REVIEW ARTICLEpublished: 29 January 2013

doi: 10.3389/fendo.2013.00003

Community-based physical activity interventions fortreatment of type 2 diabetes: a systematic review withmeta-analysisRonald C. Plotnikoff 1*, Sarah A. Costigan1, Nandini D. Karunamuni 2 and David R. Lubans1

1 Priority Research Centre in Physical Activity and Nutrition, University of Newcastle, Callaghan, NSW Australia2 Faculty of Physical Education and Recreation, University of Alberta, Edmonton, AB, Canada

Edited by:Catherine Chan, University of Alberta,Canada

Reviewed by:Mini R. Abraham, Overland ParkMedical Specialists, USAJohan H. Jendle, Karlstad CentralHospital, Sweden

*Correspondence:Ronald C. Plotnikoff , PriorityResearch Centre in Physical Activityand Nutrition, Advance TechnologyCentre, Level 3, The University ofNewcastle, University Drive,Callaghan, NSW 2308, Australia.e-mail: ron.plotnikoff@newcastle.edu.au

Evidence suggests engaging in regular physical activity (PA) can have beneficial outcomesfor adults with type 2 diabetes (TD2), including weight loss, reduction of medication usageand improvements in hemoglobin A1c (HbA1c)/fasting glucose. While a number of clinical-based PA interventions exist, community-based approaches are limited. The objective ofthis study is to conduct a systematic review with meta-analysis to assess the effective-ness of community-based PA interventions for the treatment of TD2 in adult populations.A search of peer-reviewed publications from 2002 to June 2012 was conducted acrossseveral electronic databases to identify interventions evaluated in community settings.Twenty-two studies were identified, and 11 studies reporting HbA1c as an outcome mea-sure were pooled in the meta-analysis. Risk of bias assessment was also conducted. Thefindings demonstrate community-based PA interventions can be effective in producingincreases in PA. Meta-analysis revealed a lowering of HbA1c levels by −0.32% [95% CI−0.65, 0.01], which approached statistical significance (p < 0.06). Our findings can guidefuture PA community-based interventions in adult populations diagnosed with TD2.

Keywords: type two diabetes, physical activity, community-based intervention, treatment, HbA1c

INTRODUCTIONThe estimated global prevalence of diabetes in 2010 was 6.4%(equating to ∼285 million adults). Type two diabetes (T2D) hascontributed to the majority of these cases (Shaw et al., 2010).T2D is largely related to weight gain associated with a combina-tion of low physical activity (PA) levels and a consumption of anenergy dense diet (Nolan et al., 2011). Evidence suggests engagingin regular PA can have beneficial outcomes for adults with T2D(Sigal et al., 2007), including improved self-management of T2D,weight loss, increased fitness, reduction of medication usage andimprovements in HbA1c/fasting glucose (Church et al., 2010).

Researchers have investigated a variety of PA interventionstrategies to encourage individuals with T2D to be more active.One such strategy is supervised facility-based exercise training,which has potential to improve glycemic control and other car-diovascular risk factors (Sigal et al., 2007; Balducci et al., 2010;Church et al., 2010). However, these programs are often resource-intensive, only available in metropolitan areas, and their long-termsustainability is indeterminate.

Other strategies to promote PA in T2D adults includeindividual-based approaches such as medication use and behav-ior change. These strategies have become increasingly common,with information on diet, exercise, and medication generally pro-vided by health care professionals (e.g., GP, pharmacist, specialist).However, encouraging individuals with T2D to adopt behaviormodification during short visits to their GP, is challenging. Theseself-management approaches also have modest efficacy in the short

term,and long-term assessments are often very limited (Plotnikoff,2006). Further, these types of interventions can be “out of reach”for individuals with a low-income, low education, lack of access tocare, and cultural and linguistic barriers.

On the other hand, community-based approaches may helpimprove self-management of T2D by addressing barriers encoun-tered in both facility-based approaches and individual-basedapproaches. For example, community-based interventions candeliver culturally appropriate health education which can improveself-care compliance and adherence to self-management prac-tices (Two Feathers et al., 2005; Vachon et al., 2007). Further,community-based interventions can be more cost-effective andpractical, may have better long-term effectiveness, and the poten-tial to reach a large proportion of individuals who are in most needof treatment (Two Feathers et al., 2005; Plotnikoff, 2006; Vachonet al., 2007).

The objective of this paper is to conduct a systematic reviewwith meta-analysis to assess the effectiveness of community-based,PA interventions for the treatment of T2D in adult populations.The meta-analysis focused on interventions that assessed changesin hemoglobin A1c (HbA1c) as an outcome measure. This reviewincludes interventions that employ community-based approaches(e.g., community centers, local facilities, community-based edu-cators), rather than those delivered in workplace or traditionalclinical settings. To the best of our knowledge, this is the first sys-tematic review to examine the efficacy of community-based PAapproaches for treatment of T2D in adults.

www.frontiersin.org January 2013 | Volume 4 | Article 3 | 1

Plotnikoff et al. Community-based physical activity interventions

METHODSSEARCH STRATEGY AND DATA SOURCESStudies published in peer-reviewed journals were identified bytwo authors (SAC & NK) through a structured electronic databasefrom January 2002 until June 2012 in CINAHL, Web of Science,Scopus, and Medline. The following search strings were used: (PAOR exercise) AND (Type two diabetes OR Type 2 Diabetes ORT2D) AND (Intervention OR Program) AND Community. Thesestrings were further limited to subjects 18+ years and English lan-guage. In the first stage of the literature search, titles, and abstractsof identified articles were checked for relevance and additionalarticles known to the authors were assessed for possible inclusion.In the second stage, full-text articles were retrieved and consid-ered for inclusion. In the final stage, the reference lists of retrievedfull-text articles were searched for additional studies.

STUDY SELECTION CRITERIAQuantitative community-based PA interventions for treatment ofT2D were included in this review. Interventional studies were con-sidered for inclusion. Studies were eligible for inclusion if they:(1) included participants 18+ years; (2) employed a PA basedintervention [i.e., studies were classified by proportion of theintervention which was PA based (i.e., ≥50%)]; (3) employedcommunity-based approaches (e.g., community centers, localfacilities, community-based educators; (4) implemented strategiesfor treating existing cases of T2D (e.g., increased PA, weight loss,reduction of medication usage, improvements in HbA1c); and (5)were quantitative studies.

Studies not included in this review consist of publicationswhich: (1) examined children or adolescents (aged < 18 years); (2)were non-interventional; (3) had an intervention that was not atleast 50% focused on PA; (4) did not include strategy for < 50%of the total intervention; (5) were conducted in traditional clinicalsettings; (6) employed preventative measures to adults at risk ofT2D (i.e., cases of pre-diabetes); (7) used qualitative studies; and(8) were conference abstracts, dissertations, theses, and articlespublished in non-peer-reviewed journals.

DATA EXTRACTIONInitially, articles were assessed for eligibility by a single reviewer(SAC) based on the study title. After this initial cull, study abstractswere assessed by two authors (SAC & NK) independently in anunblinded standardized manner. Findings were compared and dif-ferences were resolved by consensus or by a third author (RCP).Specific study characteristics were identified and extracted by twoauthors (SAC & NK). These characteristics included the countryof origin, study-design, size/source of study population, details ofthe PA intervention and community components of the study (seeTable 1).

SYNTHESIS OF RESULTSMeta-analysis was conducted for studies that reported their effecton HbA1c using RevMan version 5.1 [The Nordic Cochrane Cen-tre, The Cochrane Collaboration, Copenhagen, Denmark; ReviewManager (RevMan), 2011]. As recommended by the CochraneCollaboration, post-test means (i.e., Less et al., 2009; Plotnikoffet al., 2010, 2012; Piette et al., 2011) or change scores (i.e.,

Goldhaber-Fiebert et al., 2003; Dunstan et al., 2006; Engel andLindner, 2006; Brooks et al., 2007; Davies et al., 2008; Skoro-Kondza et al., 2009; Plotnikoff et al., 2011) and their SD were usedin the analysis. One study compared multiple treatment groupswith a single control group (n= 1), to avoid double counting, thesample size of the control group was divided by two. All data wereconsidered continuous and therefore the mean difference (MD)with 95% confidence intervals was used to determine effect mea-sures. Statistical heterogeneity was examined via Chi-squared andthe I2-Index tests. A guide to the interpretation of heterogeneitybased on the I2-Index is as follows: 0–40% might not be impor-tant; 30–60% may represent moderate heterogeneity; 50–90% mayrepresent substantial heterogeneity; and 75–100% considerableheterogeneity (Deeks et al., 2008).

RISK OF BIASRisk of bias was assessed using a nine-item checklist tool adaptedfrom the Consolidated Standards of Reporting Trials (CONSORT)statement and previous reviews (van Sluijs et al., 2007; Younget al., 2012). The following items were assessed: (1) baselineresults reported separately for each group; (2) randomizationclearly described and adequately completed; (3) dropout≤ 20%for≤6 months follow-up and≤30% for >6 months follow-up; (4)assessor blinding; (5) intention-to-treat analysis; (6) confoundersaccounted for in analyses; (7) summary results presented andestimated effect sizes and precision estimates; (8) power calcula-tion reported and study adequately powered; and (9) an objectivemeasure of PA was used.

Each item was scored as“present”(X),“absent”(×), or“unclearor inadequately described” (?). Depending on the study-design,some items were not applicable (n/a). Unweighted sum totals werecalculated for each study using a predefined scoring system (X= 1|×= 0 | ?= 0 | n/a= 0). Low risk of bias studies were regarded asthose with a 8–9, a moderate risk of bias presented scores of 4–7,high risk of bias scored 0–3.

