi.org/10.1016/j.maturitas.2015.04.005 G Model MAT-6389; No of Pages 13 10 ARTICLE IN PRESS K.D Hill et al / Maturitas xxx (2015) xxx–xxx Fig Sensitivity analysis – forest plot of comparison: intervention vs control for number of fallers Fig Forest plot of comparison: intervention vs control for physical activity measured by PASE 12-month follow-up, two studies had shorter durations of two [17] and four months [21] 3.6.1 Falls Six studies reported number of falls [15,18–20,23,24], four number of fallers [17,25–27] and two reported fall incidence rates [21,22] for the intervention and control groups The total sample size for intervention and control groups in these studies was 1466 and 1054 participants, respectively The total number of falls reported in the intervention and control groups was 752 and 818 respectively The results of the meta-analyses for the outcomes number of fallers, number of injuries requiring medical attention and number of fractures resulting from a fall are shown in Figs and The study by Suttanon et al [25] reported data on number of fallers, however their data were not included in the meta-analysis because of a significant between-group difference in number of fallers at baseline Overall, at follow-up, there was no significant between-group difference in number of fallers (RR [95% CI] = 0.93 [0.72–1.21]) (Fig 2) No significant between-group difference in number of injuries requiring medical attention (RR [95% CI] = 0.96 [0.78–1.19]) (Fig 3A) and number of fractures (RR [95% CI] = 0.75 [0.40–1.41]) (Fig 3B) were found 3.6.1.1 Sensitivity analysis For the outcomes number of fallers and number of injuries requiring medical attention, sensitivity analysis was performed to explore possible changes on the meta-analyses results Specifically, we excluded the study by Sherrington et al [15] which included exclusively older people following hospital discharge This patient group has been shown to have a high falls rate [16] and may require different intervention approaches in isolation or together with an exercise program as they adjust to returning home often with changed function Removal of that study resulted in a significant between-group difference in number of fallers, with number of fallers at follow-up statistically lower in the exercise group compared to the control group (RR [95% CI] = 0.84 [0.72–0.99]) (Fig 4) Removal of this study was supported as the I2 value reduced to zero in the sensitivity analysis For the sensitivity analysis for number of injuries requiring medical attention, removal of the study by Sherrington et al [15] did not change the results (RR [95% CI] = 0.88 [0.59–1.10]) Sensitivity analyses were not performed for the other outcomes due to the limited number of studies included in the meta-analyses of these outcomes 3.6.2 Physical activity Two studies [18,20] reported data (at all assessment points) on physical activity using the PASE At 12 months follow-up, physical activity levels measured by the PASE were significantly higher in the intervention group compared to the control group (MD [95% CI] = 15.88 [7.80–27.02]) (Fig 5) 3.6.3 Balance Four studies [17,21,25,26] reported data on balance At follow up, functional reach was significantly higher in the intervention group compared to the control group (MD [95% CI] = 1.57 [0.37–2.76]) (Fig 6A) Three studies [15,25,26] reported data on the step test At follow-up, there was no significant difference between groups on performance using the step test (MD [95% CI] = 0.88 [−0.01–1.77]) (Fig 6B) 3.6.4 Muscle strength Three studies [15,20,26] reported data on knee extensor force At follow-up, knee extensor force was greater in the intervention group compared to the control group (SMD [95% CI] = 0.16 [0.00–0.33]) (Fig 7) 3.6.5 Mobility Two studies reported data on sit to stand [25,26] and two reported data on Timed Up and Go [22,25] Performance during the sit to stand test was better in the intervention group compared to the control group (MD [95% CI] = 0.71 [−1.42 to −0.00]) (Fig 8A) There was no significant difference between groups on performance during the Timed Up and Go (MD [95% CI] = 0.88 [−0.01–1.77]) (Fig 8B) Discussion In this review we have focussed on individualized home-based exercise programs to reduce falls Overall, there was no significant Please cite this article in press as: Hill KD, et al Individualized home-based exercise programs for older people to reduce falls and improve physical performance: A systematic review and meta-analysis Maturitas (2015), http://dx.doi.org/10.1016/j.maturitas.2015.04.005 G Model MAT-6389; No of Pages 13 ARTICLE IN PRESS K.D Hill et al / Maturitas xxx (2015) xxx–xxx 11 Fig Forest plot of comparison: intervention vs control for (A) functional reach and (B) step test Fig Forest plots of comparison: intervention vs control for (A) knee extensor force Fig Forest plots of comparison: intervention vs control for (A) sit to stand test and (B) Timed Up and Go difference between exercise and control groups for number of fallers, although issues of heterogeneity among the studies was a concern The sensitivity analysis which excluded one study investigating a high risk group (recent discharge from hospital) identified a significant reduction in number of fallers across the other four studies in this analysis Significant differences in favor of the intervention (home-based exercise) group were also identified for improved physical activity (PASE), balance (functional reach), quadriceps strength, and performance on sit to stand tests Although there were no significant