pollock 07 phys treatment
Post on 29-Sep-2015
215 Views
Preview:
DESCRIPTION
TRANSCRIPT
-
Clinical Rehabilitation 2007; 21: 395410
Physiotherapy treatment approaches for therecovery of postural control and lower limbfunction following stroke: a systematic reviewAlex Pollock Stroke Therapy Evaluation Programme, Academic Section of Geriatric Medicine, Glasgow Royal Infirmary,Glasgow, Gillian Baer Department of Physiotherapy, Queen Margaret University College, Edinburgh, Peter LanghorneAcademic Section of Geriatric Medicine, Glasgow Royal Infirmary, Glasgow and Valerie Pomeroy Rehabilitation andAgeing, Geriatric Medicine, St Georges University of London, London, UK
Received 30th September 2006; accepted 6th October 2006.
Objectives: To determine whether there is a difference in global dependency andfunctional independence in patients with stroke associated with differentapproaches to physiotherapy treatment.Data sources: We searched the Cochrane Stroke Group Trials Register (lastsearched May 2005), Cochrane Central Register of Controlled Trials (CENTRAL)(Cochrane Library Issue 2, 2005), MEDLINE (1966 to May 2005), EMBASE (1980to May 2005) and CINAHL (1982 to May 2005). We contacted experts andresearchers with an interest in stroke rehabilitation.Review methods: Inclusion criteria were: (a) randomized or quasi-randomized controlled trials; (b) adults with a clinical diagnosis of stroke; (c) physiotherapytreatment approaches aimed at promoting postural control and lower limb func-tion; (d) measures of disability, motor impairment or participation. Two independentreviewers categorized identified trials according to the inclusion/exclusion criteria,documented the methodological quality and extracted the data.Results: Twenty trials (1087 patients) were included in the review. Comparisonsincluded: neurophysiological approach versus other approach; motor learningapproach versus other approach; mixed approach versus other approach for theoutcomes of global dependency and functional independence. A mixed approachwas significantly more effective than no treatment control at improving functionalindependence (standardized mean difference (SMD) 0.94, 95% confidence interval(CI) 0.08 to 1.80). There were no significant differences found for any othercomparisons.Conclusions: Physiotherapy intervention, using a mix of components from differentapproaches is more effective than no treatment control in attaining functionalindependence following stroke. There is insufficient evidence to conclude that anyone physiotherapy approach is more effective in promoting recovery of disabilitythan any other approach.
Address for correspondence: Alex Pollock, Stroke TherapyEvaluation Programme, Academic Section of Geriatric Medicine,Room 34, Level 3, University Block, Queen Elizabeth Building, 10Alexandra Parade, Glasgow Royal Infirmary University NHS Trust,Glasgow G31 2ER, UK. e-mail: alex@strokerehab.fsnet.co.uk 2007 SAGE Publications 10.1177/0269215507073438
-
396 A Pollock et al.
Background
There are several different approaches to physiothera-py treatment following stroke. Prior to the 1940s theseprimarily consisted of corrective exercises based onorthopaedic principles related to the contraction andrelaxation of muscles, with the emphasis placed onregaining function by compensating with the unaf-fected limbs.1,2 In the 1950s and 1960s techniquesbased on available neurophysiological knowledgewere developed, including the methods of Bobath,3,4Brunnstrm,5 Rood6 and the ProprioceptiveNeuromuscular Facilitation (PNF) approach.7,8 In the1980s the potential importance of neuropsychologyand motor learning was highlighted9,10 and the motorlearning, or relearning, approach was proposed.11This approach suggested that active practice of con-text-specific motor tasks with appropriate feedbackwould promote learning and motor recovery.1117
The practical application of these approachesresults in substantial differences in patient treatment.Approaches based on neurophysiological principleshave traditionally primarily involved the physiothera-pist moving the patient through patterns of movement,with the therapist acting as problem solver and deci-sion maker and the patient being a relatively passiverecipient.18 In direct contrast, the motor learningapproaches emphasize the importance of activeinvolvement by the patient,11 while orthopaedicapproaches emphasize muscle strengthening tech-niques and compensation with the non-paretic side. Atpresent, the Bobath Approach, based on neurophysio-logical principles, remains probably the most widelyused approach in Sweden,19 Australia20 and theUK.2123
Physiotherapists often seek evidence relating toglobal approaches to treatment of stroke patients,rather than evidence in support of individual treat-ments. The evaluation of this research evidence isoften difficult due to poor description and documenta-tion of the approaches investigated. Often the treat-ment approach is vaguely described as conventionalor traditional (e.g. refs 2430) and few other detailsare available. Although questionnaire-based studiesdo demonstrate that physiotherapists often have apreference for a particular approach, there is present-ly no convincing evidence to support any specificphysiotherapy treatment approach.21,31
This report details key findings from a CochraneSystematic Review.32 Readers are referred to the full
Cochrane Review for additional details. The objectivewas to determine if there is a difference in the recov-ery from disability in patients with stroke if physio-therapy treatment is based on any one of orthopaedicor neurophysiological or motor learning principles, oron a mixture of these treatment principles.
Methods
Types of studiesControlled trials were included if participants were
randomly or quasi-randomly assigned to one of two ormore treatment groups. Trials with or without blind-ing of participants, physiotherapists and assessorswere included.
Types of participantsTrials were included if participants were adults
(over 18 years) with a clinical diagnosis of stroke.33Participants with diagnosis of either ischaemic strokeor haemorrhagic stroke were included (confirmationof the clinical diagnosis using imaging was notcompulsory).
Types of interventionsPhysiotherapy treatment approaches that were
aimed at promoting the recovery of postural control(balance during the maintenance of a posture, restora-tion of a posture or movement between postures) andlower limb function (including gait) were included.Interventions that had a more generalized stated aim,such as improving functional ability, were alsoincluded. Treatment approaches that were primarilyaimed at promoting recovery of upper limb movementor function were excluded. Studies of specific inter-ventions, such as electrical stimulation, biofeedbackand treadmill training, were excluded.
