Difference between groups in FR and quality of life was significant at 6 months 2 and 6 months Campbell et al., 1997 New Zealand effectiveness of a home exercise program compared with usu
Trang 1Maturitas
j ourna l h o me pa g e :w w w e l s e v i e r c o m / l o c a t e / m a t u r i t a s
Review
Keith D Hilla,c,∗, Susan W Hunterb, Frances Batchelorc, Vinicius Cavalheria,
Elissa Burtona
a r t i c l e i n f o
Keywords:
Exercise
Community
Elderly
Personalized
a b s t r a c t Thereisconsiderablediversityinthetypesofexerciseprogramsinvestigatedtoreducefallsinolder peo-ple.Thepurposeofthispaperwastoreviewtheeffectivenessofindividualized(tailored)home-based exerciseprogramsinreducingfallsandimprovingphysicalperformanceamongolderpeoplelivinginthe community.Asystematicreviewandmeta-analysiswasconductedofrandomizedorquasi-randomized trialsthatutilizedanindividualizedhome-basedexerciseprogramwithatleastonefallsoutcome mea-surereported.Singleinterventionexercisestudies,andmultifactorialinterventionswhereresultsforan exerciseinterventionwerereportedindependentlywereincluded.Tworesearchersindependentlyrated thequalityofeachincludedstudy.Of16,871papersidentifiedfromsixdatabases,12metallinclusion criteria(11randomizedtrialsandapragmatictrial).Studyqualityoverallwashigh.Samplesizesranged from40to981,participantshadanaverageage80.1years,andalthoughthemajorityofstudiestargeted thegeneralolderpopulation,severalstudiesincludedclinicalgroupsastheirtarget(Parkinson’sdisease, Alzheimer’sdisease,andhipfracture).Themeta-analysisresultsforthefivestudiesreportingnumber
offallersfoundnosignificanteffectoftheintervention(RR[95%CI]=0.93[0.72–1.21]),althoughwhen
asensitivityanalysiswasperformedwithonestudyofparticipantsrecentlydischargedfromhospital removed,thisresultwassignificant(RR[95%CI]=0.84[0.72–0.99]).Themeta-analysisalsofoundthat interventionledtosignificantimprovementsinphysicalactivity,balance,mobilityandmusclestrength Therewerenosignificantdifferencesformeasuresofinjuriousfallsorfractures
©2015ElsevierIrelandLtd.Allrightsreserved
Contents
1 Introduction 00
2 Methods 00
2.1 Objectives 00
2.2 Eligibilitycriteria 00
2.3 Informationsourcesandsearchstrategy 00
2.4 Studyselection 00
2.5 Datacollectionprocess 00
2.6 Studyquality 00
2.7 Dataanalysis 00
http://dx.doi.org/10.1016/j.maturitas.2015.04.005
Trang 23 Results 00
3.1 Studyselection 00
3.2 Interventions 00
3.3 Outcomemeasures 00
3.4 Dropoutandadherencetohomeexercises 00
3.5 Qualityofstudies 00
3.6 Effectivenessofinterventionprograms 00
3.6.1 Falls 00
3.6.2 Physicalactivity 00
3.6.3 Balance 00
3.6.4 Musclestrength 00
3.6.5 Mobility 00
4 Discussion 00
5 Conclusion 00
Ethicalapproval 00
Contributors 00
Competinginterests 00
Funding 00
Provenanceandpeerreview 00
References 00
1 Introduction
Fallsaretheleadingcauseofinjuryrelatedhospitalizationsin
Australia[1]andmanyothercountries,andthegreatest
percent-ageofthoseinjuredfromfallsarepeopleagedover65years.Falls
arealso the19thleading causeof disability-adjusted life years
lostgloballyacrossallhealthconditions,andtrendsindicatethey
willbecomeaneven strongercontributor todisability-adjusted
lifeyearsinthefuture [2].Hospitalizationfiguresonlyreflect a
minorityoftheimpactof fallsamongolderpeople.In a recent
epidemiologicalstudyinScotland,only20%ofthe294,000
peo-pleagedover65livinginthecommunitywhofellina12month
periodpresentedtomedicalservices,withonly16%presentingto
theEmergencyDepartment,and6%oftheseadmittedto
hospi-tal[3].Evenintheabsenceofinjury,othercommonsequelaeof
fallsthatcanimpactsubstantiallyonqualityoflifeandabilityto
liveindependentlyincludelossofconfidenceinmobility,reduced
activitylevel,depression,andimpairedbalanceandfunction[4,5]
Exerciseis awell-researched areaof fallsprevention.A