RESULTSDESCRIPTION OF STUDIESFigure 1 describes the progress through the stages of study selec-tion. The electronic database search strategy provided 1015 refer-ences; 38 additional records were identified through other sources.About 1053 records were screened based on titles and abstracts,and of these 992 records were excluded. The remaining 61 recordswere assessed as full-text articles. Of these 61 studies, 39 did notmeet eligibility criteria. The remaining 22 studies were includedin our review (see Figure 1).

STUDY CHARACTERISTICSTable 1 reports selected characteristics of all eligible studies. Oursearch identified 22 studies that examined community-based PAinterventions for treatment of T2D, of which a PA component of≥50% of the overall intervention was employed.

Most studies (16/22, 73%) employed a RCT design. The activeintervention periods ranged from 4 weeks (Dutton et al., 2008)to 24 months (Brandon et al., 2003). Study populations rangedfrom 19 (Martyn-Nemeth et al., 2010) to 652 (Davies et al., 2008).Most of the studies were conducted in the United States (12/20),

Frontiers in Endocrinology | Diabetes January 2013 | Volume 4 | Article 3 | 2

Plotnikoff et al. Community-based physical activity interventions

Tab

le1

|Stu

dy

char

acte

rist

ics.

Au

tho

r

(dat

e)

Stu

dy

des

ign

NN

per

gro

up

Rec

ruit

men

tIn

terv

enti

on

len

gth

Co

mm

un

ity

bas

ed

inte

rven

tio

n

com

po

nen

t

PAas

pec

tP

rim

ary

Ou

tco

me(

s)

Res

ult

s

Plo

tnik

off

etal

.

(201

1),

Can

ada

RC

T96

N(C

on)=

49 N(In

t)=

47

Dia

bete

scl

inic

s16

wee

ksR

ecru

itmen

t

Use

of

com

mun

ity-

base

dfa

cilit

ies

Str

uctu

red

resi

stan

t

trai

ning

prog

ram

HbA

1c

PA(M

ETm

ins)

Fitn

ess

(mC

AFT

)

HbA

1c[b

asel

ine

mea

n(S

D)]

Con

trol=

7.8%

(2.0

);61

.75

mm

ol/m

ol

Inte

rven

tion=

7.3%

(1.3

);56

.28

mm

ol/m

ol

HbA

1c[1

2-m

onth

sm

ean

(SD

)]

Con

trol=

7.2%

(1.6

);55

.19

mm

ol/m

ol

Inte

rven

tion=

7.2%

(1.4

);55

.19

mm

ol/m

ol

HbA

1cm

ean

chan

geat

12m

onth

s

Con

trol

grou

p=−

0.5;

95%

CI:−

0.9

to−

0.2)

Inte

rven

tion

grou

p=−

0.4;

95%

CI:−

0.7

to0.

0

PA(M

ETm

in)

Mea

nch

ange

at12

mon

ths

Con

:−33

.9(C

I:−

213.

6–14

5.8)

Int:

654.

2

(CI:4

66.9−

841.

6)**

Fitn

ess

(mC

AFT

)

Mea

nch

ange

at12

mon

ths

Con

:−1.

9(−

16.3

–12.

5)

Int:

28.2

(14.

2–42

.2)*

*

Pie

tte

etal

.

(201

1),

USA

RC

T29

1N

(Con

)=

146

N(In

t)=

145

Com

mun

ity-b

ased

non-

profi

the

alth

care

syst

em;A

dditi

onal

patie

nts

wer

e

self-

refe

rred

base

d

onad

vert

isem

ents

and

new

slet

ters

12m

onth

sR

ecru

itmen

t

Prim

ary

care

prov

ider

s

iden

tified

for

each

part

icip

ant

Pedo

met

er-b

ased

wal

king

prog

ram

HbA

1cH

bA1c

[bas

elin

em

ean

(SD

)]C

ontr

ol=

7.7%

(1.7

);60

.66

mm

ol/m

ol

Inte

rven

tion=

7.5%

(1.7

);58

.47

mm

ol/m

ol

HbA

1c[1

2-m

onth

sm

ean

(SD

)]

Con

trol=

7.7%

(1.7

);60

.66

mm

ol/m

ol

Inte

rven

tion=

7.7%

(1.8

);60

.66

mm

ol/m

ol

HbA

1cB

etw

een-

grou

pD

iffer

ence

(95%

CI)=

0.07

(−0.

26to

0.40

)

Kru

se

etal

.

(201

0),

USA

RC

T79

N(C

on)=

38 N(In

t)=

41

Prog

ram

open

to

anyo

nem

eetin

gth

e

crite

ria

12m

onth

sR

ecru

itmen

t

Com

mun

ity

sett

ing

Indi

vidu

alse

ssio

nsw

ith

phys

ical

ther

apis

t

focu

sing

onle

gst

reng

th

and

prom

otin

gba

lanc

e

Str

engt

han

d

bala

nce

BM

I

Falls

No

sign

ifica

ntdi

ffere

nces

betw

een

cont

rola

nd

inte

rven

tion

grou

psfo

rth

epr

imar

you

tcom

es.

Wee

kly

1-h

sess

ions

to

impr

ove

leg

stre

ngth

and

bala

nce

Add

ing

100

step

sto

thei

rda

ilyac

tiviti

es

ever

y2

wee

ks

(Con

tinue

d)

www.frontiersin.org January 2013 | Volume 4 | Article 3 | 3

Plotnikoff et al. Community-based physical activity interventions

Tab

le1

|Co

nti

nu

ed

Au

tho

r

(dat

e)

Stu

dy

des

ign

NN

per

gro

up

Rec

ruit

men

tIn

terv

enti

on

len

gth

Co

mm

un

ity

bas

ed

inte

rven

tio

n

com

po

nen

t

PAas

pec

tP

rim

ary

Ou

tco

me(

s)

Res

ult

s

Less

etal

.

(200

9),

Jam

aica

Pros

pect

ive

coho

rt

stud

y

293

N(C

on)=

135

N(In

t)=

158

Hea

lthce

nter

s6

mon

ths

Rec

ruitm

ent

Com

mun

ity

sett

ing

Phy

sica

lact

ivity

log,

they

trac

ked

thei

r

succ

ess

inac

hiev

ing

the

goal

of30

-min

phys

ical

activ

ityea

chda

y

HbA

1cH

bA1c

(Bas

elin

eM

ean)

Con

trol=

8.03

%;6

4.26

mm

ol/m

ol

Inte

rven

tion=

7.95

%;6

3.39

mm

ol/m

ol

HbA

1c(P

ost

Mea

n)

Con

trol=

8.62

%;7

0.71

mm

ol/m

ol

Inte

rven

tion=

7.32

%**

;56.

50m

mol

/mol

Mar

tyn-

Nem

eth

etal

.

(201

0),

USA

Pret

est/

post

-

test

desi

gn

16O

ne

grou

p

N=

16

Com

mun

ity-b

ased

clin

icth

atpr

ovid

es

care

for

low

-inco

me,

inde

pend

ent-

livin

g

com

mun

ity

resi

dent

s.

12w

eeks

Inpu

tpr

ovid

ed

from

:

com

mun

ity-

base

d

nurs

es

Reg

ular

low

-impa

ct

exer

cise

rout

ine

led

by

nurs

es

Wee

kly

exer

cise

sess

ions

wer

ele

dfo

r

12w

eeks

60-m

in,l

ow-im

pact

exer

cise

rout

ine

HbA

1c

Lipi

dpa

nel

BM

I

No

sign

ifica

ntdi

ffere

nces

betw

een

cont

rola

nd

inte

rven

tion

grou

psfo

rth

epr

imar

you

tcom

es.

diab

etic

nurs

e

educ

ator

med

ical

staf

f

exer

cise

phys

iolo

gist

Mat

hieu

etal

.

(200

8),

Can

ada

RC

T58

N(C

on)=

25 N(In

t)=

33

Kin

esio

logi

sts

in

com

mun

ity-b

ased

faci

litie

sw

ere

resp

onsi

ble

for

recr

uitin

g

part

icip

ants

10w

eek

Rec

ruitm

ent

Gro

uptr

aini

ng

sess

ions

Dia

betA

ctio

npr

ogra

m

intr

oduc

edpa

rtic

ipan

ts

toca

rdio

vasc

ular

,

resi

stan

ce,b

alan

ce,a

nd

flexi

bilit

yex

erci

ses

PAN

osi

gnifi

cant

diffe

renc

esbe

twee

nco

ntro

land

inte

rven

tion

grou

psfo

rth

epr

imar

you

tcom

es

10gr

oup

sess

ions

cons

istin

gof

a60

-min

PApe

riod

ofae

robi

c

exer

cise

sat

light

to

mod

erat

e-in

tens

ity,

resi

stan

ceex

erci

ses,

and

bala

nce

and

flexi

bilit

yex

erci

ses

(Con

tinue

d)

Frontiers in Endocrinology | Diabetes January 2013 | Volume 4 | Article 3 | 4

Plotnikoff et al. Community-based physical activity interventions

Tab

le1

|Co

nti

nu

ed

Au

tho

r

(dat

e)

Stu

dy

des

ign

NN

per

gro

up

Rec

ruit

men

tIn

terv

enti

on

len

gth

Co

mm

un

ity

bas

ed

inte

rven

tio

n

com

po

nen

t

PAas

pec

tP

rim

ary

Ou

tco

me(

s)

Res

ult

s

Tayl

or

etal

.

(200

9),

USA

RC

T24

N(C

on)=

13 N(In

t)=

11

Loca

lcom

mun

ity2

mon

ths

Rec

ruitm

ent

Use

of

com

mun

ity

fitne

ssce

nter

faci

litie

s

Phy

sica

lthe

rapi

st–

dire

cted

exer

cise

coun

selin

g

Fitn

ess

cent

er–b

ased

exer

cise

trai

ning

Mus

cula

r

stre

ngth

Exe

rcis

eca

paci

ty

No

sign

ifica

ntdi

ffere

nces

betw

een

cont

rola

nd

inte

rven

tion

grou

psfo

rth

epr

imar

you

tcom

es

4se

tsof

upto

8

repe

titio

nsof

each

exer

cise

usin

g80

%of

the

1-re

petit

ion

max

imum

(1-R

M)o

n2

non-

cons

ecut

ive

days

per

wee

kfo

r2

mon

ths

Sko

ro-

Kond

za

etal

.