Types of outcome measuresPrimary outcomes were defined as measures of dis-
ability. Relevant measures of disability were prestatedas (1) global dependency scales or (2) functional inde-pendence in mobility. (A number of secondary out-comes were also included; these are detailed in theCochrane Review.32)
-
Physiotherapy treatment approaches for stroke 397
Search strategyFull details of the search strategy, including the full
bibliographic search history for electronic databases,are provided within the Cochrane Review.32
The searching was based on the strategy developedby the Cochrane Stroke Group and was done usingintervention-based search strategies developed in con-sultation with the Cochrane Stroke Group TrialsSearch Co-ordinator. Searching included: CochraneStroke Group Trials Register (May 2005); CochraneCentral Register of Controlled Trials (CENTRAL)(Cochrane Library Issue 2, 2005); MEDLINE (1966to May 2005); EMBASE (1980 to May 2005);CINAHL (1982 to May 2005). Experts and authorswere contacted and asked if they knew of any addi-tional, unpublished or on-going trials, and the refer-ence lists of all trials found using the above searchmethods were searched.
Search resultsThe electronic searching resulted in 8408 potentially
relevant trials. One reviewer (AP) eliminated obvious-ly irrelevant studies based on titles and, where avail-able, abstracts. This eliminated 8161 studies, leaving247 potentially relevant trials. Examination of thereference lists of these potential trials, and communi-cation with known experts and colleagues, added afurther 18 studies, making a total of 265 potentiallyrelevant trials.
Two independent reviewers (AP, GB) read theabstracts for these 265 studies. Of these, 184 (69%)were classified as relevant or possibly relevant.
Titles, introduction and methods sections of the184 possible trials were independently scrutinized bytwo reviewers (AP, GB). Based on a detailed writtendescription (which was based on the available litera-ture, and which had been discussed between allreviewers to ensure consensus) of the classification ofphysiotherapy approaches based on motor learning,neurophysiological or orthopaedic principles (seeCochrane Review32 for details) reviewers independ-ently classified the interventions administered in eachtrial. Any disagreements were resolved through dis-cussion involving a third reviewer (PL), and furtherinformation was obtained from trialists where neces-sary (and possible).
This review process led to the identification of 20relevant trials to be included in this review: Dean andShepherd,34 Dean et al.,35 Duncan et al.,36 Duncan
et al.,37 Gelber et al.,38 Green et al.,39 Hesse et al.,40Howe et al.,41 Langhammer and Stranghelle,42Lincoln,43 McClellan and Ada,44 Mudie et al.,45Ozdemir et al.,46 Pollock,47 Richards et al.,48 Salbachet al.,49 Stern et al.,24 Wade et al.,50 Wang et al.51 andWellmon and Newton.52
Brief descriptions of the included studies can befound in Table 1; detailed descriptions are publishedin the Cochrane Review.32
Methodological quality and data extractionTwo independent reviewers (AP, GB) judged the
methodological quality of studies and extracted data,with any disagreements resolved through discussioninvolving a third reviewer. Trial authors were contactedfor clarification where necessary.
The following quality criteria were documented:randomization (allocation concealment); baselinecomparison of groups; blindness of recipients andproviders of care to treatment group/study aims;blindness of outcome assessor; possibility of contam-ination/co-intervention by therapists providing inter-vention; completeness of follow-up; other potentialconfounders.
Data extraction documented (where possible): trialsetting (e.g. hospital, community); details of partici-pants (e.g. age, gender, side of hemiplegia, strokeclassification, comorbid conditions, premorbid dis-ability); inclusion and exclusion criteria; all assessedoutcomes.
Details of the methodological quality and dataextraction from individual trials are fully presentedwithin the Cochrane Review.32
Data analysis was carried out using CochraneRevMan software. Standardized mean differences(SMD) and 95% confidence intervals (CI) were calcu-lated, using a random effects model, for all outcomesanalysed.
Results
The 20 included trials randomized 1087 patients.Three of these studies (78 patients) have no dataincluded in any of the review analyses: we wereunable to obtain the data from the first phase of thestudy by Hesse et al.40 (n 22); and Wellmon andNewton52 (n 21) and Howe et al.41 (n 35) reported
-
398 A Pollock et al.Ta
ble
1S
umm
ary
of c
hara
cter
istic
s of
incl
uded
stu
dies
Tria
lD
esig
nA
lloca
tion
Num
ber
of
Incl
usio
n O
utco
me
Trea
tmen
t O
utco
me:
O
utco
me:
N
otes
conc
ealm
ent
part
icip
ants
cr
iteria
: Tim
e as
sess
or
grou
psgl
obal
fu
nctio
nal
rand
omiz
edsi
nce
stro
kebl
inde
d to
de
pend
ency
inde
pend
ence
grou
p al
loca
tion?