sys-tematic review and meta-analysis published in 2011 included
54 randomized trials covering community and residential care
settings(85%community)[6].Thisreviewandmeta-analysis
com-binedallstudiesirrespectiveofsettingandidentifiedthreemain
characteristicsofexerciseprogramsthatincreasedthelikelihood
oftheprogrambeingeffectiveinreducingfalls:(1)moderateor
highchallengetobalance;(2)atleast50hofexerciseequatingto
2haweekintensity;and(3)theexercisemustbeongoing,once
stoppedtheeffectislostquickly.Itisimportanttonotethatthere
aresubstantialdifferencesbetweensettingsintermsofpopulation,
environment,andhealth andcarestaffsupportthatnecessitate
considering settings separately An exploratory sub-analysis of
studiesundertakenonlyinresidentialcaresettingsinSherrington
andcolleagues’meta-analysisidentifiedanon-significant
reduc-tioninfallsfollowingtheexerciseintervention,highlightingthe
needforseparateanalysesbysetting.Thefocusofthecurrentpaper
islimitedtoexerciseprogramswithinthecommunitysetting(that
isexcludinghospitalandresidentialcarefacilities,bothlowand
highcare)
Evenin the community setting, there is considerable
diver-sityinthetypesofexerciseprogramsavailable,whereandhow
theyareimplemented,andtheirassociatedoutcomes.Oneofthe
importantdistinctionsaboutexerciseprogramsforolderpeopleis
whethertheyarecentre-based(i.e.theolderpersonneedstotravel
toanexternalvenuetoparticipateintheexerciseprogram,and
theprogramisoftengroup-based)orhome-based(i.e.the exer-ciseprogramisabletobeundertakenindividuallyintheirown home).Differentiatinghome-basedfromgroup-basedexercise pro-gramsisimportant,asthereissomeevidenceofdifferingoutcomes [7,8],differentadherencerates[7]anddifferentfactors influenc-ingpreferenceandparticipationinthesetwotypesofprogramby olderpeople[9,10].Afurtherimportantdistinctioniswhetherthe exerciseprogramisindividualized(i.e.tailoredtomeetthe spe-cificneedsofanindividual,intermsofbalance,mobility,function andco-morbidities)oragenericprogramwherethesameexercise programisprovidedtoallexercisers.Individualizedprogramsare morelikelytobeatasuitableleveltosafelystressbalanceand functioninamannerlikelytoachievehealthbenefits,andtobe monitoredandprogressedifperformancechangesovertime.The focusof thisreview isindividualizedhome-basedexercise pro-gramsforolderpeopleinthecommunitysetting,aimingtoreduce falls
2 Methods
2.1 Objectives Thepurposeofthissystematicreviewandmeta-analysiswasto determinetheeffectivenessofindividualizedhome-basedexercise programsforolderpeopleinthecommunitysettinginreducing falls,andimprovingsecondaryoutcomesofphysicalperformance includingphysicalactivity,balance,mobilityandstrength 2.2 Eligibilitycriteria
Thereviewwaslimitedtostudiesmeetingthefollowing eligi-bilitycriteria:
• Studyparticipants:
◦aged60yearsandover(atleast50%ofthesample),
◦livinginthecommunity
• A home-basedexercise program that is personalized or indi-vidualizedtotheolderperson’scapabilities(differentexercises selected for each participant based on assessment, exercises modified based onindividual progress or needs) and targets
areductionin falls(and/or)riskoffalls Home-basedexercise programswereincludediftheywereasingleintervention;or
ifa home-basedexerciseprogramwaspartofamultifactorial
Trang 3Table 1
intervention,usingafactorialdesign,withresultsfortheexercise
interventionreportedseparately
• Outcomemeasuresincludedoneormoreof:numberoffalls,rate
offalls,numberoffallers,ortimetofirstfall.Otheroutcomes
mayincludefearoffalling,function,physicalperformance(e.g
balanceorstrength),oradherencetotheexerciseintervention
• Study design: randomized controlled trials (RCT) and
quasi-experimentaltrials
• StudieswritteninEnglish
Wheretwoormorestudiesreportdatafromthesamesample,
onlyoneofthesestudieswasincludedinthemeta-analysis