(200

9),

UK

RC

T59

N(C

on)=

30 N(In

t)=

29

GP

’s12

wee

ksC

omm

unity

-

base

d

sess

ions

Reg

ular

yoga

clas

ses

24×

90-m

inyo

ga

clas

ses

over

12w

eeks

HbA

1cH

bA1c

Con

trol

grou

p:

Bas

elin

em

ean=

7.03

%;5

3.33

mm

ol/m

ol

Mea

nat

6m

onth

s=

7.30

%;5

6.28

mm

ol/m

ol(m

ean

chan

ge−

0.28

;95%

CI−

0.10

–0.6

6)

Inte

rven

tion

grou

p:

Bas

elin

em

ean=

7.06

%;5

3.66

mm

ol/m

ol

Mea

nat

6m

onth

s=

7.01

%;5

3.11

mm

ol/m

ol(m

ean

chan

ge−

0.05

;95%

CI−

0.26

–0.1

6).

Dut

ton

etal

.

(200

8),

USA

RC

T85

N(C

on)=

39 N(In

t)=

46

Com

mun

itydi

abet

ic

clin

ic

4w

eeks

Rec

ruitm

ent

Sta

ge-t

arge

ted

book

let

addr

essi

ng:

PAW

eekl

ym

inut

esof

PA

Con

trol

grou

p:

Bas

elin

em

ean=

107.

3(1

02.2

)

Mea

nch

ange

from

base

line

to

post

-inte

rven

tion=−

1.9,

95%

CI=−

37.0

to33

.2

Inte

rven

tion

grou

p:

Bas

elin

em

ean=

148.

5(1

21.3

)

Mea

nch

ange

from

base

line

to

post

-inte

rven

tion=

21.8

,95%

CI=−

16.3

to59

.9

Mot

ivat

ion

Sel

f-ef

ficac

y

Goa

l-set

ting

Soc

ials

uppo

rt

Prob

lem

-sol

ving

Two-

page

lett

erta

ilore

d

toth

eir:

Sta

geof

chan

ge

Sel

f-ef

ficac

y

Pros

/con

sfo

rac

tivity

Cur

rent

activ

ityle

vels

/

pref

eren

ces

Act

ivity

goal

s

(Con

tinue

d)

www.frontiersin.org January 2013 | Volume 4 | Article 3 | 5

Plotnikoff et al. Community-based physical activity interventions

Tab

le1

|Co

nti

nu

ed

Au

tho

r

(dat

e)

Stu

dy

des

ign

NN

per

gro

up

Rec

ruit

men

tIn

terv

enti

on

len

gth

Co

mm

un

ity

bas

ed

inte

rven

tio

n

com

po

nen

t

PAas

pec

tP

rim

ary

Ou

tco

me(

s)

Res

ult

s

Klu

g

etal

.

(200

8),

USA

Non

-

rand

omiz

ed

one

grou

p

befo

rean

d

afte

rde

sign

243

One

grou

pA

t

4m

onth

s

N=

147

Ann

ounc

emen

tsan

d

pres

enta

tions

to

inte

rnal

and

exte

rnal

com

mun

itygr

oups

,

post

edfli

ers,

plac

ed

bulle

tins

in

new

slet

ters

,

cont

acte

del

igib

le

olde

rad

ults

dire

ctly

,

depe

nded

on“w

ord

ofm

outh

,”w

orke

d

with

loca

lmed

ia,

and

part

nere

dw

ith

othe

rag

enci

es

6m

onth

sR

ecru

itmen

t

Set

ting

Peer

lead

er

Wee

kly

supp

ort

sess

ions

for

PAan

ddi

et

Sel

f-ef

ficac

y

Sel

f-ra

ted

heal

th

Sel

f-ef

ficac

yan

dse

lf-ra

ted

heal

thsc

ores

sign

ifica

ntly

impr

oved

Spe

er

etal

.

(200

8),

USA

Con

veni

ence

sam

ple

260

One

grou

p

N=

144

Sen

iors

cent

ers

4m

onth

sR

ecru

itmen

t

Exp

erts

from

the

loca

l

univ

ersi

ty

Phy

sica

lact

ivity

inco

rpor

ated

into

sess

ions

HbA

1cH

bA1c

Bas

elin

e7.

00%

(SD=

1.39

);53

.01

mm

ol/m

ol

Post

-inte

rven

tion

6.7%

(SD=

1.17

)*;

50.3

8m

mol

/mol

40–6

0m

inw

eekl

y

sess

ions

Bro

oks

etal

.

(200

7),

USA

RC

T62

N(C

on)=

31 N(In

t)=

31

Not

repo

rted

16w

eeks

Exe

rcis

e

trai

ning

cond

ucte

din

com

mun

ity

sett

ing

Reg

ular

exer

cise

and

stre

ngth

trai

ning

HbA

1cH

bA1c

[bas

elin

em

ean

(SE

)]

Bod

yco

mpo

sitio

nC

ontr

ol=

7.8%

(1.6

);61

.75

mm

ol/m

ol

Mus

cle

stre

ngth

Inte

rven

tion=

8.7%

(1.8

);71

.58

mm

ol/m

ol

Mus

cle

qual

ityH

bA1c

(pos

t)

3x/w

eek

for

16w

eeks

for

exer

cise

trai

ning

.

Con

trol=

8.3%

(1.3

);67

.21

mm

ol/m

ol

Inte

rven

tion=

7.6%

(1.5

)**;

59.5

6m

mol

/mol

Con

trol

grou

p:

Mea

nch

ange

from

base

line

topo

st-in

terv

entio

n

[mea

n(S

E)]=

0.4

(0.3

)

Ses

sion

sin

clud

ed

35-m

inst

reng

th

trai

ning

3se

tsof

8

repe

titio

nson

each

mac

hine

(Con

tinue

d)

Frontiers in Endocrinology | Diabetes January 2013 | Volume 4 | Article 3 | 6

Plotnikoff et al. Community-based physical activity interventions

Tab

le1

|Co

nti

nu

ed

Au

tho

r

(dat

e)

Stu

dy

des

ign

NN

per

gro

up

Rec

ruit

men

tIn

terv

enti

on

len

gth

Co

mm

un

ity

bas

ed

inte

rven

tio

n

com

po

nen

t

PAas

pec

tP

rim

ary

Ou

tco

me(

s)

Res

ult

s

Trai

ning

inte

nsity

durin

gw

ks1–

8

wer

e60

–80%

of

base

line

1-re

petit

ion

max

imum

(1-R

M),

and

durin

gw

ks

10–1

4w

ere

70–8

0%of

mid

-stu

dy1-

RM

Inte

rven

tion

grou

p:

Mea

nch

ange

from

base

line

topo

st-in

terv

entio

n

[mea

n(S

E)]=−

1.0

(0.2

)

Sig

nific

ant

chan

gesc

ore

diffe

renc

esfo

rth

e

inte

rven

tion

grou

pvs

.con

trol

son

uppe

rbo

dy

mus

cle

stre

ngth

,low

erbo

dym

uscl

est

reng

th,

Who

le-b

ody

lean

tissu

em

ass,

Mus

cle

qual

ity,T

ype

1an

dTy

pe2

mus

cle

fire

area

Dun

stan

etal

.

(200

6),

Aus

tral

ia

RC

T57

N(h

ome

trai

n-

ing)=

26

N(c

ente

r

trai

n-

ing)=

27

Rec

ruite

dfr

omth

e

clin

ics

ofth

e

Inte

rnat

iona

l

Dia

bete

sIn

stitu

te

and

bya

loca

lmed

ia

cam

paig

n.

12m

onth

sC

omm

unity

fitne

ssce

nter

faci

litie

sus

ed

Res

ista

nce

trai

ning

prog

ram

2da

ys/w

eek

HbA

1cH

bA1c

Hom

egr

oup:

Bas

elin

em

ean=

7.5%

(0.5

);58

.47

mm

ol/m

ol

Cen

ter

grou

p:

Bas

elin

em

ean=

7.8%

(0.9

);61

.75

mm

ol/m

ol

Hom

egr

oup:

Mea

nch

ange

from

base

line:

M(S

D):−

0.1(

1.1)

Cen

ter

grou

p:

Mea

nch

ange

from

base

line:

M(S

D):−

0.4(

1.0)

**

Low

-inte

nsity

stat

iona

ry

cycl

ing

plus

stre

tchi

ng

exer

cise

san

d45

min

of

high

inte

nsity

resi

stan

cetr

aini

ng

Goa

lwas

toac

hiev

e

betw

een

75an

d85

%of

the

curr

ent

1-R

M.T

hree

sets

ofei

ght

repe

titio

ns

wer

epe

rfor

med

for

all

exer

cise

s

Boy

d

etal

.

(200

6),

USA

Non

-

rand

omiz

ed

one

grou

p

befo

rean

d

afte

rde

sign

48O

ne

grou

p

N=

35

Rec

ruite

dan

d

refe

rred

by

phys

icia

ns,

phys

icia

ns

assi

stan

ts,a

dvan

ce

nurs

epr

actit

ione

rs

12m

onth

sPa

rtne

rshi

p

betw

een

a

com

mun

ity

heal

thce

nter

s

toim

prov

e

acce

ssto

exer

cise

for

low

-inco

me

patie

nts

with

type

2

diab

etes

3m

onth

YM

CA

mem

bers

hip

exer

cise

clas

ses

twic

e

per

wee

kla

stin

gup

to

12m

onth

s

HbA

1cH

bA1c

Blo

odpr

essu

reB

asel

ine:

8.6%

;70.