Dea
n 19
9734
RC
T (b
lock
ed
Ade
quat
e20
1
year
Yes
Mot
or
(non
e)(n
one)
Sec
onda
ry
rand
omiz
atio
n)le
arni
ng
outc
omes
P
lace
boin
clud
ed in
C
ochr
ane
revi
ewD
ean
2000
35R
CT
(mat
ched
Ade
quat
e12
3
mon
ths
Yes
Mot
or
(non
e)(n
one)
Sec
onda
ry
pairs
le
arni
ng
outc
omes
ra
ndom
ized
)P
lace
boin
clud
ed in
C
ochr
ane
revi
ewD
unca
n R
CT
(met
hod
Ade
quat
e20
309
0 da
ysU
ncle
arM
ixed
B
arth
el
Fugl
-Mey
er
1998
36of
ran
dom
-C
ontr
olIn
dex
mot
or s
core
izat
ion
not
Law
ton
stat
ed)
Inst
rum
enta
lA
DL
Dun
can
RC
T (b
lock
ed
Ade
quat
e10
030
150
day
sYe
sM
ixed
(n
one)
Fugl
-Mey
er
2003
37ra
ndom
-C
ontr
ol(lo
wer
lim
b)iz
atio
n)G
elbe
r R
CT
(met
hod
Unc
lear
27
1 m
onth
No
Neu
roph
ysio
logi
cal
(non
e)Fu
nctio
nal
1995
38of
ran
dom
-(N
DT)
Inde
pend
ence
izat
ion
not
Ort
hopa
edic
Mea
sure
stat
ed)
(tra
ditio
nal
(FIM
)fu
nctio
nal
retr
aini
ng)
Gre
en
RC
T (b
lock
ed
Ade
quat
e17
0
1 ye
arYe
sM
ixed
Bar
thel
Riv
erm
ead
2002
39ra
ndom
-co
mm
unity
Inde
xM
obili
ty
izat
ion)
phys
ioth
erap
y,In
dex
usin
g Fr
ench
ay
prob
lem
A
ctiv
ities
so
lvin
g
Inde
x.ap
proa
ch
Con
trol
no in
terv
entio
nH
esse
S
ingl
e-su
bjec
tIn
adeq
uate
22ch
roni
c U
ncle
arB
obat
h (n
one)
(non
e)N
o da
ta
1998
40de
sign
, with
stro
ke
(wal
king
av
aila
ble
rand
om o
rder
with
of
allo
catio
nB
obat
h to
3 in
terv
en-
faci
litat
ion)
tions
Con
trol
(w
alki
ngw
ith n
o
-
Physiotherapy treatment approaches for stroke 399in
terv
entio
n)A
id (w
alki
ngw
ith
wal
king
stic
k)H
owe
RC
T (b
lock
A
dequ
ate
35A
cute
Ye
sM
ixed
(n
one)
(non
e)N
o ou
tcom
es20
0541
rand
omi-
(Neu
roph
y-in
clud
ed in
za
tion)
siol
ogic
alan
alys
is
all
m
otor
ou
tcom
esle
arni
ng)
wer
e sp
ecifi
c N
euro
phy-
to g
oal o
f si
olog
ical
late
ral w
eigh
t tr
ansf
eren
ceLa
ngha
mm
erR
CT
Ade
quat
e61
Not
sta
ted
Yes
Neu
roph
ysio
logi
cal
Bar
thel
Mot
or20
0042
(str
atifi
ed
(Bob
ath)
In
dex
Ass
essm
ent
acco
rdin
g M
otor
S
cale
Sod
ring
to g
ende
rle
arni
ngM
otor
an
d si
de
Eva
luat
ion
of le
sion
)Li
ncol
n R
CT
Ade
quat
e12
0
2 w
eeks
Yes
Neu
roph
ysio
logi
cal
Bar
thel
Mot
or20
0343
(blo
cked
(B
obat
h)
Inde
xA
sses
smen
tra
ndom
i-M
otor
E
xten
ded
Sca
le
zatio
n)le
arni
ngA
ctiv
ities
Riv
erm
ead
of D
aily
Mot
orLi
ving
Ass
essm
ent
scal
eM
cCle
llan
RC
TA
dequ
ate
26
18Ye
sM
otor
(non
e)M
otor
2004
44m
onth
sle
arni
ngA
sses
smen
tP
lace
boS
cale
cont
rol
(item
5)
(mot
orle
arni
ng,
uppe
rlim
b)M
udie
R
CT
Ade
quat
e40
Rec
ent
Yes
Feed
back
Bar
thel
(n
one)
Dat
a fr
om20
0245
only
Inde
xfe
edba
ckM
otor
only
gro
uple
arni
ngno
t us
ed(t
ask
rela
ted
trai
ning
)N
euro
phys
iolo
gica
l
(Bob
ath)
Con
trol
(no
trea
tmen
t)
(conti
nued
)
-
400 A Pollock et al.
Ozd
emir
Qua
si-r
ando
mIn
adeq
uate
60N
ot s
tate
dN
oM
ixed
(non
e)Fu
nctio
nal
Inte
nsity
of
2001
46co
ntro
lled
(ort
hopa
edic
Inde
pend
ence
the
2tr
ial (
alte
rnat
e
neur
ophy
siol
ogic
al)
Mea
sure
inte
rven
tions
allo
catio
nC
ontr
ol(F
IM)
varie
d ac
cord
ing
toor
der
ofen
try
tost
udy)
Pol
lock
R
CT
(blo
cked
A
dequ
ate
28
6 w
eeks
No
Neu
roph
ysio
logi
cal
Bar
thel
(n
one)
Con
side
rabl
e19
9847
rand
omi-
(Bob
ath)
Inde
xnu
mbe
rs o
fza
tion,
with
M
ixed
dr
op-o
uts
not
2 co
ntro
l (n
euro
phys
iolo
gica
lfo
llow
ed-u
p(n
euro
phys
):1
mot
or
inte
rven
tion
lear
ning
)(m
ixed
))R
icha
rds
RC
T A
dequ
ate
270
7 da
ysYe
sE
arly
M
ixed
Bar
thel
Fu
gl-M
eyer
Ana
lysi
s19
9348
(str
atifi
ed
Early
N
euro
-In
dex
mot
orba
sed
onbl
ocke
d ph
ysio
logi
cal
asse
ssm
ent
com
paris
onra
ndom
i-(B
obat
h)of
2 e
arly
za
tion,
C
onve
ntio
nal
grou
ps, a
sst
ratifi
ed
Neu
roph
ysio
logi
cal
thes
e 2
acco
rdin
g gr
oups
are
to p
rogn
ostic
(Bob
ath)
com
para
ble
cate
gory
)in
ter
ms
oftim
ing
and
inte
nsity
Sal
bach
R
CT
Ade
quat
e91
1
year
Yes
Mot
or
(non
e)(n
one)
Sec
onda
ry20
0449
(str
atifi
ed
lear
ning
outc
omes
bloc
ked
(mob
ility
)in
clud
ed in
rand
omi-
Pla
cebo
C
ochr
ane
zatio
n,
cont
rol
revi
ewst
ratifi
ed
(mot
or
acco
rdin
g to
le
arni
ng,
3 le
vels
of
uppe
r w
alki
ng
limb)
defic
it)S
tern
Q
uasi
-ran
do-
Inad
equa
te62
Not
sta
ted
Unc
lear
Ort
hopa
edic
(non
e)Fu
nctio
nal
1970
24m
ized
tria
l M
ixed
st
atus
(50
patie
nts
(Ort
hopa
edic
(ada
pted
wer
e
Neu
roph
ysio
logi
cal)
from
rat
ing
orig
inal
ly
scal
e)
Tab
le 1
cont
inue
d
Tria
lD
esig
nA
lloca
tion
Num
ber
of
Incl
usio
n O
utco
me
Trea
tmen
t O
utco
me:
O
utco
me:
N
otes
conc
ealm
ent
part
icip
ants
cr
iteria
: Tim
e as
sess
or
grou
psgl
obal
fu
nctio
nal
rand
omiz
edsi
nce
stro
kebl
inde
d to
de
pend
ency
inde
pend
ence
grou
p al
loca
tion?