2.3 Informationsourcesandsearchstrategy
Databases searched included Medline (ProQuest), CINAHL,
PubMed, PsycInfo, EMBASE and Scopus, from January 1974 to
December 2014 Reference lists of the identified papers were
scannedandonlypapersinEnglishwereincluded,nounpublished
data, books,conference proceedings, theses or poster abstracts
wereincluded.Thesearchstrategywasconductedusinganumber
ofkeywordsthatweretobeidentifiedinthetitleorabstractofthe
paper.Table1outlinesthesearchstrategyundertakeninCINAHL
Languageandsyntaxwereadapteddependentonthedatabase:for
example,PubMedallowedtitle/abstracttobesearched
simulta-neouslyhowevernotalldatabasesallowedthisandinthesecases
theabstractwassearched
2.4 Studyselection
The study selection was a three stage process: stage one
involvedoneauthor(EB)initiallyscanningthetitlesandabstracts
toexcludearticlesnotmeetingthecriteria.Stagetwowasafull
screeningof theabstractsby EB.Full articleswerescreenedby
twoauthors(EBandFB)toidentifywhicharticlesmetthe
inclu-sioncriteria,wheredisagreementoccurred,EBandFBreferredto
theinclusioncriteriaandstudyprotocolandcommunicateduntil
consensuswasachieved.Referencelists ofincludedpapers and
recentreviews(inparticular Sherringtonetal.’s[6]review and
meta-analysisofexerciseRCTswithfallsrelatedoutcomes)were
screenedforadditionalarticles,andthreeadditionalstudiesnot
foundduringthesearchwerealsoincludedastheymetthecriteria
WeusedthePRISMAchecklisttoensurethattheresultswere
reportedsystematically[11]
2.5 Datacollectionprocess
Eachoftheincludedstudieshadthefollowingdataextracted:
design,purpose, detailsof theintervention,samplesize,gender
percentages, average age, withdrawal rate, outcome measures, numberoffalls,effectoftheinterventionandlengthoffollow-up 2.6 Studyquality
TheCochrane’sCollaboration’s“riskofbiastool”wasusedby threeindependentresearchers(EB,SWH,FB)toassessthe method-ologicalqualityofeachpaper[12].Categoriesthatwereassessed includedsequencegeneration,allocationconcealment,participant and staffblinding, blindingof outcome assessment, incomplete outcomedata,selectiveoutcomereporting,andothersourcesof bias[12].Riskofbiasincludedthreedifferentlevelsofassessment:
“lowrisk”,“unclearrisk”,or“highrisk”ofbias[12] 2.7 Dataanalysis
Eachstudy wasdescribed outliningtheircharacteristics, the interventionandoutcomemeasuresused,adherencetothe exer-ciseinterventions,qualityofthestudiesandeffectivenessofthe interventionprograms(Tables2and3)
TheReviewManager(RevMan)version5.3wasusedto con-ductstatisticalanalysesandcreateforestplots[13].BoththeI2
statistic and visual inspectionof the forest plots were used to assessheterogeneity.Initially,arandom-effectsmodelwasused
tocalculatesummaryestimates.Whenstudieswerefoundtobe homogeneous,afixed-effectmodelwasapplied.WhenI2was>50%
arandom-effectmodelwasapplied.Fordichotomousoutcomes, resultsofstudiesweremeta-analyzedusingtheMantel-Haenszel’s fixedeffectsmodel[13],andriskratios(RR)withtheirrespective 95%confidenceintervals(CI)werecalculated.Forcontinuous out-comes,theresultsofstudiesweremeta-analyzedusingtheinverse varianceDerSimonianandLairdmethod[14],andeitherthemean differences (MD)orstandardized meandifferences(SMD), with theirrespective95%CI,werecalculated.Forcontinuousoutcomes,
ifmeansandstandarddeviationsfordifferencesonoutcome meas-urescollected atbaseline and atfollow up were not available, themeta-analysesofcontinuousoutcomeswereperformedusing follow-up dataonly,however, studieswithbaselinedifferences betweenthecontrolgroupandinterventiongroupwereomitted fromtheseanalyses
3 Results
