Download as pdf or txt
Download as pdf or txt
You are on page 1of 15

Preventive Medicine 47 (2008) 354–368

Contents lists available at ScienceDirect

Preventive Medicine
j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / y p m e d

Review

Long-term effectiveness of interventions promoting physical activity:


A systematic review
Falk Müller-Riemenschneider ⁎, Thomas Reinhold, Marc Nocon, Stefan N. Willich
Institute for Social Medicine, Epidemiology, and Health Economics, Charité University Medical Center, 10098 Berlin, Luisenstraße 57, Germany

a r t i c l e i n f o a b s t r a c t

Article history: Background. This systematic review aims to evaluate the long-term effectiveness of physical activity
Available online 15 July 2008 interventions targeted at healthy adults and to identify effective intervention components.
Methods. The systematic search in Pubmed, Embase, the Cochrane Library and Psycinfo identified
Keywords:
literature published in English and German between January 2001 and June 2007. We included randomised
Physical activity
controlled trials investigating physical activity interventions for healthy adults with a minimum study
Exercise
Intervention
duration of 12 months. Two researchers independently assessed publications according to pre-defined
Systematic review inclusion criteria and methodological quality was rated according to the SIGN criteria. Study characteristics
Meta-analysis and outcome measures were extracted, and pooled effect estimates with 95% confidence intervals calculated.
Randomised controlled trials Results. Of 5508 identified publications 25 studies met the inclusion criteria. There was substantial
heterogeneity in study quality, intervention strategies and intervention effects. Compared to no-intervention
and minimal-intervention control, gains in weekly energy expenditure and physical fitness of up to 975 kcal
and 11% were achieved, respectively. Booster interventions were used in 16 studies.
Conclusions. There is evidence for long-term increases in physical activity behaviour and physical fitness.
To improve uptake of physical activity additional tailored exercise prescription strategies seem promising.
Booster interventions such as phone, mail or internet can help to facilitate long-term effectiveness.
© 2008 Elsevier Inc. All rights reserved.

Contents

Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 354
Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 355
Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363
Long-term effectiveness of physical activity interventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 363
Sustainability of intervention effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 364
Intervention effectiveness in specific population groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
Effectiveness of specific Intervention components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
Initial intervention intensity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
Written exercise prescription . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
Tailored vs. standard intervention messages . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
Effectiveness of maintenance strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 365
Conflict of interest statement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 367
Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 367
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 367

Background

The positive effects of physical activity on health and mortality


when performed on a regular basis have been well established. In
⁎ Corresponding author.
E-mail address: falk.mueller-riemenschneider@charite.de
addition, physical activity plays an important role in the prevention of
(F. Müller-Riemenschneider). various chronic diseases, such as cardiovascular diseases, ischemic

0091-7435/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.ypmed.2008.07.006
F. Müller-Riemenschneider et al. / Preventive Medicine 47 (2008) 354–368 355

