Individualized home based exercise programs for older people to reduce falls and improve physical performance a systematic review and meta analysis

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
1.
2.

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 8 92663618; fax: +61 8 92663699.
E-mail addresses: (K.D. Hill), (S.W. Hunter), (F. Batchelor),
(V. Cavalheri), (E. Burton).

/>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), />

G Model
MAT-6389; No. of Pages 13

ARTICLE IN PRESS
K.D. Hill et al. / Maturitas xxx (2015) xxx–xxx

2

3.

4.
5.

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 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1. 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
2 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.

2. 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), />

G Model

ARTICLE IN PRESS

MAT-6389; No. of Pages 13

K.D. Hill et al. / Maturitas xxx (2015) xxx–xxx
Table 1

Search strategy (according to CINAHL terminology).
1
2
3
4
5
6
7
8
9
10
11
12
13
14

person* exercise ti,ab.
person* activity ti,ab.
individ* exercise ti,ab.
individ* active* ti,ab.
1 or 2 or 3 or 4
fall* ti,ab.
accidental fall ti,ab.
fall* prevent* ti,ab.
6 or 7 or 8
older ti,ab.
age* ti.ab
elder ti,ab
10 or 11 or 12
5 and 9 and 13

3

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 1 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 2 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.
3. Results
3.1. Study selection
Fig. 1 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 9 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), />

G Model

ARTICLE IN PRESS

MAT-6389; No. of Pages 13

K.D. Hill et al. / Maturitas xxx (2015) xxx–xxx

4
Table 2
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
1 h over 6 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 2
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 8
weeks and 6 months
for exercise group.

2 and 6
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).

6 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.

6 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 min
3×/week, and walk
outside 3×/week. Four,
1 h PT visits in first 8
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 4
times in 8 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 6 months
E2: Structured program
involved 7 exercise for
balance and 6 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 6 months.
C: two sessions, one
booster and 6
follow-up phone calls
comprised 12 gentle
exercises no change or
increase in intensity
was provided.

Fallers at 6
months:
E = 46/63 (73%).
C = 49/63 (78%).

Difference between
groups in FR and
quality of life was
significant at 6 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), />

G Model

ARTICLE IN PRESS

MAT-6389; No. of Pages 13

K.D. Hill et al. / Maturitas xxx (2015) xxx–xxx

5

Table 2 (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 6 months for
fall risk or functional
mobility.
There was a significant
difference between
groups for the response
inhibition (part of
Stroop Test).

6 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 0 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 6 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 2 weeks for 4 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 2
weeks and were provided
with falls prevention
pamphlets.
E: The Otago Exercise
program. First four visits
every 2 weeks and a final
(fifth visit at 6 months).
Exercises performed
3×/week, and walk for
30 min 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 3
trainer-supervised exercise

sessions per week for the
first 2 months, and then 2
per week for the next 2
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.

2 and 4
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), />

G Model

ARTICLE IN PRESS

MAT-6389; No. of Pages 13

K.D. Hill et al. / Maturitas xxx (2015) xxx–xxx

6
Table 2 (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 5
home visits, including a
booster visit at month 6.
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 6
months E: 5
(47%),
C: 6 (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 min 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 6 PT visits,
and encouraged to exercise
5×/week. Caregivers were
also instructed how to do
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.

6 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), />

G Model

ARTICLE IN PRESS

MAT-6389; No. of Pages 13

K.D. Hill et al. / Maturitas xxx (2015) xxx–xxx

7

Table 2 (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, 4 weeks and 8
weeks. Participants were
asked to complete the
exercises or walking
5×/week for 6 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 6
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.

6 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 min of balance and strength activities three
times a week and 30 min 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 3
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), />

G Model
MAT-6389; No. of Pages 13
8

ARTICLE IN PRESS
K.D. Hill et al. / Maturitas xxx (2015) xxx–xxx

Fig. 1. 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 2 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), />

G Model
MAT-6389; No. of Pages 13

ARTICLE IN PRESS
K.D. Hill et al. / Maturitas xxx (2015) xxx–xxx

9

Fig. 2. 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 3 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 5 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. 3. 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), />

G Model
MAT-6389; No. of Pages 13

10

ARTICLE IN PRESS
K.D. Hill et al. / Maturitas xxx (2015) xxx–xxx

Fig. 4. Sensitivity analysis – forest plot of comparison: intervention vs control for number of fallers.

Fig. 5. 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. 2 and 3.
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).
4. 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), />

G Model
MAT-6389; No. of Pages 13

ARTICLE IN PRESS
K.D. Hill et al. / Maturitas xxx (2015) xxx–xxx

11

Fig. 6. Forest plot of comparison: intervention vs control for (A) functional reach and (B) step test.

Fig. 7. Forest plots of comparison: intervention vs control for (A) knee extensor force.

Fig. 8. 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), />

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.

5. 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 2 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), />

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), />