3.1 Studyselection Fig.1showsthestudyselectionflowchart.Thesearch gener-ated16,871articlesfromthesixdatabases.Afterduplicatearticles wereremoved,3889articlesremained.Abstractsandsubsequently full text articles of those remaining at each latter stage were reviewedagainstreviewinclusioncriteria,followingwhich9 arti-clesremainedinthereview.Referencelistsforincludedpapersand recentexerciseandfallspreventionreviewswerescanned, includ-ingSherringtonetal.’s[6]updatedmeta-analysis,andthreefurther articleswereincluded
Elevenofthe12articlesincludedinthesystematicreviewwere RCTs[15,17–26].Theotherarticlewasapragmatictrial[27].The samplesizesrangedfrom40[25]to981[27],withanaverage sam-plesizeof250(Table2).Theaverageageacrossthe12 studies was80.1years,withanaverageagerangebetween72.2and84.1 years.Twothousand,ninehundredandninetynineparticipants completedbaselinetestingand2570completedpost-testingacross the12 studies,anaverageretentionrateof 82.24%.The largest dropoutratewasfoundforCampbelletal.’s(1999)studywhere only67.76%ofthestudypopulationwereretained[18].However, giventhestudyperiodwastwoyearsandthewomenparticipating wereaged,onaverageover80yearsofagethisretentionrateseems
Trang 4Table 2
Design
female; age (years); specific population;
drop out
Number of falls/fallers
Ashburn et al.,
2007
UK
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.
C: Usual care (contact with local PD nurse).
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%), Fallers at 6 months:
E = 46/63 (73%).
C = 49/63 (78%).
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.
Difference between groups in FR and quality of life was significant at 6 months
2 and 6 months
Campbell et al.,
1997
New Zealand
effectiveness of
a home exercise program compared with usual care.
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.
C: Research nurse made social visit 4 times in 8 weeks and phoned regularly.
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
Campbell et al.,
1999
New Zealand
effectiveness of
a home-based exercise program for older women over two years
See above intervention and sample same as Campbell et al., 1997.
Follow-up results are over the second year and two years combined.
n = 152 (E: 71, C: 81);
female: 100%;
age: 83.9yrs;
older women ≥
80 years;
drop out:
32.24% (n = 49).
2nd year only, E: 50, control:
68.
Total falls for two years, E:
138, C: 220.
Exercise program showed a significant reduction in falls.
24 months
Clemson et al.,
2012
Australia
Randomized Parallel Trial
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.
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).
Falls at 12 months: E1:
172; E2: 193;
C: 224.
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
Trang 5Table 2 (Continued)
Design
female; age (years); specific population;
drop out
Number of falls/fallers
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.
C: usual care.
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%).
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.
12 months
Lin et al., 2007
Taiwan
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.
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.
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.
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.
2 and 4 months
Liu-Ambrose
et al., 2008
Canada
effects of the Otago Exercise program on falls risk, mobility and executive functioning after 6 months
in older adults with a history
of falling.
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.
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).
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
Orwig et al.,
2011
United States
a 12 month home-based exercise program could improve outcomes for people with hip fracture.
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.
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.
No significant differences for any other outcome measures including falls.