stroke, hypertension, obesity, diabetes mellitus, osteoporosis, colon reviews specifically investigating the long-term effectiveness of
cancers and fall related injuries (Camacho et al., 1991; Giovannucci et physical activity interventions in a rigorous way are currently not
al., 1995; Hu et al., 2001a; Sesso et al., 2000; Lee et al., 1995; Hu et al., available.
2000; Wannamethee and Shaper 1999; Hu et al., 2001b; Warburton et The present systematic review therefore aims to summarise the
al., 2006a; Wannamethee and Shaper 2001). While health benefits evidence for the long-term effectiveness of physical activity interven-
seem to increase with physical activity levels, already moderate levels tions and to identify efficacious intervention components. In the
have proved to be beneficial. Current guidelines therefore recommend context of this review long-term is defined as a study or follow-up
to achieve a weekly exercise based energy expenditure of about duration of at least 12 months.
1000 kcal (4200 kJ) or to accumulate 30 min of moderate intensity
physical activity over one day on most days of the week (Warburton et Methods
al., 2006a; US Department of Health and Human Services: Healthy
People 2010, 2000; Warburton et al., 2006b; Haskell et al., 2007). The focus of this systematic review is on current literature,
Despite the positive effects associated with moderate levels of published between January 2001 and June 2007. We identified
physical activity, sedentary life style remains a common problem in relevant publications by means of a structured search (see Table 1)
industrialised countries and the overall levels of physical activity of the following databases: Cochrane Library, PubMed, Embase and
continue to be low or even declined in recent years (Owen and PsycInfo. In addition, we conducted a manual search of reference lists
Bauman 1992; Brownson et al., 2005; Lampert et al., 2005). The included in the articles identified as part of the structured database
initiation and maintenance of exercise in adults, adolescents and search described above. Titles and abstracts of identified publications
children are therefore important objectives for the promotion of were screened for relevance by one researcher without restrictions on
health and crucial in the prevention of the increasing burden of study design. Two researchers independently assessed full-texts of
chronic diseases. In view of the public health importance, many included studies with regard to relevance and methodological quality.
countries are aiming to increase the uptake of overall physical activity Disagreements were resolved in discussion. We included studies if
levels and the participation in leisure time activity within the they met the following criteria.
population (Department of Culture Media and Sport & Strategy Unit, Inclusion criteria:
2002; US Department of Health and Human Services: Healthy People
– study design: randomised controlled trials
2010, 2000).
– outcome measures: participation in or level of physical activity or
In these attempts to promote and maintain higher levels of
physical fitness (e.g. proportion of participants meeting recom-
physical activity, a variety of methods have been developed, including
mended targets, energy expenditure, maximum oxygen uptake)
interventions based on informational-, behavioural- and/or environ-
– interventions: exercise programs or interventions aimed to
mental approaches, targeting either the individual, groups of
promote and maintain physical activity regardless of the setting
individuals and/or the whole community. While many approaches
(e.g. worksite, primary health care, community)
proved to be effective and successfully increased physical activity in
the short-term, subjects might fall back into their routine during long-
term interventions or once the intervention period is completed
(Robison and Rogers 1994; Rhodes et al., 1999). Maintenance and
long-term adherence to physical activity, however, is essential to
achieve sustainable public health effects. More recent intervention
studies therefore tried to investigate the long-term maintenance of
increased physical activity levels. To help people exercise continuously
following the initial intervention, booster and reminder strategies via
follow-up workshops, print materials, newsletters, phone or new
information technologies have been developed (Simons-Morton et al.,
2001; Burke et al., 2003; Cox et al., 2003). However, systematic

Table 1
Search strategy

#1 Motor activity (Mesh)


#2 Exertion (Mesh)
#3 Leisure activities (Mesh)
#4 Exercise therapy (Mesh)
#5 Physical education and training (Mesh)
#6 Physical fitness (Mesh)
#7 Preventive health services (Mesh)
#8 Health promotion (Mesh)
#9 Primary prevention (Mesh)
#10 Behavior therapy (Mesh)
#11 Primary health care (Mesh)
#12 Workplace (Mesh)
#13 Risk-reduction behaviour (Mesh)
#14 Community health services (Mesh)
#15 Health care evaluation mechanisms (Mesh)
#16 Randomized controlled trial (Mesh)
#17 Review literature (Mesh)
#18 #1 or #2 or #3 or #4 or #5 or #6
#19 #7 or #8 or #9 or #10 or #11 or #12 or #13 or #14
#20 #15 or #16 or #17
Fig. 1. Flow-chart of study selection for the systematic review and meta-analysis of the
#21 #18 and #19 and #20
long-term effectiveness of physical activity interventions compared to no-intervention
Mesh: Medical Subject Heading. control, minimal-intervention control and alternative-intervention control.
356
Table 2
Long-term effectiveness of physical activity interventions

Author Year Study Intervention components Participants (Age, Sex) Follow-up Self-reported Self-reported physical activity Physical fitness
Quality baseline comparability Maintenance physical activity (Energy expenditure)
of groups strategy
Elley et al., (2003) Green prescription (I): Sedentary adults FU: 12 months Targets (2.5 h/wk): TEE (kcal/kg/wk), change
(40–79 yrs). (I: 85%, C: 85%) to BL
High • physician counselling, exercise planning, • I: 23 practices, n = 451 Booster • I: 14.6% (66/451) • I: 9.76 (5.85–13.68)
exercise prescription and phone counselling (57 ± 11, F: 67%)
• newsletters and feedback • C: 23 practices, n = 427 • C: 4.9% (21/427) • C: 0.37 (−3–39–4.14)
(58 ± 12F: 66%)
Control (C): Groups comparable: • BGD: p = 0.003 • BGD: 9.28 (3.96–14.81) =
975 kcal/wk (p = 0.001)
• delayed intervention • Yes
Stewart et al. (2001) Champs II Intervention (I): Healthy sedentary older FU: 12 months (95%) TEE (kcal/wk), change to BL:
adults (65+).
High • staff supported PA planning • I: n = 81 (74 ± 6, F: 69%) Booster • I: 687, p b 0.001