differences on the other outcomes, including injurious falls, fractures, and other balance measures, almost all of these were in the direction of favoring the intervention group The included studies investigated a diverse range of samples, including older people with no specific health problems but increased falls risk, as well as a number of studies selected for populations with health conditions known to affect balance performance and increase falls risk, such as Parkinson’s disease, Alzheimer’s disease, and hip fracture These findings reinforce the value of individualized home-based exercise programs being promoted widely to older people living at home as one exercise approach to reduce falls risk and improve balance, strength and function Sherrington and colleagues [15] finding of a significant increase in number of fallers in the exercise group was in contrast to all of the other individualized home-based exercise studies This was Please cite this article in press as: Hill KD, et al Individualized home-based exercise programs for older people to reduce falls and improve physical performance: A systematic review and meta-analysis Maturitas (2015), http://dx.doi.org/10.1016/j.maturitas.2015.04.005 G Model MAT-6389; No of Pages 13 12 ARTICLE IN PRESS K.D Hill et al / Maturitas xxx (2015) xxx–xxx a complex, high risk group, with unique needs associated with recovering from an acute health problem, and the need to adjust to potential new additional functional limitations associated with their hospitalization This period immediately after discharge home has been shown to be associated with high risk of falls and re-admissions [16], and highlights that this particular patient group may require a multifactorial approach to reduce risk of falls Specifically, other important approaches including home assessment and modification, and education regarding safe integration back home while encouraging gradual increase in mobility and independence, warrant consideration rather than introducing a home-based exercise program in isolation Individualized home-based exercise programs have benefits over group-based exercise approaches, including the tailoring of exercise to individual needs and lifestyle preferences, participant autonomy, flexibility in timing of exercise, ability to break up exercise throughout the day, low cost to the individual, and no need for travel Innovative programs such as the LiFE program [20] have also aimed to break away from the concept of formal exercise time, to structuring exercise around activities of daily living, spread throughout the day Factors influencing higher levels of full or partial adherence to home-based exercise programs include programs with balance or walking exercise, moderate levels of home visit support, a physiotherapist guiding the intervention, and having home visit or telephone support [10] Use of intermittent re-assessment and feedback to participants, recording methods of participation that are intermittently reviewed (e.g exercise diaries) [28], structuring the exercises with lifestyle preferences of the individual, seeing a perceived effect of the program on physical and mental health and creating participant autonomy can also help improve participation [29] However, there are also some limitations to home-based programs, including the need for strong self-discipline to adhere to the program, and lack of a social element of the program (the latter is considered a positive element of group exercise program) There remain a number of important areas for future research to improve effective implementation of individualized home-based exercise programs to reduce falls for older people living in the community This review and meta-analysis has highlighted the diversity of measures utilized to report falls and other related outcomes, which limits the ability to incorporate multiple studies into meta-analyses and to produce specific exercise prescription guidelines There is a need for future researchers to consider using a core set of measures that are common with other studies, to enable incorporation in future meta-analyses Participation and adherence with home-based programs remains generally low, and innovative approaches are needed to support improved uptake and sustained participation in these programs Conclusion Individualized home-based exercise programs appear to reduce number of fallers (except in older people returning home from hospital), and to improve a number of other physical performance outcomes including improved balance, leg strength, function and physical activity for older people living in the community Strategies are needed to increase the reach of individualized home-based exercises, and subsequent participation and sustainability of these programs to achieve a range of positive outcomes Ethical approval No ethics approval was required for this research work, as it is a systematic review Contributors The study was conceptualized by KH and EB Literature search was done by EB SWH, FB and EB contributed towards the screening of papers for inclusion and quality rating of included papers Analysis was done by VC and EB KH, VC and EB drafted the manuscript All authors gave feedback on manuscript drafts They also reviewed and approved the final version of the manuscript Competing interests The authors declare no conflict of interest Funding No funding was received to support this review Provenance and peer review Commissioned and externally peer reviewed References [1] Tovell A, Harrison J, Pointer S Hospitalised injury in older Australians, 