49m

mol

/mol

Post

-inte

rven

tion:

6.9%

*;51

.91

mm

ol/m

ol

LDL

Blo

odpr

essu

re

Bas

elin

esy

stol

ic:1

30m

mH

g

Post

-inte

rven

tion:

126

mm

Hg

Bas

elin

edi

asto

lic:7

9m

mH

g

Post

-inte

rven

tion

dias

tolic

:75

mm

Hg*

LDL

Bas

elin

e:11

6m

g/dL

Post

-inte

rven

tion:

99m

g/dL

*

(Con

tinue

d)

www.frontiersin.org January 2013 | Volume 4 | Article 3 | 7

Plotnikoff et al. Community-based physical activity interventions

Tab

le1

|Co

nti

nu

ed

Au

tho

r

(dat

e)

Stu

dy

des

ign

NN

per

gro

up

Rec

ruit

men

tIn

terv

enti

on

len

gth

Co

mm

un

ity

bas

ed

inte

rven

tio

n

com

po

nen

t

PAas

pec

tP

rim

ary

Ou

tco

me(

s)

Res

ult

s

Wei

ght

HD

L

Bas

elin

e:41

mg/

dL

Post

-inte

rven

tion:

44m

g/dL

Wei

ght

Bas

elin

e:21

8lb

Post

-inte

rven

tion:

206

lb

Eng

el

and

Lind

ner

(200

6),

Aus

tral

ia

RC

T57

N(c

oach

ing

only

)=28

Loca

lmed

ia

cam

paig

n

6m

onth

sR

ecru

itmen

t

Gro

up

coac

hing

sess

ions

Gro

upco

achi

ng

sess

ions

HbA

1c

Blo

odpr

essu

re

BM

I

PA Fitn

ess

(shu

ttle

run)

HbA

1c

Coa

chin

gon

lygr

oup:

N(P

edo-

met

er)=

22

Pedo

met

erus

e

valu

esof

betw

een

6000

and

8500

step

s

per

day

(hea

lthy

olde

r

adul

ts)a

ndbe

twee

n

3500

and

5500

step

s

per

day

(old

erad

ults

)

Bas

elin

em

ean=

7.70

%(1

.39)

;60.

66m

mol

/mol

Mea

nch

ange

from

base

line

to

post

-inte

rven

tion=−

0.2(

1.1)

Pedo

met

ergr

oup:

Bas

elin

em

ean=

7.14

%(0

.98)

;54.

54m

mol

/mol

Mea

nch

ange

from

base

line

to

post

-inte

rven

tion=

0.02

(1.0

)

Syst

olic

bloo

dpr

essu

re

Pedo

met

er

Mea

nch

ange

from

base

line=

2.4

(18.

3)

Coa

chin

g=

–2.9

(14)

Dia

stol

icbl

ood

pres

sure

Mea

nch

ange

from

base

line

Pedo

met

er=

1.6

(9.9

)

Coa

chin

g=

–4.1

(9.9

)

BM

I

Mea

nch

ange

from

base

line

Pedo

met

er=

–0.7

(2.3

)

Coa

chin

g=

–0.7

(1.5

)

Fitn

ess

Mea

nch

ange

from

base

line

Pedo

met

er=

65(1

02)

Coa

chin

g=

49(7

2)

(Con

tinue

d)

Frontiers in Endocrinology | Diabetes January 2013 | Volume 4 | Article 3 | 8

Plotnikoff et al. Community-based physical activity interventions

Tab

le1

|Co

nti

nu

ed

Au

tho

r

(dat

e)

Stu

dy

des

ign

NN

per

gro

up

Rec

ruit

men

tIn

terv

enti

on

len

gth

Co

mm

un

ity

bas

ed

inte

rven

tio

n

com

po

nen

t

PAas

pec

tP

rim

ary

Ou

tco

me(

s)

Res

ult

s

Two

Feat

hers

etal

.

(200

5),

USA

Non

-

rand

omiz

ed

one

grou

p

befo

rean

d

afte

rde

sign

91O

ne

grou

p

N=

91

Hea

lthca

resy

stem

s

inde

troi

t

5m

onth

sLo

cally

base

d

com

mun

ity

resi

dent

mee

tings

Som

em

eetin

gs

focu

sed

onin

crea

sing

PAto

impr

ove

patie

nts

self-

man

agem

ent

of

diab

etes

HbA

1cH

bA1c

Bas

elin

e:8.

4%(S

D=

2.3)

;68.

31m

mol

/mol

Post

-inte

rven

tion:

7.6%

(SD=

1.9)

*;

59.5

6m

mol

/mol

Bra

ndon

etal

.

(200

3),

USA

RC

T31

N(C

on)=

15 N(In

t)=

16

Vete

rans

affa

irs

diab

etic

clin

ic,a

com

pute

rized

rese

arch

cent

er

data

base

,loc

al

diab

etic

clin

ics,

and

seni

orce

nter

s,an

d

byw

ord

ofm

outh

24m

onth

sR

ecru

itmen

tR

egul

artr

aini

ngfo

r

24m

onth

s

Mob

ility

Str

engt

h-hi

pfle

xors

Com

mun

ityPA

faci

litie

s

3se

ssio

nspe

rw

eek

Str

engt

hC

ontr

ol:0

.38(

0.07

)S

tren

gth

trai

ning

was

com

plet

edon

an

11-s

tatio

nN

autil

us

mac

hine

.The

subj

ects

com

plet

ed3

sets

of

8–12

repe

titio

nspe

r

exer

cise

.The

inte

nsity

was

50,6

0,an

d70

%

for

sets

1,2,

and

3,

resp

ectiv

ely.

Eac

h1-

h

sess

ion

cons

iste

dof

50m

inof

stre

ngth

-tra

inin

g

exer

cise

san

d10

min

of

war

m-u

p,fle

xibi

lity,

and

cool

-dow

nex

erci

ses

Inte

rven

tion:

0.44

(0.1

2)

Mob

ility

-tim

edup

and

go:

Con

trol

:8.3

(1.0

)

Inte

rven

tion:

7.6(

1.8)

Ther

ew

asa

grou

p×

time

effe

ctas

the

Inte

rven

tion

grou

pin

crea

sed

31.4

%(p

<.0

01)i

nst

reng

thfo

ral

l

mus

cle

grou

psaf

ter

the

first

6m

onth

s,an

dth

e

stre

ngth

gain

sw

ere

reta

ined

at24

mon

ths.

Ther

e

was

also

agr

oup×

time

effe

ctfo

rm

obili

tyas

perf

orm

ance

incr

ease

d8.

6an

d9.

8%(p

<03

2an

d

p<

0.03

1)fo

rth

efir

st6

and

12m

onth

s,

resp

ectiv

ely,

but

decr

ease

dto

4.6%

abov

e

base

line

atth

een

dof

the

inte

rven

tion.

Gol

dhab

er-

Fieb

ert

etal

.

(200

3),

Cos

ta

Ric

a

RC

T75

N(C

on)=

28 N(In

t)=

33

Com

mun

ityG

P’s

12w

eeks

Inte

rven

tion

clas

ses

cond

ucte

dby

:

Nut

ritio

nist

s

Loca

l

volu

ntee

rs

/lead

ers

60m

inw

alki

nggr

oup

sess

ions

thre

etim

es

per

wee

kfo

r12

wee

ks

HbA

1c

BM

I

Wei

ght

Fast

ing

plas

ma

gluc

ose

Ser

umlip

ids

Blo

odpr

essu

re

HbA

1c

Con

trol

grou

p:

Bas

elin

em

ean=

8.6%

(3.9

);70

.49

mm

ol/m

ol

Mea

nch

ange

from

base

line:−

0.4(

2.3)

Inte

rven

tion

grou

p:

Bas

elin

em

ean=

8.6%

(3.7

);70

.49

mm

ol/m

ol

Mea

nch

ange

from

base

line:−

1.8(

2.3)

**

Sig

nific

ant

decr

ease

sin

wei

ght,

BM

I,an

dfa

stin

g

plas

ma

bloo

dgl

ucos

efo

rin

terv

entio

nvs

.con

trol

grou

p

(Con

tinue

d)

www.frontiersin.org January 2013 | Volume 4 | Article 3 | 9

Plotnikoff et al. Community-based physical activity interventions

Tab

le1

|Co

nti

nu

ed

Au

tho

r

(dat

e)

Stu

dy

des

ign

NN

per

gro

up

Rec

ruit

men

tIn

terv

enti

on

len

gth

Co

mm

un

ity

bas

ed

inte

rven

tio

n

com

po

nen

t

PAas

pec

tP

rim

ary

Ou

tco

me(

s)

Res

ult

s

Keys

erlin

g

etal

.

(200

2),

USA

RC

T20

0N

(clin

ic

and

com

-

mun

ity)

b=

54

Com

mun

ityG

P’s

6m

onth

sP

hone

calls

to

part

icip

ants

and

grou

p

sess

ions

PAco

mpo

nent

deve

lope

dto

incr

ease

mod

erat

e-in

tens

ityPA

toa

cum

ulat

ive

tota

lof

30m

ina

day

PAPA

(kca

l/day

)

Clin

ican

dco

mm

unity

:364

(SE

:168

.7)

Clin

icon

ly:3

22(S

E:2

07.1

)

N(c

linic

only

)=59

Min

imal

inte

rven

tion:

297

(SE

:167

.2)

Cha

nge

scor

esov

er24

mon

ths

Clin

ican

dco

mm

unity

vs.c

ontr

ol:>

44.1

(SE

:15.

8)

[CI:1

3.1

to75

.1]*

*N

(min

imal

inte

rven

-

tion)=

58

Clin

icon

lyvs

Con

trol

:>33

.1(S

E:1

5.1)

[CI:3

.3to

62.8

]**

Plo

tnik

off

etal

.