F. Müller-Riemenschneider et al. / Preventive Medicine 47 (2008) 354–368


• print materials and activity log • C: n = 83 (75 ± 6, F: 63%) • C: −10, NS
• 10 group sessions (monthly) Groups comparable: Data for Meta-Analysis⁎:
• regular counselling calls • Yes • I: 374 (SD:260)
Control (C): • C: 292 (SD:244)
• delayed intervention
Hillsdon et al. (2002) All intervention participants: Sedentary adults FU: 12 months % change in EE:
(45–64 yrs). (I1: 32%, I2: 33%, C: 57%)
Good • 30 min primary care based health check • I1: n = 302 (55 ± 6, F: 53%) Booster • I1+2:124 (110–137)
• 6 telephone contacts over 34 weeks • I2: n = 285 (55 ± 6, F: 50%) • C: 113 (95–133), BGD:
3.7, p = 0.39
• exercise logs • C: n = 553 (55± 5, F: 50%) % change in EE by
intervention group:
Direct advice (I1): Groups comparable: • I2: 148 (117–183)
• benefits of PA and risks of inactivity • Yes • I1: 98 (75–125). BGD: 10.2,
p = 0.16
• recommend 30 min walking 5d/wk
Brief negotiation (I2):
• negotiation about positive and negative
outcomes of attempts to increase activity
Control (C):
• repeat health check after 11 months
Napolitano et al. (2006) Choose to move (I1): Healthy sedentary women FU: 12 months PA (min/wk):
(overall: 92.9%)
Good • Standard 12 wk program (AHA) with • I1: n = 93 (47 ± 10) None • I1: 154.48 (SE: 19.51)
changing topics
• booklet and letter • I2: n = 95 (48 ± 11) • I2: 148.87 (SE: 19.13)
Jumpstart (I2): • C: n = 92 (47 ± 11) • C: 139.52 (SE: 19.61)
• BL, 1, 3 and 6 months questionnaire Groups comparable: • BGD: NS
and tailored feedback report addressing
self-efficacy, barriers, benefits, social
support, goal setting and a booklet
Contact control (C): • N.a.
• One mailing including general health
information material
Aittasalo et al. (2004) Counselling (I1): Healthy sedentary FU: 12 months (99%) Targets (moderate: LTPA-EE (kcal/wk): overall
employees (volunteered) 30 min 4d/wk, strenuous: slight increase (p = 0.011)
30 min 2d /wk):
Fair • BL, 8wks, 6 and 12 months. • I1: n = 52 (45 ± 9, F: 61%) Repeat Intervention • Overall slight increase • I1+2 vs. C: −17.9%
(p = 0.049) (−44.4–21.3)
• PA assessment and activity log • I2: n = 51 • I1 vs. I2 8.9% (−30.3−70.2)
(44 ± 10, F: 53%)
• Individual goal setting and PA planning • C: n = 52 (42 ± 9, F: 54%)
Counselling + fitness test (I2): Groups comparable:
• additional fitness tests BL, 6 and • No
12 months.
• discussion and PA-plan adaptation
Control (C):
• usual care
21
Burke et al. (2003) 16 week intervention for all participants Healthy couples FU: 12 months Targets (30 min 5d/wk), Physical fitness
completers: (75% max heart rate):
Fair High-level (I1): • I1: n = 47 (age M/F: 31/27) (I1: 57%, I3: 42%, C: 72%) • C: 42% (13/31) • Greatest increase and
level in high-level group
• biweekly modules, alternating personal • I2: n = 47 (age M/F: 30/28) None • I1:35% (7/20) • BGD: significant
contact session and mail
Low-level (I2): • C: n = 43 (age M/F: 32/29) • I1: 42% (11/27)
• one initial personal contact session Groups comparable: • BGD: NS
followed by biweekly mail
• Control (C): • No
• no-intervention