2, 6, 12 months
Trang 6Table 2 (Continued)
Design
female; age (years); specific population;
drop out
Number of falls/fallers
Robertson
et al., 2001
New Zealand
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.
C: usual care.
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.
No hospital admission from injurious falls for the exercise group, five for the control group.
12 months
Sherrington
et al., 2014
Australia
effects of a home-based exercise program on mobility and falls among people recently discharged from hospital.
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.
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.
Exercise group fell significantly more than the control group at 12 months.
Using the Short Physical Performance Battery mobility was significantly better for the intervention group compared to the control group at 12 months.
3, 12 months
Suttanon et al.,
2012
Australia
effectiveness of
a home-based exercise program for people with Alzheimer’s Disease to improve balance, mobility and reduce risk of falls.
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.
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%).
Functional Reach improved significantly
in the exercise group compared to the control group as did the Falls Risk for Older People – Community Score.
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.
6 months
Trang 7Table 2 (Continued)
Design
female; age (years); specific population;
drop out
Number of falls/fallers
Yang et al.,
2012
Australia
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.
C: Control group were provided with a falls prevention booklet.
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
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.Sherringtonandcolleagueshadthelowestdropoutrateof
7.4%(25/340)overtheoneyearstudyperiod.Samplepopulations
fromthestudiesinthereviewincluded:Parkinson’sDisease[17],
peoplelivingwithdementia[25],hipfracture[23],olderpeople
recentlydischargedfromhospital[15]andolderpeoplewithno
specifichealthproblem[18–22,24,26,27]
3.2 Interventions
Theinterventionperiodrangedfromsixweekstotwoyears,
althoughlongerterminterventionswerepredominantlybyphone
calltocontinuemotivation.Participantsrandomizedtothe
inter-vention(exercise)groupwereaskedtocompletetheexercisesdaily
[17,20], threetofive timesa week(includes strengthdays and
aerobicdays)[18,19,21–27],andsixdaysaweek[15]
SevenofthestudieswerebasedontheOtagoExerciseprogram,
whichincludesstrengtheningexercises,balanceexercisesand a
walkingprogram[18,19,22,24–27].Ankle weightswereusedto
progressthestrengthexercisesover timeandparticipantswere
given a booklet with illustrationsof the exercises and instruc-tionsonhowto completethem incase theyhad forgotten the explanationfromthephysiotherapistornursewhodeliveredthe program The Otago program required participants to perform approximately 30min of balance and strength activities three timesaweekand30minwalkingtwiceaweek(aftertheoriginal studystarted withthree timesa week)[19] Strengthexercises were predominantly lower body, balance was both static and dynamic, and stair climbing and range of movement exercises werealsoincluded.OnestudycombinedtheOtagoprogramwith another commercially available program to provide a greater rangeofbalancechallengingexercises(theOtagoPlusprogram) [26]
TheWeightBearingExercise forBetterBalance(WEBB) pro-gramwas utilized inone studyincluded in this review[15].A physiotherapistindividuallyprescribeduptosixexercisesbasedon theparticipant’sphysicalperformanceassessment.Exercisesagain were predominantly lowerbody specific includingsit tostand, calfraises,step-ups,differentstancesthatreducebaseofsupport
Table 3
Sequence generation
Allocation concealment
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 (䊉) Gardner et al was not included because it was not a RCT study design.
Trang 8Fig 1.Study selection flow chart.