(201

2),

Can

ada

RC

T28

7N

(Con

)=

84 N(p

rint

mat

eri-

als)=

71

N(t

elep

hone

coun

sel-

ing)=

73

Gen

eral

adve

rtis

ing

stra

tegi

es

12m

onth

sR

ecru

itmen

t

Pho

neca

llsto

part

icip

ants

PAgu

idel

ines

Pers

onal

ized

prin

t

mat

eria

ls

Tele

phon

eco

unse

ling

Use

ofa

pedo

met

er

HbA

1c

PA

HbA

1c[B

asel

ine

mea

n(S

E)v

alue

s]

Con

trol

grou

p=

7.08

%(0

.07)

;53.

88m

mol

/mol

Prin

tm

ater

ials=

7.08

%(0

.07)

;53.

88m

mol

/mol

Tele

phon

eco

unse

ling=

7.11

%(0

.07)

;

54.2

1m

mol

/mol

[12-

mon

ths

mea

n(S

E)v

alue

s]

Con

trol=

7.07

%(0

.07)

;53.

77m

mol

/mol

Prin

tm

ater

ials=

7.00

%(0

.08)

;53.

01m

mol

/mol

Tele

phon

eco

unse

ling=

7.28

%(0

.08)

;

56.0

7m

mol

/mol

Adj

uste

ddi

ffere

nce

betw

een-

grou

ps

Prin

tm

ater

ials

grou

p–

Con

trol

grou

p:−

0.07

(95

%

CI−

0.31

to0.

17)

Tele

phon

eco

unse

ling

grou

p−

Con

trol

grou

p:0.

18

(95

%C

I−0.

07to

0.42

)

MV

PA

Con

trol=

137.

8(S

E:1

68.4

)

Prin

tm

ater

ials=

174.

5(S

E:1

68.0

)

Tele

phon

eco

unse

ling=

153.

8(S

E:1

66.6

)

Ste

ps

Con

trol=

20,0

97(S

E:7

313)

Prin

tm

ater

ials=

20,8

21(S

E:7

308)

Tele

phon

eco

unse

ling=

21,4

93(S

E:7

249) (C

ontin

ued)

Frontiers in Endocrinology | Diabetes January 2013 | Volume 4 | Article 3 | 10

Plotnikoff et al. Community-based physical activity interventions

Tab

le1

|Co

nti

nu

ed

Au

tho

r

(dat

e)

Stu

dy

des

ign

NN

per

gro

up

Rec

ruit

men

tIn

terv

enti

on

len

gth

Co

mm

un

ity

bas

ed

inte

rven

tio

n

com

po

nen

t

PAas

pec

tP

rim

ary

Ou

tco

me(

s)

Res

ult

s

Plo

tnik

off

etal

.

(201

0),

Can

ada

RC

T48

N(C

on)=

21 N(In

t)=

27

Dia

bete

scl

inic

s16

wee

ksR

ecru

itmen

t

Exe

rcis

e

prog

ram

cond

ucte

din

part

icip

ants

hom

e

RT

grou

ppe

rfor

med

regu

lar

exer

cise

prog

ram

Mus

cle

stre

ngth

Glu

cose

cont

rol

Mus

cle

stre

ngth

Ben

chpr

ess

1-R

M(k

g)

Str

uctu

red

exer

cise

prog

ram

on3

nonc

onse

cutiv

eda

ys

per

wee

k.E

ight

exer

cise

sw

ere

perf

orm

edpe

rse

ssio

n

16-w

eek

prog

ram

Two

sets

of10

–12

repe

titio

nsat

50–6

0%

of1-

RM

Mea

ndi

ffere

nce:

9.3

kg**

(CI:

3.7,

15.0

)Int

>C

on

Con

trol=

44.8

(15.

9)

Inte

rven

tion=

55.5

(26.

3)

Leg

pres

s1-

RM

(kg)

Mea

ndi

ffere

nce:

41kg

**(C

I:15

.1,6

7.0)

Int>

Con

Con

trol=

93.8

(63.

4)

Inte

rven

tion=

175.

0(14

1.5)

Upr

ight

row

1-R

M(k

g)

Mea

ndi

ffere

nce:

12.4

kg**

(CI:

7.3,

17.5

)Int

>C

on

Con

trol=

56.6

(20.

6)

Inte

rven

tion=

69.8

(27.

8)

HbA

1c(b

asel

ine

valu

es)

Con

trol=

6.81

%(0

.83)

;50.

93m

mol

/mol

Inte

rven

tion=

6.89

%(1

.50)

;51.

80m

mol

/mol

HbA

1c(p

ost-

inte

rven

tion

valu

es)

Con

trol=

6.77

%(0

.80)

;50.

49m

mol

/mol

Inte

rven

tion=

6.97

%(1

.35)

;52.

68m

mol

/mol

Adj

uste

ddi

ffere

nce

betw

een-

grou

ps:

(RT

grou

p–

cont

rolg

roup

)=0.

3;95

%C

I0.1

to0.

4

Dav

ies

etal

.

(200

8),

UK

Clu

ster

RC

T65

2N

(Con

)=

248

N(In

t)=

314

Com

mun

ityG

P

clin

ics

12m

onth

sD

eliv

ered

in

the

com

mun

ity

Gro

uped

ucat

ion

prog

ram

asso

ciat

ed

with

bene

fits

ofPA

HbA

1cH

bA1c

Con

trol

grou

p:

Bas

elin

em

ean=

7.9%

(2.0

);62

.84

mm

ol/m

ol

Mea

nch

ange

from

base

line

to

post

-inte

rven

tion=−

1.21

;95%

CI−

1.40

to−

1.02

Inte

rven

tion

grou

p:

Bas

elin

em

ean=

8.3%

(2.2

);67

.21

mm

ol/m

ol

Mea

nch

ange

from

base

line

to

post

-inte

rven

tion=−

1.49

;95%

CI−

1.69

to−

1.29

Con

,con

trol

;Int

,int

erve

ntio

n;B

MI,

body

mas

sin

dex;

GP,

gene

ralp

ract

ition

ers;

HbA

1c,h

emog

lobi

nA

1c;M

ETm

ins,

met

abol

iceq

uiva

lent

min

utes

;RC

T,ra

ndom

ized

cont

rolt

rial;

MV

PA,m

oder

ate

tovi

goro

usph

ysic

al

activ

ity.

Not

e:N

umbe

rsin

pare

nthe

sis

()re

fer

toS

Dun

less

othe

rwis

est

ated

.

*Sig

nific

ant

(p<

0.5)

with

ingr

oup

chan

ge(o

negr

oup

desi

gn).

**S

igni

fican

t(p

<0.

5)be

twee

n-gr

oup

chan

ge(In

terv

entio

nvs

.Con

trol

).

www.frontiersin.org January 2013 | Volume 4 | Article 3 | 11

Plotnikoff et al. Community-based physical activity interventions

FIGURE 1 | Flow of study selection through the phases of the review.

the remaining studies were conducted in Canada (Mathieu et al.,2008; Plotnikoff et al., 2010, 2011, 2012), Australia (Dunstan et al.,2006; Engel and Lindner, 2006), the United Kingdom (Davies et al.,2008; Skoro-Kondza et al., 2009), Costa Rica (Goldhaber-Fiebertet al., 2003), and Jamaica (Less et al., 2009).

The majority of studies recruited participants from clinics orduring visits to general practitioners (14/22), however other com-munity recruitment strategies included local media campaigns(Dunstan et al., 2006; Engel and Lindner, 2006; Klug et al., 2008;Piette et al., 2011; Plotnikoff et al., 2012), intervention open tocommunity (Kruse et al., 2010), and a senior’s center’s (Speer et al.,2008). Four studies employed a combined approach of strategies.

For the purposes of this review, only community-basedapproaches were considered. The majority of studies (14/22)recruited participants by employing a multi-strategy approachincluding a combination of the following components: inter-vention recruitment, delivery, use of facilities, group sessions,and expert advice. In addition studies used of community facil-ities (Dunstan et al., 2006; Brooks et al., 2007; Davies et al.,

2008;, Skoro-Kondza et al., 2009) and expert advice from special-ists within the community (Goldhaber-Fiebert et al., 2003; Speeret al., 2008).

Physical activity interventions for treating T2D were eligi-ble only if the PA component/approach accounted for≥ 50% ofthe overall intervention. PA approaches including general exer-cise programs were employed by the majority of studies (9/22),PA counseling/information (Davies et al., 2008; Dutton et al.,2008; Klug et al., 2008; Plotnikoff et al., 2012), walking programs(Goldhaber-Fiebert et al., 2003; Piette et al., 2011), resistance train-ing programs (Dunstan et al., 2006; Brooks et al., 2007; Mathieuet al., 2008; Kruse et al., 2010; Plotnikoff et al., 2011), gym member-ship provision (Boyd et al., 2006), and yoga classes (Skoro-Kondzaet al., 2009).

META-ANALYSIS OF INTERVENTION EFFECTSResults from RCTs were pooled to establish the effects of inter-ventions on HbA1c levels (see Figure 2). As there was consid-erable heterogeneity among interventions [χ2

= 92.16, df= 11

Frontiers in Endocrinology | Diabetes January 2013 | Volume 4 | Article 3 | 12

Plotnikoff et al. Community-based physical activity interventions

FIGURE 2 | Forest plot for HbA1c.

(p < 0.00001); I2= 88%], the random effects models were used.

The impact of interventions on HbA1c levels approached statis-tical significance (−0.32% (mmol/mol=−26.99) [−0.65, 0.01],Z= 1.88 [p < 0.06]).

RISK OF BIASRisk of bias assessment was conducted (see Table 2). One studywas identified as low risk of bias (Kruse et al., 2010), 15 studieswere classified as moderate risk of bias, the remaining six stud-ies were rated as high risk of bias. Only three studies reportedassessor blinding (Item D), only four studies presented summaryresults, estimated effect sizes and precision estimates (Item G) and12 studies employed an objective measure of PA (Item I ).