F. Müller-Riemenschneider et al. / Preventive Medicine 47 (2008) 354–368


Campbell et al. (2002) PA as part of multi component worksite Female rural workers. FU: 18 months (75.6%) Frequency of PA TEE (MET):
health promotion program (frequency/wk)
Fair Intervention (I): • I: 4 worksites, n = 362 Repeat Intervention • I: 1.6 ± 2.0 I: 4.9 ± 7.7
(17% N 50 yrs)
• individualised health messages • C: 5 worksites, n = 298 • C: 1.5 ± 1.6 C: 4.1 ± 5.7
(e.g. feedback, strategies for change, (20% N 50 yrs)
community resource) after baseline and
6 months
• lay helpers support Groups comparable: • BGD: p = 0.51 BGD: p = 0.46
Control (C): • No
• after 6 months one magazine
Eiben and Lissner (2006) Health Hunters (I): Healthy sedentary women FU: 12 months Change in self-reported Change in time on
(18–28yrs) (I: 78%, C: 73%) EE (kcal/wk) treadmill (min)
Fair • BL examination and PA counselling • I: n = 18 Booster • I: 1464 ± 96) • I: 0.7 ± 0.4
• Individualised support package including • C: n = 22 • C: 200 ± 383) • C: −0.3 ± 0.4
informational and self-help material
focussing on PA, diet and weight control
• Continuous contact: personal, e-mail Groups comparable: • BGD: p = 0.03 • BGD: p = 0.08
and phone
Control (C): • No
• Delayed intervention
Yancey et al. (2006) Intervention (I): Healthy afro-american FU: 12 months Self-reported PA: Change in 1 mile walk
women (overall: N 70%) time (min)
Fair • 8 weekly (120 min) group sessions related • I: n = 188 (45 ± 11) Gym • BGD: NS • I: 1.9 (n = 72)
to exercise skills training and dietary advice
• Free gym membership for participant and • C: n = 178 (47 ± 11) • C: 2.3 (n = 61)
one other person
• Incentives: pedometers, exercise bands etc. Groups comparable: • BGD: p = 0.1
Control (C): • Yes
• 8 weekly sessions focussing on general
health
• Delayed free gym membership

Studies comparing intervention vs. no-intervention control.


PA: physical activity, M: male, F: female, FU: follow-up, BL: baseline, BGD: between group difference, NS: not significant, MWT: minute walk test, AHA: American Heart Association, n.a: not available, SE: standard error, LTPA: leisure time
physical activity, TEE: total energy expenditure, EE: energy expenditure, MET: metabolic equivalents, ⁎ data not from original publication but previous systematic review.

357
358
Table 3

F. Müller-Riemenschneider et al. / Preventive Medicine 47 (2008) 354–368


Long-term effectiveness of physical activity interventions. Studies comparing intervention vs. minimal-intervention control