-
Physiotherapy treatment approaches for stroke 401re
crui
ted
and
rand
omiz
ed.
Add
ition
al 1
2pa
tient
s th
ense
lect
ivel
y as
sign
ed t
o ev
en o
ut
diff
eren
ces
in im
port
ant
char
acte
r-is
tics
)W
ade
RC
T (b
lock
edA
dequ
ate
94
1 ye
arYe
sM
ixed
B
arth
el
Riv
erm
ead
Com
men
ts
1992
50ra
ndom
i-(p
robl
em
Inde
x M
obili
ty
from
pee
r za
tion)
.so
lvin
g,
Fren
chay
Ass
essm
ent
revi
ewer
s co
mm
unity
Act
iviti
es
for
the
phys
ioth
er-
Inde
x up
date
dap
y)
Not
tingh
am
vers
ion
led
No
trea
tmen
tE
AD
L sc
ale
to t
he
incl
usio
n of
this
tria
lW
ang
RC
T (s
trat
ified
Ade
quat
e44
Not
sta
ted
Yes
Neu
roph
ysio
logi
cal
(non
e)M
otor
D
ata
from
20
0551
into
pat
ient
s(B
obat
h)
Ass
essm
ent
patie
nts
with
O
rtho
paed
icS
cale
Str
oke
with
sp
astic
ity
Ass
essm
ent
spas
ticity
(B
runn
stro
m
Impa
irmen
t en
tere
d st
age
2 or
3)
Set
unde
r W
ang
and
pat
ient
s20
05, a
nd
with
rel
ativ
eda
ta f
rom
re
cove
ry
patie
nts
with
(B
runn
stro
m
rela
tive
stag
e 4
or 5
))re
cove
ry
ente
red
unde
r W
ang
2005
aW
ellm
onR
CT
Ade
quat
e21
15
0 da
ysN
oM
otor
(n
one)
(non
e)N
o ou
tcom
es19
9752
(met
hod
of
lear
ning
incl
uded
inra
ndom
izat
ion
Con
trol
an
alys
isno
t st
ated
)(n
o tr
eatm
ent)
-
402 A Pollock et al.
no outcomes which were included in the analysis. Afurther three studies (123 patients)34,35,49 did notreport a measure of disability, although they didinclude secondary measures that are analysed in theCochrane review.32 These six studies are not includedin the analysis, results or discussion of this paper.
The remaining 14 trials were analysed within thecomparisons of (1) neurophysiological versus other, (2)motor learning versus other, (3) mixed versus others.
Neurophysiological (Bobath) versus otherComparisons of neurophysiological approaches
with other approaches were reported in seven studies,with one of these studies45 comparing a neurophysio-logical approach to both another approach (motorlearning) and to a control group. The neurophysiolog-ical approach used for all seven studies was describedas Bobath. Time of follow-up was four weeks forLincoln43 and Wang et al.51, six weeks for Pollock47and Richards et al.,48 three months for Langhammerand Stranghelle,42 two weeks after the end of the inter-vention for Mudie et al.45 and at the time of dischargefor Gelber et al.38
The analyses are displayed in the figures and arebriefly described below.
Global dependency (Figure 1)The Barthel Index was reported by six of the
trials.42,43,45 (2 comparisons),47,48 No trials compared theneurophysiological approach with the orthopaedicapproach for global dependency. There were nosignificant differences between neurophysiologicalapproach and motor learning approach (SMD 0.12,95% CI 0.56 to 0.33), mixed approach (SMD 0.13,95% CI 0.87 to 0.61) or no treatment/placebo(SMD 0.71, 95%, CI 0.79 to 0.36), indicating thatthere are no significant differences between neuro-physiological and other approaches for globaldependency.
Functional independence (Figure 2)Five trials reported measures of functional inde-
pendence Gelber et al.38: Functional IndependenceMeasure (FIM); Langhammer and Stranghelle42 andWang et al.51: Motor Assessment Scale (MAS);Lincoln43 and Richards et al.48: Fugl-Meyer motorassessment lower limb score. No trials compared theneurophysiologial approach with no treatment/placebo for functional independence. There were nosignificant differences between neurophysiologicalapproach and orthopaedic approach (SMD 0.02,95% CI 0.55 to 0.59), motor learning approach
Figure 1 Neurophysiological versus other approaches; global dependency scale.
-
Physiotherapy treatment approaches for stroke 403
(SMD 0.08, 95% CI 0.60 to 0.75) or mixedapproach (SMD 0.12, 95% CI 1.16 to 0.91), indi-cating that there are no significant differencesbetween neurophysiological and other approaches forfunctional independence.