RESULTS OF INCLUDED STUDIESA summary of results is presented in Table 1. A brief descrip-tion of results comparing intervention to control group effects ispresented below.

HbA1cThe association between community-based PA interventions andHbA1c was examined in 12 studies (see Table 1). HbA1c improve-ments were observed in eight studies. Four studies reported no dif-ference in HbA1c levels between intervention and control groups(Plotnikoff et al., 2011, 2012), or between baseline and follow-up(Engel and Lindner, 2006; Piette et al., 2011). Interestingly, in onestudy, a reduction in HbA1c levels was evident after the interven-tion phase, however this increased after 6 months (Skoro-Kondzaet al., 2009).

WeightWeight outcomes associated with the community-based PA inter-ventions were examined in four studies. Reduction in weightand BMI were observed in three of the four studies. Two stud-ies (Goldhaber-Fiebert et al., 2003; Boyd et al., 2006) reportedweight loss in comparison to the control group. While significantreductions in waist circumference and weight were observed inone study (Engel and Lindner, 2006), significant changes in BMIwas not evident until 6-month follow-up. One study reported nochange in BMI (Martyn-Nemeth et al., 2010).

Physical activityPhysical activity was reported in nine studies as a primary out-come. Increases in overall PA was observed (Keyserling et al., 2002;Engel and Lindner, 2006; Speer et al., 2008); specifically increasedtime spent in leisure time PA (Mathieu et al., 2008), number of par-ticipants meeting the PA guidelines (Martyn-Nemeth et al., 2010),increased amount of days engaging in PA (Klug et al., 2008), andincreased step counts (Plotnikoff et al., 2012). Two studies (Brookset al., 2007; Plotnikoff et al., 2011) also found improvements inmuscles strength and muscle quality. Three studies reported nochange or improvement in PA (Plotnikoff et al., 2012), muscularstrength (Taylor et al., 2009), or balance (Kruse et al., 2010).

Other outcomesOther health outcomes included improved quality of life (Mathieuet al., 2008), improved psychological well-being (Martyn-Nemethet al., 2010), reduced diastolic blood pressure (Boyd et al., 2006;Mathieu et al., 2008), reduced systolic blood pressure (Boyd et al.,2006; Mathieu et al., 2008), improvements in dietary and PAknowledge (Two Feathers et al., 2005), and decreased fastingplasma glucose (Goldhaber-Fiebert et al., 2003).

DISCUSSIONThe objective of this paper was to conduct a systematic review withmeta-analysis to investigate the effectiveness of community-basedPA interventions for the treatment of T2D in adult populations.Effective treatment/management of T2D can prevent the develop-ment of microvascular complications and risk of cardiovasculardiseases, which are the leading cause of death in diabetic patients(Stamler et al., 1993; Gilmer et al., 1997). In this review, 22 eligi-ble studies were identified in which the PA component/approachaccounted for≥ 50% of the overall intervention. The studies wereconducted across different countries, and represented differentethnic/cultural groups. Although most studies were adequatelypowered, not all studies provided power calculations. The majorityof studies (16/22) employed a RCT design. The primary outcomesvaried across the studies; HbA1c, weight and PA were the primaryoutcomes in 11, 4, and 9 studies, respectively. Overall, only onestudy had a low risk of bias, and the vast majority of studies had

www.frontiersin.org January 2013 | Volume 4 | Article 3 | 13

Plotnikoff et al. Community-based physical activity interventions

Tab

le2

|Met

ho

do

log

ical

qu

alit

ysc

ore

san

dri

sko

fb

ias

inco

mm

un

ity-

bas

edPA

inte

rven

tio

ns

totr

eatT

2D.

Stu

dy

A(S

epar

ate

bas

elin

ere

sult

s)

B (ran

do

miz

atio

n)

C(d

rop

ou

t

rate

)

D(a

sses

sor

blin

din

g)

E(i

nte

nti

on

-

to-t

reat

)

F (co

nfo

un

der

s)

G(e

ffec

t

size

s)

H (pow

er)

I(o

bje

ctiv

e

PA)

Sco

re/9

Plo

tnik

off

etal

.(20

11)

XX

X(?

)X

X×

X×

6/9

Pie

tte

etal

.(20

11)

XX

X×

XX

×X

X7/

9

Kru

seet

al.(

2010

)X

XX

XX

XX

XX

9/9

Mar

tyn-

Nem

eth

etal

.(20

10)

n/a

n/a

Xn/

a×

××

××

1/9

Mat

hieu

etal

.(20

08)

XX

××

X×

××

X4/

9

Tayl

oret

al.(

2009

)X

X×

×X

X×

XX

6/9

Sko

ro-K

ondz

aet

al.(

2009

)X

X(?

)×

X(?

)×

×X

4/9

Dut

ton

etal

.(20

08)

XX

(?)

×X

X×

××

4/9

Klu

get

al.(

2008

)n/

an/

a×

n/a

×X

××

×1/

9

Spe

eret

al.(

2008

)n/

an/

a×

n/a

××

×X

×1/

9

Bro

oks

etal

.(20

07)

X×

XX

XX

××

X6/

9

Dun

stan

etal

.(20

06)

X×

X×

XX

××

X5/

9

Boy

det

al.(

2006

)n/

an/

aX

n/a

X×

××

X3/

9

Eng

elan

dLi

ndne

r(2

006)

XX

X×

×X

××

×4/

9

Two

Feat

hers

etal

.(20

05)

n/a

n/a

×n/

a×

X×

××

1/9

Bra

ndon

etal

.(20

03)

X×

××

××

×X

X3/

9

Gol

dhab

er-F

iebe

rtet

al.(

2003

)X

XX

XX

××

XX

7/9

Keys

erlin

get

al.(

2002

)X

XX

×X

X×

XX

7/9

Plo

tnik

off

etal

.,20

10X

XX

×X

×X

XX

7/9

Plo

tnik

off

etal

.(20

12)

XX

X×

×X

X×

×5/

9

Dav

ies

etal

.(20

08)

XX

××

XX

××

×4/

9

Less

etal

.(20

09)

X×

X×

XX

XX

×6/

9

Crit

eria

:A,

base

line

resu

ltsre

port

edse

para

tely

for

each

grou

p;B

,ra

ndom

izat

ion

clea

rlyde

scrib

edan

dad

equa

tely

com

plet

ed;

C,

drop

out≤

20%

for≤

6m

onth

sfo

llow

-up

and≤

30%

for>

6m

onth

sfo

llow

-up;

D,

asse

ssor

blin

ding

;E,i

nten

tion-

to-t

reat

anal

ysis

;F,c

onfo

unde

rsac

coun

ted

fori

nan

alys

es;G

,sum

mar

yre

sults

pres

ente

dan

des

timat

edef

fect

size

san

dpr

ecis

ion

estim

ates

;H,p

ower

calc

ulat

ion

repo

rted

and

stud

y

adeq

uate

lypo

wer

ed;I

,an

obje

ctiv

em

easu

reof

PAw

asus

ed.

Frontiers in Endocrinology | Diabetes January 2013 | Volume 4 | Article 3 | 14

Plotnikoff et al. Community-based physical activity interventions

a moderate risk of bias. Overall, this review demonstrates thatcommunity-based interventions utilizing a large PA componentcan be effective in treating T2D in terms of decreasing HbA1c lev-els, reducing weight (including BMI and waist circumference) andincreasing PA levels.

Eleven of the studies that reported HbA1c as an outcome mea-sure were pooled in the meta-analysis. Meta-analysis revealedcommunity-based PA interventions contribute to a lowering ofHbA1c by−0.32% (−26.99 mmol/mol) which approached statis-tical significance effect (p= 0.06). The results of the meta-analysesmay have indeed been stronger than what has been provided inthis report, as the control groups in three of the studies used in themeta-analysis (i.e., Dunstan et al., 2006; Engel and Lindner, 2006;Plotnikoff et al., 2011) received a greater intervention dose than“true controls” receiving no intervention and/or those receiving“standard practice.”

It is interesting to note that in a meta-analyses of 14 clini-cally based exercise trials in the T2D population, Boulé et al.(2001) reported exercise training decreased HbA1c by −0.66%(−30.71 mmol/mol) which is considered to significantly decreasethe clinical risk of diabetes complications in this population. Smallchanges in HbA1c may still be important from a public healthperspective, considering small changes at the individual levelcan translate to substantial changes within the population if thechanges are distributed across the entire target (T2D) population.Health promotion experts advocate that practical, low/minimalintensity interventions that might not have large clinical effects,but can be delivered to large numbers of participants, are morelikely to have a broader health impact (Tunis et al., 2003).

A variety of factors may have contributed to the limited changein HbA1c observed in the meta-analysis. A floor effect may haveoccurred in several studies because individuals may have had well-controlled diabetes at baseline. Some studies employed inclusioncriteria, where only individuals with high HbA1c levels at baselinewere admitted. The ADAPT study (Plotnikoff et al., 2012), one ofthe larger studies in this review that reported no impact on HbA1c,had a relatively low baseline level of HbA1c.

Different methods used to recruit individuals into the studiesmay also have had an impact on outcomes. Several of the studies inthis review had participant recruitment in the form of advertising(as employed by 6 of 22 studies). It is possible those who respond topublic advertising (e.g., media campaigns) may be a biased groupwho are more active and take “better care” of themselves. Thereis evidence that individuals who volunteer to participate in thesetypes of studies are healthier on average compared to those whodo not voluntarily participate in studies (Plotnikoff et al., 2008;Rosal et al., 2011).

Adherence to intervention studies may have also played a role.For example, a study conducted by Skoro-Kondza et al. (2009),included in this review, report a very low adherence resulting frommotivational barriers to program attendance. Also it is possiblethat individuals who are most in need of intervention treatmentdo not adhere to programs. For example, in the ADAPT study,Plotnikoff et al. (2012), found that many dropouts initially had lowlevels of PA, and it was also individuals most needy of PA inter-ventions that did not participate in the full length of the study.Future research should examine strategies to reach and encourage

inactive individuals who are genuinely in need of interventions toparticipate in research studies.