witheyesopenandclosed,steppingoverobjects,foottaps,lateral
sidestepsandsidewayswalking[15]
Otherexerciseprogramincludedinthereviewwere:theLiFE
program [20], the Exercise Plus program [23], and two others
thatwereunnamed[17,21].Similartotheaboveprogramsthey
predominantlyconcentratedonlowerbodystrength,balanceand
mobilityexercises.ThephilosophybehindClemsonetal’sLiFE
pro-gram[20]wastoincludeexercisethatwasnotstructuredinnature
andthephilosophybehindcompletingtheexerciseswasto
incor-poratethemintousualdailyactivitysuchasstandingontiptoes
toreachforacupinthekitchen,orbendingkneestopick
some-thingupofftheground.Becauseofthis,participantswereasked
toperformtheseexercises dailyinorderfor themtobecomea
habit[20]
TheExercise Plusprogramwasintensively supervisedbyan exercisetrainer,includingupto56sessionsintotal(seeTable2for moredetails).Theexerciseswereacombinationofstrength, aero-bicandstretchingandwerecompletedthreetimesaweekforthe aerobicandtwiceaweekforthestrengthexerciseswhichutilized thera-bands,ankleandwristcuffweights[23].Theexercise inter-ventionusedbyAshburnetal.includedsixlevelsofprogressive strength,rangeofmovement,balanceandwalkingexercises,again basedonimprovinglowerbodyperformance[17].Noequipment wasdescribed forthisintervention.Lin etal’sexerciseprogram includedstretchingofallthemajorjoints,andstrengthandbalance exercisesofthelowerbody.Ankleweightswereusedtoincrease resistanceandtheexerciseswerecompletedthreetimesaweek [21]
Trang 9Fig 2.Forest plot of comparison: intervention vs control for number of fallers (studies with 12 month follow-up included).
3.3 Outcomemeasures
Therewere17outcome measuresrelating tofalls,including
numberoffallsorfallers(11studies)[15,17–20,22–27],fallrate
(perpersonyearorweek;fourstudies)[15,18,24,25],injuriousfalls
[15,17–19,24],repeatfalling[17,20],locationoffallsandfallsby
timeperiodinstudy[15].Twostudiesmeasuredphysicalactivity
usingthePhysicalActivityScalefortheElderly(PASE)[18–20,27]
Balance,mobilityandstrengthwereallmeasuredusingmany
dif-ferenttests
3.4 Dropoutandadherencetohomeexercises
Studydropoutratesrangedbetween7.4%(n=25)[15]and32.2%
(n=49)[18].Elevenstudiesevaluatedadherencetotheexercise
program,fourmethods tocollectthedatawereusedacrossthe
studies.Anaverageof51.6%oftheparticipantsfrom11studies
(whoreportedadherence)adheredtoatleast50%oftheexercises
prescribed,thisrangedbetween25%completingthreeormoredays
ofexercise[22]throughto81%ofparticipantsfullycomplyingto
exercising5daysaweek[25]
3.5 Qualityofstudies Table3showstheassessedpotentialbiasineachstudyexcept Gardneretal.[27].ThisstudywasnotaRCT,thereforeusingthe riskofbiastoolwasnotappropriate.Ninestudieswereassessedas havinglowriskofbiasacrossalldomains[15,17–20,23–26] How-ever,theassessorsdeemedLinetal.[21]andLiu-Ambroseetal [22]asbothunclearonallocationconcealment.Thestudiesstated thatrandomizationhadoccurredafterbaselineassessment, how-evernofurtherdetailwasprovidedforhowthegroupassignment wasperformedandbywhom.Overall,the11RCTswereregarded
ashighqualitystudies[15,17–26] 3.6 Effectivenessofinterventionprograms Notallstudiescouldcontributedatatothemeta-analysisdue
toincompletereportinginthepublisheddata.Forcontinuous out-comes,meansandstandarddeviationsfordifferencesonoutcome measurescollectedatbaselineandatfollowupwerenotavailable
inmostofthestudies.Therefore,themeta-analysesofcontinuous outcomeswereperformedusingfollow-updataforthosestudies
inwhichnosignificantdifferencesbetweenthecontrolgroupand interventiongroupwerereportedatbaseline.Tenstudieshada
Trang 10Fig 4. Sensitivity analysis – forest plot of comparison: intervention vs control for number of fallers.