Motor learning (Carr and Shepherd) versus otherComparisons of motor learning approaches with
other approaches were reported in seven studies. Themotor learning approach used for all seven studieswas described as, or referenced to, Carr andShepherd. Time of follow-up was two weeks forDean and Shepherd,34 four weeks for Lincoln,43 sixweeks for McClellan and Ada,44 two months for Deanet al.35 and Salbach et al.49 and three months forLanghammer and Stranghelle.42
The analyses are displayed in the figures and arebriefly described below.
Global dependency (Figure 3)Langhammer and Stranghelle,42 Lincoln,43 and
Mudie et al.45 reported the Barthel Index. No trials com-pared the motor learning approach with the orthopaedicapproach or mixed approach for global dependency.There were no significant differences between themotor learning approach and neurophysiological
approach (SMD 0.12, 95% CI 0.33 to 0.56) or notreatment/placebo (SMD 0.24, 95% CI 1.26 to0.78), indicating that there are no significant differ-ences between motor learning and other approaches forglobal dependency.
Functional independence (Figure 4)Langhammer and Stranghelle,42 Lincoln,43 and
McClellan and Ada44 reported the Motor AssessmentScale. No trials compared the motor learningapproach with the orthopaedic approach or mixedapproach for functional independence. There were nosignificant differences between the motor learningapproach and neurophysiological approach (SMD0.08, 95% CI 0.75 to 0.60) or no treatment/placebo(SMD 0.34, 95% CI 1.21 to 0.53), indicating thatthere are no significant differences between motorlearning and other approaches for functionalindependence.
Mixed versus otherEight studies reported comparisons using a mixed
approach. Time of follow-up was six weeks forPollock47 and Richards et al.,48 12 weeks for Duncanet al.,36,37 Green et al.39 and Wade et al.50 and at thetime of discharge from rehabilitation for Stern et al.24
Figure 2 Neurophysiological versus other approaches; functional independence scale.
-
404 A Pollock et al.
and Ozdemir et al.46 There is considerable hetero-geneity in these data. Stern et al.24 and Ozdemiret al.46 are both quasi-randomized trials and sensitivityanalyses were therefore planned to explore the effectof including these studies. As Stern et al.24 was theonly trial comparing a mixed approach with anorthopaedic approach, it is not combined with anyother trials, and sensitivity analysis was therefore notnecessary.
The analyses displayed in the figures and are brieflydescribed below:
Global dependency (Figure 5)Six of the nine studies included measures of global
dependency. The Kenny Institute of RehabilitationScale was used by Stern et al.,24 and the Barthel Indexby Pollock,47 Richards et al.,48 Duncan et al.,36 Greenet al.39 and Wade et al.50 No trials compared the mixed
Figure 3 Motor learning versus other approaches; global dependency scale.
Figure 4 Motor learning versus other approaches; functional independence scale.
-
Physiotherapy treatment approaches for stroke 405
approach with the motor learning approach for globaldependency. There were no significant differencesbetween mixed approach and neurophysiologicalapproach (SMD 0.13, 95% CI 0.61 to 0.87),orthopaedic approach (SMD 0.08, 95% CI 0.42 to0.58) or no treatment/placebo (SMD 0.05, 95%
CI 0.28 to 0.19), indicating that there are no signifi-cant differences between mixed and other approaches.
Functional independence (Figure 6)The Fugl-Meyer motor assessment lower limb
score was used by Richards et al.,48 Duncan et al.36,37;
Figure 5 Mixed versus other approaches; global dependency scale.
Figure 6 Mixed versus other approaches; functional independence scale.
-
406 A Pollock et al.
the Rivermead Mobility Index was used by Greenet al.39 and Wade et al.50 and the FunctionalIndependence Measure was used by Ozdemir et al. 46A mixed approach was significantly more favourablethan a no treatment control (SMD 0.94, 95% CI 0.08to 1.80)(data from Duncan et al.,36,37 Green et al.,39Ozdemir et al.46 and Wade et al.50). If Ozdemiret al.,46 which uses quasi-randomization is removedfrom the analysis the result ceases to shows a signifi-cant effect, although there is a trend towards signifi-cance (SMD 0.28, 95% CI 0.03 to 0.58). No trialscompared the mixed approach with the motor learningor orthopaedic approach for functional independence.There was no significant difference between themixed approach and neurophysiological approach(SMD 0.12, 95% CI 0.91 to 1.16).
Discussion
This review was carried out with the specific aim ofinvestigating the efficacy of different treatmentapproaches, based on a historical perspective. This wasin direct response to a consultation exercise carried outin Scotland which aimed to identify the burning ques-tions of Scottish stroke rehabilitation workers, andwhich identified different treatment approaches to beamongst the most burning questions of physiothera-pists.53 Hence this review was driven by an identifiedclinical question, rather than originating from a scien-tific and logical standpoint. While the results of thisreview may lead to the conclusion that no one physio-therapy treatment approach appears to be more advan-tageous to the promotion of recovery of lower limbfunction or postural control, the difficulties encoun-tered in the methodology of the review highlight theabsence of a scientific rationale for basing physiother-apy interventions on named approaches.
A statistically significant result was found in thecomparison of a mixed approach with a no treatmentcontrol, for the recovery of functional independence.Data from five trials (427 participants) demonstratedthat a mixed approach was significantly morefavourable than no treatment control in the recoveryof functional independence (SMD 0.94, 95% CI 0.08to 1.80). One of the five trials did have a number ofmethodological limitations.46 Ozdemir et al.,46 whichreported a much more significant result, did not userandom allocation to groups and did not have a blinded
outcome assessor: these methodological limitationscould have allowed the introduction of bias into thedata collected. With Ozdemir et al.46 removed fromthe analysis the result ceases to be significant,although there is a trend towards significance (SMD0.28, 95% CI 0.03 to 0.58).
The data analysed in this review provide evidencethat a mixed physiotherapy intervention is signifi-cantly favourable to no treatment intervention in therecovery of functional independence following stroke.This significant effect arguably demonstrates thatany physiotherapy is better than none.