There was considerable heterogeneity in terms of the differ-ent intervention strategies used to promote PA. Some techniquesincluded structured group education programs, yoga classes, tele-phone counseling, YMBA membership, motivational techniques,providing multi-gym apparatus for home use and the use of laydiabetes facilitators. For adults with T2D, studies have shown bet-ter metabolic outcomes when interventions focus exclusively onPA compared to strategies that encompass multiple diabetes carebehaviors (Conn et al., 2007). However, this difference was notobserved in our review.

Considering that this review examined interventions thatused≥ 50% PA component, other factors of the program besidesPA may also have influenced HbA1c levels in participants, such aschanges in diet and sitting time. Fourteen of the 22 studies hada dietary component and/or measured dietary changes. However,studies have shown when diet and exercise approaches were com-bined, the effect on HbA1c was similar to the effect of exercisealone (Boulé et al., 2001). Further, no studies included the pro-motion of reducing sitting time behavior. Future studies shouldinclude strategies to promote the reduction of this sedentary activ-ity given its negative impact on metabolic health (van der Ploeget al., 2012).

In the set of studies reviewed, five studies (Dunstan et al.,2006; Brooks et al., 2007; Mathieu et al., 2008; Kruse et al., 2010;Plotnikoff et al., 2011) incorporated resistance training mode ofactivity either as the main focus of the intervention or as part of theintervention. Resistance training has been recognized as a usefultherapeutic tool for the treatment of a number of chronic diseasesand has been demonstrated as safe and efficacious for elderly andobese individuals (Eves and Plotnikoff, 2006). Resistance traininghas the potential for increasing muscle strength, lean muscle mass,and bone mineral density, which could enhance functional statusand glycemic control.

It is generally accepted that behavior change programs for PAthat are theoretically grounded are more efficacious/effective thana theoretical strategies (Biddle et al., 2007). In the reviewed stud-ies, six studies had interventions based in social-cognitive theories,or interventions that were tailored based on theories. Differenttheoretical frameworks utilized include Social Ecological Model,Health Belief Model, Theory of Planned Behavior, and Social-Cognitive Theory. In this regard, it should be noted that thereare limited PA theory-based, long-term studies with large samplestargeting the adult T2D population. Few studies have exploredthe mechanisms of behavior change in community-based inter-ventions for individuals with T2D. One notable exception is theADHERES study (Plotnikoff et al., 2010) which reported in a sec-ondary analysis that RT planning strategies mediated the effect ofthe intervention on RT behavior (Lubans et al., 2011). Researchersare encouraged to operationalize their interventions and test thehypothesized mediators of behavior change.

In this review, 10 studies reviewed employed self-report mea-sures for PA, and 12 studies employed objective measures.Although assessing objective measures of PA may not always be fea-sible in community-based interventions, objective measurementtechniques using recent technologies such as accelerometers and

www.frontiersin.org January 2013 | Volume 4 | Article 3 | 15

Plotnikoff et al. Community-based physical activity interventions

pedometers as well as GPS devices wherever possible should beused to assess levels of PA given the over-reporting in self-reportPA measures.

In terms of study follow-up, the maximum followed-up timewas 2 years (Brandon et al., 2003). In the literature, there is gener-ally limited information on long-term outcomes of interventions,and sharp declines in intervention benefits after a few months havealso been observed (Deakin et al., 2005). Future studies need toinvestigate the effectiveness of interventions to produce long-termchange in PA levels, and how effectiveness could be sustained byusing strategies such as booster sessions following interventions tofacilitate longer-term behavior change.

Considering that the studies included in this review representa variety of ethnic groups (Mexican, Latino, Caucasian, African-American), a range of age groups (ages 52 to 73), and were 64%female, study findings are perhaps by and large generalizable. How-ever, the variability between the economic and health status ofcountries included in this review (for example, USA and Canadavs. Jamaica and Costa Rica), must also be considered in the inter-pretation of the study’s results. In the meta-analysis, we did notconduct additional analyses to investigate intervention featuresthat may moderate intervention effectiveness (e.g., mode of deliv-ery, objective PA vs. self-report PA, age, sex) due to the relativelyfew number of studies that qualified for the analysis. Further stud-ies which collect such information could be useful to inform thedevelopment of future research and interventions. As the numberof trials of community-based interventions to increase PA grows,future research should attempt to isolate the impact of specificintervention features on PA change as well as implement highquality study-designs that will allow such investigations.

In terms of a health economical aspect, there is evidence oflower health care utilization and costs in T2D individuals whomeet minimum PA guidelines (Plotnikoff et al., 2008). Otherstudies have shown community-based programs that target PAappear to be cost-effective for individuals with diabetes (Jacobs-van der Bruggen et al., 2007; Roux et al., 2008). Community-basedapproaches to increase PA have the potential to improve program

adherence by using intervention strategies that directly addressfactors such as socio-economic, psychological, social, and culturalelements (Rosal et al., 2011). Although resource-intensive, clin-ical approaches can be effective. Future research may considerthe combination of clinical and community-based approaches tomaximize treatment effects.

Strengths of the study include being the first study, to ourknowledge, to meta-analyze solely on community-based interven-tions to increase PA. Studies representative of a variety of culturalgroups from different countries is an additional strength. Limi-tations include the difficulty in generalizing the contribution ofmulti-component interventions, and the relatively small numberof studies that were included in the meta-analysis. The reviewedstudies have moderate levels of risk bias in the methodologicalquality which should also be considered in the interpretation ofthese findings. Future community-based trials should be method-ologically more rigorous. Publication bias of studies should alsobe taken into consideration when interpreting the results of thismeta-analysis. Studies that do not find statistically significantresults may not be published either due to authors’ not attempt-ing to publish or journals not accepting the article for publica-tion, and therefore, this meta-analysis should be interpreted withcaution.

Overall, our findings demonstrate community-based PA inter-ventions can be effective for increasing PA and decreasing HbA1clevels. The change in HbA1c was clinically significant in the meta-analysis. Producing even small effects can be meaningful at a‘population level’ for adults living with T2D. As the number oftrials of community-based interventions to increase PA grows inthe future, the findings of this meta-analysis will need to be repli-cated, and intervention features and elements that may moderateintervention effectiveness can be further investigated.

ACKNOWLEDGMENTSRonald C. Plotnikoff is supported by a senior research fellow-ship salary award from the National Health and Medical ResearchCouncil (NHMRC), Australia.

REFERENCESBalducci, S., Zanuso, S., Nicolucci, A.,

De Feo, P., Cavallo, S., Cardelli, P.,et al. (2010). Effect of an inten-sive exercise intervention strategy onmodifiable cardiovascular risk fac-tors in subjects with type 2 diabetesmellitus: a randomized controlledtrial: the Italian diabetes and exer-cise study (IDES). Arch. Intern. Med.170, 1794–1803.

Biddle, S. J. H., Hagger, M. S., Chatzis-arantis, N. L. D., and Lippke,S. (2007). “Theoretical frameworksin exercise psychology,” in Hand-book of Sport Psychology, 3rdEdn, eds G. Tenenbaum and R.C. Eklund (New York: Wiley),537–559.

Boulé, N. G., Haddad, E., Kenny, G. P.,Wells, G. A., and Sigal, R. J. (2001).Effects of exercise on glycemic con-trol and body mass in type 2 diabetes

mellitus: a meta-analysis of con-trolled clinical trials. JAMA 286,1218–1227.

Boyd, S. T., Scott, D. M., and Augus-tine, S. C. (2006). Exercise for low-income patients with diabetes. Dia-betes Educ. 32, 385–393.

Brandon, L., Gaasch, D. A., Boyette,L. W., and Lloyd, A. M. (2003).Effects of long-term resistive train-ing on mobility and strength inolder adults with diabetes. J. Geron-tol. A Biol. Sci. Med. Sci. 58,M740–M745.

Brooks, N., Layne, J. E., Gordon, P.L., Roubenoff, R., Nelson, N. E.,and Castaneda-Sceppa, C. (2007).Strength training improves musclequality and insulin sensitivity in his-panic older adults with type 2 dia-betes. Int. J. Med. Sci. 4, 19–27.

Church, T. S., Blair, S. N., andCocreham, S. (2010). Effects of

aerobic and resistance training onhemoglobin A1c levels in patientswith type 2 diabetes. JAMA 302,2253–2262.

Conn, V., Hafdahl, A., Mehr, D.,LeMaster, J. W., Brown, S. A.,and Nielsen, P. J. (2007). Meta-bolic effects of interventions toincrease exercise among adults withtype 2 diabetes. Diabetologia 50,913–921.

Davies, M. J., Heller, S., Skinner, T.C., Campbell, M. J., Carey, M. E.,Cradock, S., et al. (2008). Effective-ness of the diabetes education andself-management for ongoing andnewly diagnosed (DESMOND) pro-gramme for people with newly diag-nosed type 2 diabetes: cluster ran-domised controlled trial. BMJ 336,491–495.

Deakin, T., McShane, C., Cade, J., andWilliams, R. (2005). Group based

training for self-management strate-gies in people with type 2 dia-betes. Cochrane Database Syst. Rev.2, CD003417.

Deeks, J. J., Altman, D. G., andBradburn, M. J. (2008). “Statisti-cal methods for examining hetero-geneity and combining results fromseveral studies in meta-analysis,” inSystematic Reviews in Health Care:Meta-Analysis in Context, 2nd Edn,Chap. 15, eds M. Egger, G. D. Smith,and D. G. Altman (London: BMJPublishing Group), doi: 10.1002/9780470693926.

Dunstan, D. W., Vulikh, E., Owen,N., Jolley, D., Shaw, J., and Zim-met, P. (2006). Communitycenter–based resistance train-ing for the maintenance ofglycemic control in adults withtype 2 diabetes. Diabetes Care 29,2586–2591.