2011–12 Injury research and statistics series no 90 Cat no INJCAT 166 Canberra: AIHW; 2014 [2] Murray C, Vos T, Lozano R, et al Disability-adjusted life years (DALYs) for 291 diseases and injuries in 21 regions, 1990–2010: a systematic analysis for the Global Burden of Disease Study 2010 Lancet 2012;380(9859):2197–223 [3] Craig J, et al The high cost to health and social care of managing falls in older adults living in the community in Scotland Scott Med J 2013;58(4): 198–203 [4] Russell M, et al Falls risk and functional decline in older fallers discharged directly from emergency departments J Gerontol Ser A Biol Sci Med Sci 2006;61(10):1090–5 [5] Stel V, et al Consequences of falling in older men and women and risk factors for health service use and functional decline Age Ageing 2004;33(1):58–65 [6] Sherrington C, et al Exercise to prevent falls in older adults: an updated meta-analysis and best practice recommendations NSW Public Health Bull 2011;22(3–4):78–83 [7] Ashworth N, et al Home versus center based physical activity programs in older adults Cochrane Database Syst Rev 2005;1:CD004017 [8] Martin J, et al The effectiveness of physical therapist-administered groupbased exercise on fall prevention: a systematic review of randomized controlled trials J Geriatr Phys Ther 2013;36(4):182–93 [9] McPhate L, Simek E, Haines T Program-related factors are associated with adherence to group exercise interventions for the prevention of falls: a systematic review J Physiother 2013;59(2):81–92 [10] Simek E, McPhate L, Haines T Adherence to and efficacy of home exercise programs to prevent falls: a systematic review and meta-analysis of the impact of exercise program characteristics Prev Med 2012;55(4):262–75 [11] Liberati A, et al The PRISMA statement for reporting systematic reviews and meta-analyses of studies that evaluate health care interventions: explanation and elaboration PLoS Med 2009;6(7):e1000100 [12] Higgins J, et al The Cochrane Collaboration’s tool for assessing risk of bias in randomised trials Br Med J 2011;343:d5928 [13] Higgins J, Green S Cochrane handbook for systematic reviews of interventions 5.1.0; 2011 Available from: www.cochrane-handbook.org [14] DerSimonian R, Laird N Meta-analysis in clinical trials Control Clin Trials 1986;7(3):177–88 [15] Sherrington C, et al A post-hospital home exercise program improved mobility but increased falls in older people: a randomised controlled trial PLOS ONE 2014;9(9):e104412 [16] Mahoney J, et al Temporal association between hospitalization and rate of falls after discharge Arch Intern Med 2000;160(18):2788–95 [17] Ashburn A, et al A randomised controlled trial of a home based exercise programme to reduce the risk of falling among people with Parkinson’s disease J Neurol Neurosurg Psychiatry 2007;78:678–84 [18] Campbell A, et al Falls prevention over years: a randomized controlled trial in women 80 years and older Age Ageing 1999;28(6):513–8 [19] Campbell A, et al Randomised controlled trial of a general practice programme of home based exercise to prevent falls in elderly women Br Med J 1997;315(7115):1065–9 [20] Clemson L, et al Integration of balance and strength training into daily life activity to reduce rate of falls in older people (the LiFE study): randomised parallel trial Br Med J 2012;345:e4547 [21] Lin M-R, et al A randomized, controlled trial of fall prevention programs and quality of life in older fallers J Am Geriatr Soc 2007;55(4):499–506 Please cite this article in press as: Hill KD, et al Individualized home-based exercise programs for older people to reduce falls and improve physical performance: A systematic review and meta-analysis Maturitas (2015), http://dx.doi.org/10.1016/j.maturitas.2015.04.005 G Model MAT-6389; No of Pages 13 ARTICLE IN PRESS K.D Hill et al / Maturitas xxx (2015) xxx–xxx [22] Liu-Ambrose T, et al Otago home-based strength and balance retraining imporves executive functioning in older fallers: a randomized controlled trial J Am Geriatr Soc 2008;56(10):1821–30 [23] Orwig D, et al Delivery and outcomes of a yearlong home exercise program after hip fracture Arch Intern Med 2011;171(4):323–31 [24] Robertson M, et al Effectiveness and economic evaluation of a nurse delivered home exercise programme to prevent falls 1: Randomised controlled trial Br Med J 2001;322(7288):697–701 [25] Suttanon P, et al Feasibility, safety and preliminary evidence of the effectiveness of a home-based exercise programme for older people with Alzheimer’s disease: a pilot randomized controlled trial Clin Rehabil 2012;27(5): 427–738 13 [26] Yang X, et al Effectiveness of a targeted exercise intervention in reversing older people’s mild balance dysfunction: a randomized controlled trial Phys Ther 2012;92(1):24–37 [27] Gardner M, et al Application of a falls prevention program for older people to primary health care practice Prev Med 2002;34:546–53 [28] Suttanon P, et al Factors influencing commencement and adherence to a home-based balance exercise program for reducing risk of falls: perceptions of people with Alzheimer’s disease and their caregivers Int Psychogeriatr 2012;24(7):1172–82 [29] Simek E, et al What are the characteristics of home exercise programs that older adults prefer? A cross-sectional study Am J Phys Med Rehabil 2015 [Epub ahead of print] Please cite this article in press as: Hill KD, et al Individualized home-based exercise programs for older people to reduce falls and improve physical performance: A systematic review and meta-analysis Maturitas (2015), http://dx.doi.org/10.1016/j.maturitas.2015.04.005