Limitations
Identification of relevant trialsThe identification of all relevant trials was
confounded by a number of factors:
Inconsistent and poorly defined terminology:Electronic searching was difficult as the names andcontent associated with different physiotherapytreatment approaches are poorly documented,often have several derivations, and have variedover time.
Lack of detail within abstract: Lack of informa-tion on study methodology, subjects and interven-tions potentially increases the chance of excludinga relevant trial. However the method of includingall possible trials should have prevented this.
Material published in journals not included inelectronic databases and unpublished material:While substantial effort was made to identifyunpublished material and material in journals notcited on the included databases, relevant trials mayhave not been identified.
Non-English trials awaiting assessment: Twenty-six non-English (23 Chinese) trials are currentlyawaiting translation and formal assessment. Withso many studies awaiting assessment, and thepotential that they may be relevant to the currentlyincluded comparison groups, there is a possibilitythat inclusion of these trials will alter the conclu-sions made in this review.
Quality of included trialsMany of the included trials had methodological
limitations, which may have led to the introduction of
-
Physiotherapy treatment approaches for stroke 407
selection bias. Two key methodological factors whichreduced the quality of many of the included trialswere the method of randomization and blinding:
Randomization: Three of the identified studiesdid not state the method of randomization36,38,52;one study divided patients into matched pairs andthen randomly allocated the pairs35; one study usedquasi-random assignment based on order of entryinto the study46; and the method of randomizationof a fifth24 was identified to be potentially unreli-able. Questions about the quality of randomizationmust challenge the robustness of the study design,and hence the results of this review.
Blinding and contamination: In the majority ofstudies it was unclear whether or not the patientswere blinded to the study group and aims. Thenature of rehabilitation interventions and the ethi-cal requirement to obtain informed consent oftenmakes it difficult, if not impossible, to blindpatients. If the aims and objectives of the studywere apparent to the subjects this could confoundthe study results. The treating therapist(s) was notblinded in any of the trials. This was to be expectedas a treating therapist has to be familiar with theintervention that they are administering. Therapistswho strongly favoured one approach over anothercould introduce performance bias. In several of thestudies the same therapist(s) administered treat-ment to patients in both study groups: this poten-tially introduces considerable contaminationbetween groups. Pollock47 reported some reluc-tance of patients to participate in the treatmentintervention confounding variables such as thesemay be attributed to the beliefs of patients andtherapists, and are examples of effects of lack ofblinding of patients and therapists. Only 13 of the20 included trials stated that they used a blindedassessor. The lack of blinding of assessor potentiallyintroduces considerable bias into the study results.This is particularly important in studies in whichtherapists often have strong beliefs in support of aparticular approach.
Heterogeneity of included trialsIn addition to the limitations of the study method-
ology, the studies included in the review had consid-erable heterogeneity within the interventions, outcomesmeasures and patient samples.
Documentation of interventionsClear, concise documentation of complex physical
interventions is exceptionally difficult to achieve.All of the included studies either gave a briefdescription of the techniques used, or referenced atext in which techniques are described in moredetail. Where possible, authors were contacted andasked to supply any further material that was avail-able (e.g. the more detailed information used by thetreating therapists). However, although there hasbeen an attempt to describe all the administeredinterventions, the available documentation is ofteninsufficient to allow confident and accurate repeti-tion of the applied treatment approach. The prob-lems of documentation are confounded by the factthat the treatments applied are ultimately the deci-sion of a single physiotherapist, based on an indi-vidual assessment of a unique patients movementdisorders.
Furthermore, the common basis of the differentphysiotherapy approaches are that they are holistic.All body parts and movements can be assessed andtreated based on the selected approach; however aphysiotherapist may select to concentrate on thetreatment of one particular body part or movementduring a treatment session. Subsequently the treat-ments given to individual patients by individual ther-apists may vary enormously. This review attemptedto limit this variation slightly by excluding trials thathad only given interventions to the upper limb.Nevertheless, although patients receiving treatmentbased on a particular approach should receivean intervention that conforms to the stated philoso-phy/theory of the approach, it is conceivable thatthere were few similarities between the physicalinterventions given to patients in the same treatmentgroup.
The argument that a physiotherapy approach isbased on an individual assessment of a uniquepatients movement disorders has been used by sometherapists/researchers to perpetuate limited documen-tation and standardization. However recent studieshave demonstrated that clear concise documentationof a treatment intervention does not necessarily meanthe removal of the therapists ability to select a treat-ment based on an individual patients problems. Forexample, Wang et al.,51 within a detailed documenta-tion of the intervention, highlights that the treatmentsare individualised, constantly modified according tosubject response.
-
408 A Pollock et al.
Classification of treatment approachesThe classification of the treatment approaches used inthis review can potentially be criticized for combininga number of different physiotherapy approaches undervery broad classifications (i.e. neurophysiological,motor learning, orthopaedic). Subgroup analysis ofthe individual named approaches within each classifi-cation was planned. However, as all of the neuro-physiological approaches were described asBobath, and all of the motor learning approachesare referenced to Carr and Shepherd, this reviewcannot by default be criticized for combining avariety of approaches under one classification head-ing, as this has not occurred.
The Bobath concept: This review includes eighttrials which stated that they were evaluating aBobath approach to stroke therapy. It is importantto note that there is considerable debate surround-ing the content of physiotherapy interventionsbased on the Bobath concept. This debate largelyarises from the fact that the content of the Bobathapproach has changed over time, there are limitedupdated published descriptions, and that there is avariation in the content of current therapy.1921,5456It is beyond the scope of this review to determinewhether the interventions described as Bobathhad any practical or theoretical differences.