Frontiers in Endocrinology | Diabetes January 2013 | Volume 4 | Article 3 | 16

Plotnikoff et al. Community-based physical activity interventions

Dutton, G. R., Provost, B. C., Tan, F., andSmith, D. (2008). A tailored print-based physical activity interventionfor patients with type 2 diabetes.Prev. Med. 47, 409–411.

Engel, L., and Lindner, H. (2006).Impact of using a pedometer on timespent walking in older adults withtype 2 diabetes. Diabetes Educ. 32,98–107.

Eves, N. D., and Plotnikoff, R. C. (2006).Resistance training and type 2 dia-betes. Diabetes Care 29, 1933–1941.

Gilmer, T. P., O’Connor, P. J., Man-ning, W. G., and Rush, W. A. (1997).The cost to health plans of poorglycemic control. Diabetes Care 20,1847–1853.

Goldhaber-Fiebert, J. D., Goldhaber-Fiebert, S. N., Tristán, M. L., andNathan, D. M. (2003). Random-ized controlled community-basednutrition and exercise interventionimproves glycemia and cardiovascu-lar risk factors in type 2 diabeticpatients in rural Costa Rica. DiabetesCare 26, 24–29.

Jacobs-van der Bruggen, M. A., Bos,G., Bemelmans, W. J., Hoogenveen,R. T., Vijgen, S. M., and Baan, C.A. (2007). Lifestyle interventions arecost-effective in people with differ-ent levels of diabetes risk: resultsfrom a modeling study. DiabetesCare 30, 128–134.

Keyserling, T. C., Samuel-Hodge, C.D., Ammerman, A. S., Ainsworth,B. E., Henríquez-Roldán, C. F.,Elasy, T. A., et al. (2002). Arandomized trial of an interven-tion to improve self-care behaviorsof African-American women withtype 2 diabetes. Diabetes Care 25,1576–1583.

Klug, C. D., Toobert, D. J., and Fogerty,M. (2008). Healthy changes™for liv-ing with diabetes. Diabetes Educ. 34,1053–1061.

Kruse, R. L., LeMaster, J. W., and Mad-sen, R. W. (2010). Fall and bal-ance outcomes after an interventionto promote leg strength, balance,and walking in people with diabeticperipheral neuropathy: “feet first”randomized controlled trial. Phys.Ther. 90, 1568–1579.

Less, L. A., Ragoobirsingh, D., Morri-son, E. Y., Boyne, M., and John-son, P. A. (2009). A preliminaryreport on an assessment of acommunity-based intervention fordiabetes control in adults withtype 2 diabetes. Fam. Pract. 27,i46–i52.

Lubans, D. R., Plotnikoff, R. C., Jung,M., Eves, N., and Sigal, R. (2011).Testing mediator variables in a resis-tance training intervention for obeseadults with type 2 diabetes. Psychol.Health 27, 1388–1404.

Martyn-Nemeth, P. A., Vitale, G. A., andCowger, D. R. (2010). A culturallyfocused exercise program in hispanicadults with type 2 diabetes. DiabetesEduc. 36, 258–267.

Mathieu, M., Brochu, M., and Béliveau,L. (2008). Diabetaction: changes inphysical activity practice, fitness, andmetabolic syndrome in type 2 dia-betic and at-risk individuals. Clin. J.Sport Med. 18, 70–75.

Nolan, C. J., Damm, P., and Prentki, M.(2011). Type 2 diabetes across gen-erations: from pathophysiology toprevention and management. Lancet378, 169–181.

Piette, J. D., Richardson, C., Himle, J.,Duffy, S., Torres, T., Vogel, M., etal. (2011). A randomized trial oftelephonic counseling plus walkingfor depressed diabetes patients. Med.Care 49, 641–648.

Plotnikoff, R. (2006). Physical activ-ity in the management of diabetes:population-based perspectives andstrategies. Can. J. Diabetes 30, 52–62.

Plotnikoff, R. C., Eves, N., Jung, M.,Sigal, R. J., Padwal, R., and Karuna-muni, N. (2010). Multicomponent,home-based resistance training forobese adults with type 2 diabetes: arandomized controlled trial. Int. J.Obes. 34, 1733–1741.

Plotnikoff, R. C., Karunamuni, N.,Courneya, K., Sigal, R., Johnson,J., and Johnson, S. (2012). TheAlberta diabetes and physical activ-ity trial (adapt): a randomized trialevaluating theory-based interven-tions to increase physical activity inadults with type 2 diabetes. Ann.Behav. Med. doi: 10.1007/s12160-012-9405-2. [Epub ahead of print].

Plotnikoff,R. C.,Karunamuni,N., John-son, J., Kotovych, M., and Sven-son, L. (2008). Health-related behav-iours in adults with diabetes: asso-ciations with health care utilizationand costs. Can. J. Public Health 99,227–231.

Plotnikoff, R. C., Pickering, M., Glenn,N., Doze, S., Reinbold-Matthews,M., McLeod, L., et al. (2011). Theeffects of a supplemental, theory-based physical activity counselingintervention for adults with type2 diabetes. J. Phys. Act. Health 8,944–954.

Rosal, M. C., Ockene, I. S., Restrepo, A.,White, M. J., Borg, A., Olendzki, B.,et al. (2011). Randomized trial of aliteracy-sensitive, culturally tailoreddiabetes self-management interven-tion for low-income latinos. DiabetesCare 34, 838–844.

Roux, L., Pratt, M., Tengs, T., Yore,M., Yanagawa, T., Van Den Bos, J.,et al. (2008). Cost effectiveness ofcommunity-based physical activityinterventions. Am. J. Prev. Med. 35,578–588.

Shaw, J. E., Sicree, R. A., and Zimmet,P. Z. (2010). Global estimates of theprevalence of diabetes for 2010 and2030. Diabetes Res. Clin. Pract. 87,4–14.

Sigal, R. J., Tai, S., Gadelrab, R., Drince-vic, D., and Greenhalgh, T. (2007).Effects of aerobic training, resistancetraining, or both on glycemic con-trol in type 2 diabetes: a randomizedtrial. Ann. Intern. Med. 147, 357–369.

Skoro-Kondza, L., Tai, S., Gadelrab,R., Drincevic, D., and Greenhalgh,T. (2009). Community based yogaclasses for type 2 diabetes: anexploratory randomised controlledtrial. BMC Health Serv. Res. 9:33.doi:10.1186/1472-6963-9-33

Speer, E. M., Reddy, S., Lommel, T.S., Fischer, J. G., Stephens, H.,Park, S., et al. (2008). Diabetesself-management behaviors and a1cimproved following a community-based intervention in older adults inGeorgia senior centers. J. Nutr. Elder.27, 179–200.

Stamler, J., Vaccaro, O., Neaton, J. D.,and Wentworth, D. (1993). Diabetes,other risk factors, and 12-yr cardio-vascular mortality for men screenedin the multiple risk factor interven-tion trial. Diabetes Care 16, 434–444.

Taylor, J. D., Fletcher, J. P., and Tiarks, J.(2009). Impact of physical therapist–directed exercise counseling com-bined with fitness center–based exer-cise training on muscular strengthand exercise capacity in people withtype 2 diabetes: a randomized clini-cal trial. Phys. Ther. 89, 884–892.

Tunis, S. R., Stryer, D. B., and Clancy,C. M. (2003). Practical clinical tri-als: increasing the value of clinicalresearch for decision making in clin-ical and health policy. JAMA 290,1624–1632.

Two Feathers, J., Kieffer, E. C.,Palmisano, G., Anderson, M.,Sinco, B., Janz, N., et al. (2005).Racial and ethnic approachesto community health (REACH)

detroit partnership: improvingdiabetes-related outcomes amongAfrican American and Latinoadults. Am. J. Public Health 95,1552–1560.

Vachon, G. C., Ezike, N., Brown-Walker,M.,Chhay,V.,Pikelny, I., and Pender-graft, T. B. (2007). Improving accessto diabetes care in an inner-city,community-based outpatient healthcenter with a monthly open-access,multistation group visit program. J.Natl. Med. Assoc. 99, 1327–1336.

van der Ploeg, H., Chey, T., Korda,R., Banks, E., and Bauman, A.(2012). Sitting time and all-causemortality risk in 222 497 Aus-tralian adults. Arch. Intern. Med. 172,494–500.

van Sluijs, E. M., McMinn, A. M., andGriffin, S. J. (2007). Effectiveness ofinterventions to promote physicalactivity in children and adolescents:systematic review of controlled tri-als. BMJ 335, 703.

Young, M. D., Morgan, P. J., Plotnikoff,R. C., Callister, R., and Collins, C.E. (2012). Effectiveness of male-only weight loss and weight lossmaintenance interventions: a sys-tematic review with meta-analysis.Obes. Rev. 13, 393–408.

Conflict of Interest Statement: Theauthors declare that the research wasconducted in the absence of any com-mercial or financial relationships thatcould be construed as a potential con-flict of interest.

Received: 27 September 2012; paperpending published: 22 October 2012;accepted: 07 January 2013; publishedonline: 29 January 2013.Citation: Plotnikoff RC, Costigan SA,Karunamuni ND and Lubans DR(2013) Community-based physical activ-ity interventions for treatment of type2 diabetes: a systematic review withmeta-analysis. Front. Endocrin. 4:3. doi:10.3389/fendo.2013.00003This article was submitted to Frontiersin Diabetes, a specialty of Frontiers inEndocrinology.Copyright © 2013 Plotnikoff, Costigan,Karunamuni and Lubans. This is anopen-access article distributed under theterms of the Creative Commons Attribu-tion License, which permits use, distrib-ution and reproduction in other forums,provided the original authors and sourceare credited and subject to any copy-right notices concerning any third-partygraphics etc.

www.frontiersin.org January 2013 | Volume 4 | Article 3 | 17