12-monthfollow-up,twostudieshadshorterdurationsoftwo[17]
andfourmonths[21]
3.6.1 Falls
Sixstudiesreportednumberoffalls[15,18–20,23,24],four
num-ber of fallers [17,25–27] and two reported fall incidence rates
[21,22]for theintervention andcontrol groups Thetotal
sam-plesizeforinterventionandcontrolgroupsinthesestudieswas
1466and1054participants,respectively.Thetotalnumberoffalls
reportedintheinterventionandcontrolgroupswas752and818
respectively
Theresultsofthemeta-analysesfortheoutcomesnumberof
fallers,numberofinjuriesrequiringmedicalattentionand
num-beroffracturesresultingfromafallareshowninFigs.2and3
ThestudybySuttanonetal.[25]reporteddataonnumberof
fall-ers,howevertheirdatawerenot includedinthemeta-analysis
becauseof asignificantbetween-groupdifferenceinnumber of
fallersatbaseline.Overall,atfollow-up,therewasnosignificant
between-groupdifferenceinnumberoffallers(RR[95%CI]=0.93
[0.72–1.21])(Fig.2)
Nosignificantbetween-groupdifferenceinnumberofinjuries
requiring medical attention (RR [95% CI]=0.96 [0.78–1.19])
(Fig.3A)andnumberoffractures(RR[95%CI]=0.75[0.40–1.41])
(Fig.3B)werefound
3.6.1.1 Sensitivityanalysis Fortheoutcomesnumberoffallersand
numberofinjuriesrequiringmedicalattention,sensitivityanalysis
wasperformedtoexplorepossiblechangesonthemeta-analyses
results.Specifically,weexcludedthestudybySherringtonetal
[15]which includedexclusivelyolderpeoplefollowing hospital
discharge.Thispatientgrouphasbeenshowntohaveahighfalls
rate[16] and may requiredifferent intervention approaches in
isolationortogetherwithanexerciseprogramastheyadjustto
returninghomeoftenwithchangedfunction.Removalofthatstudy
resultedin asignificantbetween-groupdifferencein numberof
fallers,withnumberoffallersatfollow-upstatisticallylowerinthe
exercisegroupcomparedtothecontrolgroup(RR[95%CI]=0.84
[0.72–0.99])(Fig.4).Removalofthisstudywassupportedasthe
I2 valuereducedtozerointhesensitivityanalysis.Forthe
sensi-tivityanalysisfornumberofinjuriesrequiringmedicalattention,
removalofthestudybySherringtonetal.[15]didnotchangethe results(RR[95%CI]=0.88[0.59–1.10]).Sensitivityanalyseswere notperformedfortheotheroutcomesduetothelimitednumber
ofstudiesincludedinthemeta-analysesoftheseoutcomes 3.6.2 Physicalactivity
Twostudies[18,20]reporteddata(atallassessmentpoints)on physicalactivityusingthePASE.At12monthsfollow-up,physical activitylevelsmeasuredbythePASEweresignificantlyhigherin theinterventiongroupcomparedtothecontrolgroup(MD[95% CI]=15.88[7.80–27.02])(Fig.5)
3.6.3 Balance Fourstudies[17,21,25,26]reporteddataonbalance.Atfollow
up, functional reach was significantly higher in the interven-tion group compared to the control group (MD [95% CI]=1.57 [0.37–2.76])(Fig.6A).Threestudies[15,25,26]reporteddataonthe steptest.Atfollow-up,therewasnosignificantdifferencebetween groups onperformance using the step test (MD [95%CI]=0.88 [−0.01–1.77])(Fig.6B)
3.6.4 Musclestrength Threestudies[15,20,26]reporteddataonkneeextensorforce
Atfollow-up, kneeextensorforce wasgreater in the interven-tiongroupcompared tothecontrol 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 reporteddataonTimedUpandGo[22,25].Performanceduring thesittostandtest wasbetterin theinterventiongroup com-paredtothecontrolgroup(MD[95%CI]=0.71[−1.42to−0.00]) (Fig.8A).Therewasnosignificantdifferencebetweengroupson performance during the Timed Up and Go (MD [95% CI]=0.88 [−0.01–1.77])(Fig.8B)
4 Discussion
Inthisreviewwehavefocussedonindividualizedhome-based exerciseprogramstoreducefalls.Overall,therewasnosignificant