Motor learning and mixed approaches:Reviewers found most difficulty in distinguishingbetween a mixed approach (not a mixture of twodifferent approaches, such as Stern et al.24 mixingorthopaedic and neurophysiological approaches,but an unclassified mix) and a motor learningapproach. The mixed, intensive and focusedapproach investigated by Richards et al.48 and theproblem-solving approach investigated by Greenet al.39 and Wade et al.50 had stated philosophiesvery similar to that of motor learning approaches.However the described techniques, and the sup-porting references, led the reviewers to classifythese interventions as mixed. This highlights akey problem with the classification of the motorlearning approach. While a motor relearning pro-gramme has been described by Carr andShepherd,11,14 these authors primarily advocate anapproach based on related research in relevantareas such as medical science, neuroscience, exer-cise physiology and biomechanics. Such an
approach is arguably one of research-based practice,rather than being based on one specific philosophy.
We suggest that if physiotherapists are to practiseevidence-based stroke rehabilitation their culture, atti-tudes and beliefs will have to shift away from the useof compartmentalized approaches to judging the sci-entific and research base for each individual treatmenttechnique. This review supports this approachbecause it suggests that a mixed approach is moreeffective than no treatment and it fails to demonstrateany superiority for any single approach relating to therecovery of disability following stroke. Future ran-domized controlled trials and systematic reviewsshould concentrate on investigating clearly definedand described techniques and task-specific treatments,and not on compartmentalized approaches.
Clinical messages
No one physiotherapy approach has beenshown to be more advantageous to the promo-tion of recovery of disability following stroke.
Physiotherapy which uses a mix of compo-nents from different approaches may be moreeffective than no treatment or placebo controlin the recovery of functional independencefollowing stroke.
AcknowledgementsThe Stroke Therapy Evaluation Programme is
funded by The Big Lottery Fund, and has previouslybeen funded by Chest Heart and Stroke Scotland andThe Health Foundation.
Competing interestsNone.
References
1 Ashburn A. A review of current physiotherapy in themanagement of stroke. In Harrison MA ed.Physiotherapy in stroke management. ChurchillLivingstone, 1995.
-
Physiotherapy treatment approaches for stroke 409
2 Patridge CJ. Physiotherapy approaches to the treatment of neurological conditions an historicalperspective. In Edwards S ed. Neurological physio-therapy. A problem-solving approach. ChurchillLivingstone, 1996.
3 Davies PM. Steps to follow. A guide to the treatmentof adult hemiplegia. Springer-Verlag, 1985.
4 Bobath B. Adult hemiplegia: evaluation andtreatment, second edition. Butterworth-Heinemann,1990.
5 Brunnstrm S. Movements therapy in hemiplegia.Harper and Row, 1970.
6 Goff B. Appropriate afferent stimulation.Physiotherapy 1969; 55: 917.
7 Knott M, Voss DE. Proprioceptive neuromuscularfacilitation, second edition. Harper and Row, 1968.
8 Voss DE, Ionta MK, Myers BJ. Proprioceptive neuromuscular facilitation patterns and techniques,third edition. Harper & Row, 1985.
9 Turnbull GI. Some learning theory implications inneurological physiotherapy. Physiotherapy 1982; 68:3841.
10 Anderson M, Lough S. A psychological frameworkfor neurorehabilitation. Physiother Pract 1986; 2:7482.
11 Carr JH, Shepherd RB. A motor relearning programme for stroke, first edition. HeinemannMedical, 1982.
12 Carr JH, Shepherd RB. Physiotherapy in disorders ofthe brain. Heinemann Medical, 1980.
13 Carr JH, Shepherd RB. Movement science:foundations for physical therapy in rehabilitation,first edition. Aspen Publishers, 1987.
14 Carr JH, Shepherd RB. A motor relearning programme for stroke, second edition. HeinemannMedical, 1987.
15 Carr JH, Shepherd RB. A motor learning model forstroke rehabilitation. Physiotherapy 1989; 89:37280.
16 Carr JH, Shepherd RB. A motor learning model forrehabilitation of the movement disabled. In Ada L,Canning C eds. Key issues in neurological physiother-apy. Heinemann Medical, 1990: 124.
17 Carr JH, Shepherd RB. Neurological rehabilitation.Optimising motor performance. Butterworth-Heinemann, 1998.
18 Lennon S. The Bobath concept: a critical review ofthe theoretical assumptions that guide physiotherapypractice in stroke rehabilitation. Phys Ther Rev 1996;1: 3545.
19 Nilsson L, Nordholm L. Physical therapy in strokerehabilitation: bases Swedish physiotherapist; choiceof treatments. Physiother Theory Pract 1992; 87:4955.
20 Carr JH, Mungovan SF, Shepherd RB, Dean CM,Nordholm LA. Physiotherapy in stroke rehabilitation;
bases for Australian physiotherapists choice oftreatment. Physiother Theory Pract 1994; 10: 201209.
21 Sackley CM, Lincoln NB. Physiotherapy treatmentfor stroke patients: a survey of current practice.Physiother Theory Pract 1996; 12: 8796.
22 Davidson I, Waters K. Physiotherapists working withstroke patients: a national survey. Physiotherapy2000; 86: 6980.
23 Lennon S, Baxter D, Ashburn A. Physiotherapy basedon the Bobath concept in stroke rehabilitation: a survey within the UK. Disabil Rehabil 2001; 23:25462.
24 Stern PH, McDowell F, Miller JM, Robinson M.Effects of facilitation exercise techniques in strokerehabilitation. Arch Phys Med Rehabil 1970; 51:52631.
25 Logigian MK, Samuels MA, Flaconer J, Zagar R.Clinical exercise trial for stroke patients. Arch PhysMed Rehabil 1983; 64: 36267.
26 Dickstein R, Hocherman S, Pillar T, Shaham R.Stroke rehabilitation: Three exercise therapyapproached. Phys Ther 1986; 66: 123338.
27 Basmajian JV, Gowland CA, Finlayson AJ et al.Stroke treatment: comparison of integrated behavioural physical therapy versus traditional physical therapy programs. Arch Phys Med Rehabil1987; 68: 26772.
28 Lord JP, Hall K. Neuromuscular re-education versustraditional programs for stroke rehabilitation. ArchPhys Med Rehabil 1986; 67: 8891.
29 Brunham S, Snow CJ. The effectiveness of neuro-developmental treatment in adults with neurologicalconditions: a single subject study. Physiother TheoryPract 1992; 8: 21522.
30 Sunderland A, Tinson D, Bradley L, Fletcher D,Langton Hewer R, Wade DT. Enhanced physicaltherapy improves recovery of arm function afterstroke: a randomised controlled trial. J NeurolNeurosurg Psychiatry 1992; 55: 53035.
31 Ernst E. A review of stroke rehabilitation and physio-therapy. Stroke 1990; 21: 108185.
32 Pollock AS, Baer G, Langhorne P, Pomeroy V.Physiotherapy treatment approaches for the recoveryof postural control and lower limb function followingstroke. Cochrane Database Syst Rev Issue 4, 2006.
33 Hatano S. Experience from a multicentre stroke regis-ter: a preliminary report. Bull World Health Organ1976; 54: 54153.
34 Dean CM, Shepherd RB. Task-related trainingimproves performance of seated reaching tasks afterstroke. A randomised controlled trial. Stroke 1997;28: 72228.
35 Dean CM, Richards CL, Malouin F. Task-relatedcircuit training improves performance of locomotortasks in chronic strike: a randomised controlled pilottrial. Arch Phys Med Rehabil 2000; 81: 40917.
-
410 A Pollock et al.
36 Duncan P, Richards L, Wallace D et al. A randomisedcontrolled pilot study of a home-based exercise pro-gram for individuals with mild and moderate stroke.Stroke 1998; 29: 20052060.
37 Duncan P, Studenski S, Richards L et al. Randomisedclinical trial of therapeutic exercise in subacutestroke. Stroke 2003; 34: 217380.
38 Gelber DA, Josefcyz PB, Herrman D, Good DC,Verhulst SJ. Comparison of two therapy approachesin the rehabilitation of the pure motor hemipareticstroke patient. J Neuro Rehabil 1995; 9: 19196.
39 Green J, Forster A, Bogle S, Young J. Physiotherapyfor patients with mobility problems more that 1 yearafter stroke: a randomised controlled trial. Lancet2002; 359: 199203.
40 Hesse J, Jahnke MT, Schaffrin A, Lucke D, Reiter F,Konrad M. Immediate effects of therapeutic facilita-tion on the gait of hemiparetic patients as comparedwith walking and without a cane. ElectroencephalogrClin Neurophysiol 1998; 109: 51522.
41 Howe TE, Taylor I, Finn P, Jones J. Lateral weighttransference exercise following acute stroke: a prelim-inary study of clinical effectiveness. Clin Rehabil2005:19:4553.
42 Langhammer B, Stranghelle JK. Bobath or motorrelearning programme? A comparison of two differentapproaches of physiotherapy in stroke rehabilitation:a randomised controlled study. Clin Rehabil 2000; 14:26169.
43 van Vliet PM, Lincoln NB, Foxall A. Comparison ofBobath based and movement science based treatmentfor stroke: a randomised controlled trial. J NeurolNeurosurg Psychiatry 2005; 76: 503508
44 McClellan R, Ada L. A six-week, resource-efficientmobility program after discharge from rehabilitationimproves standing in people affected by stroke:placebo-controlled, randomised trial. Aust JPhysiother 2004; 50: 16367.
45 Mudie MH, Winzeler-Mercay U, Radwan S, Lee L.Training symmetry of weight distribution after stroke:a randomised controlled pilot study comparing task-related reach, Bobath and feedback trainingapproached. Clin Rehabil 2002; 16: 58292.
46 Ozdemir F, Birtane M, Tabatabaei R, Kokino S,Ekuklu G. Comparing stroke rehabilitation outcomes
between acute inpatient and nonintense home settings. Arch Phys Med Rehabil 2001; 82:137579.
47 Pollock AS. An investigation into independent prac-tice as an addition to physiotherapy intervention forpatient with recently acquired stroke. PhD thesis,Queen Margaret College, Edinburgh, 1998.
48 Richards CL, Malouin F, Bravo G, Dumas F, Wood-Dauphinee S. The role of technology in task-orientedtraining in persons with subacute: a randomised controlled trial. Neurorehabil Neural Repair 2004;18: 199211.
49 Salbach NM, Mayo NE, Wood-Dauphinee S, HanleyJA, Richards CL, Cote R. A task-oriented interventionenchance walking distances and speed in the first yearpost stroke: a randomised controlled trial. ClinRehabil 2004; 18: 50919.
50 Wade DT, Collen FM, Robb GF, Warlow CP.Physiotherapy intervention late after stroke andmobility. BMJ 1992; 304: 60913.
51 Wang Q-R, Gan Z-R, Lu H et al. Effect of early exer-cise therapy on the recovery of motor function inpatient switch cerebral infarction and the changes ofsomatosensory evoked potential. ZhongguoLinchuang Kangfu 2004; 8: 602325.
52 Wellmon R, Newton RA. An examination of changesin gait and standing symmetry associated with thepractice of a weight shifting task. Neurol Rep 1997;21: 5455.
53 Legg L, Pollock A, Langhorne, Sellars C. A multidis-ciplinary research agenda for stroke rehabilitation. BrJ Ther Rehabil 2000; 7: 31924.
54 DeJong G, Horn S, Gassaway J, Slavin M, Dijkers M.Toward a taxonomy of rehabilitation interventions:using an inductive approach to examine the BlackBox of rehabilitation. Arch Phys Med Rehabil 2004;85: 67886.
55 Pomeroy V, Niven D, Barrow S, Faragher E, Tallis R.Unpacking the black box of nursing and therapy prac-tice for the post-stroke shoulder pain: a precursor toevaluation. Clin Rehabil 2001; 15: 6783.
56 Turner PA, Whitfield TA. Physiotherapistsreasons for selection of treatment techniques: a cross-national survey. Physiother Theory Pract 1999;15: 23546.
-
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
top related