Author Year Study Intervention components Participants (Age, Sex) Baseline comparability of Follow-up Maintenance PA-maintenance Energy Physical Fitness
Quality groups strategy Expenditure
Petrella et al. (2003) All participants: Healthy older adults (65+) FU: 12 months (86%) Compliance (I1: ≥3 sessions/wk VO2max (ml/kg/min):
at target heart rate, C: according to
ACSM recommendations):
High • physician counselling (ACSM) • I1: n = 131 (74 ± 4, F: 50%) Repeat Intervention • I1: 71% (93/131) • I1: 24.9 ± 1.3 (+15%)
• list of exercise facilities • C: n = 110 (73 ± 6, F: 46%) • C: 56% (62/110) • C: 22.8 ± 0.9 (+3%)
• repeat visits at 3, 6 and 12 months Groups comparable: • BGD: p b 0.05 • BGD: p b 0.001
STEP prescription group (I1): • yes
• prescribed exercise intensity
(75% of VO2max) and adjustment at
FU-visits
• activity logs (including heart rate)
Guidelines group (C):
• guidelines on physical activity
• activity log (no heart rate)
Marcus et al. (2007a) Motivationally tailored internet (I1): Healthy sedentary adults (18+ yrs) FU: 12 months (87%) Targets (30 min 5d/wk) VO2 (ml/kg⁎min)
High • Program website with information, • I1: n = 81 (F: 82%, 45 ± 9) Booster • I1: 39.5% • I1: 26.1 ± 6.9
goal setting tools and links
• Regular e-mail prompts over 12 months • I2: n = 86 (F: 84%, 45 ± 10) • I2: 32.6% • I2: 26.2 ± 6.9
• Monthly online questionnaires, tailored • C: n = 82 (F: 83%, 46 ± 9) • C: 30.5% • C: 25.7 ± 6.0
feedback and activity logs, financial
incentives
Motivationally tailored print (I2): Groups comparable: • BGD:p = 0.45 • BGD: p = 0.31
• Same intervention components via • Yes
print mail
Standard internet (C):
• Identical questionnaire and activity logs
• Links to 6 standard PA websites
Dubbert et al. (2002) All participants: Sedentary older adults (60-80yrs) FU: 12 months Walking for exercise (min/wk), Physical fitness (6 MWT), overall:
in stable health. (I1: 85%, I2: 71%, C: 85%) Overall:
Good • activity counselling • I1:: n = 69 (69 ± 5) Booster • 87.2 ± 99.4, p = 0.0001 • 1504 (301) feet, p = 0.0001 (BGD:
(video, nurse advice and goal setting) NS)
• walking diaries/incentives • I2: n = 73 (68 ± 4) • I2: N I1 and C (p = 0.002)
Personal phone calls (I1): • C: n = 70 (69 ± 5)
• 20 counselling calls over 12 months • Overall: 2 /179 female
Automated phone calls (I2): Groups comparable:
• 10 counselling and 10 automated calls •No
Control (C):
• no phone calls
Resnicow et al. (2005) Tailored (I1): Healthy adults (randomised per church): FU: 12 month (86%) Any PA (min/wk):
Good • culturally tailored materials • C: n = 267 (46 ± 13, F: 74%) Bbooster • C: 826 (SE: 53.6)
• exercise video and exercise guide • I1: n = 355 (46 ± 13, F: 76%) • I1: 928 (SE: 47.8)
• walking recommended, pedometer • I2: n = 304 (47 ± 14, F: 78%) • I2: 936 (SE: 50.7)
• audiocassette to accompany exercise Groups comparable: • BGD: I1 and I2 increased
significantly more than C,
I1 vs. I2: NS
• newsletters • Yes
Tailored + phone (I2):
• 4 additional phone counselling calls by
psychologists (2 calls addressing PA)
Standard (C):
• Standard material
Bock et al. (2001) Intervention materials following BL, Healthy sedentary adults FU: 12 months (63%) Targets (30 min 5d/wk):
1, 3 and 6 months questionnaires
Fair Intervention (I): • Overall: n = 194 (44 ± 11, F: 76%) Repeat Intervention • I: 42% (26/61)
• Individualised counselling messages Groups comparable: • C: 25% (15/59)
(benefits/barriers, rewards etc.)

F. Müller-Riemenschneider et al. / Preventive Medicine 47 (2008) 354–368


• stage specific self-help manuals • n.a. • BGD: p b 0.05
Standard (C):
• 4 standard booklets (AHA)
Lowther et al. (2002) According to preference participants Disadvantaged sedentary adults FU: 12 months LTPA (min/wk):
were allocated into two separate RCTs (FA: 25%, FC: 39%,
EC: 55%, CE: 41%)
Fair All participants: • FA: n = 40 (32 ± 13, F: 40%) Repeat Intervention • FA: 281 ± 100
• ask questions relating to the study or PA • FC: n = 41(40 ± 15, F: 37%) • FC: 317 ± 159
• print materials and exercise vouchers • EC: n = 40 (45 ± 19, F: 28%) • EC: 314 ± 160
• questionnaire (4 wks, 3, 6 and • CE: n = 34 (53 ± 16, F: 21%) • CE: 248 ± 158
12 months)
• re-interventions after 3 months Groups comparable: Change to FU:
1. Fitness assessment (FA): • No • Only EC increased PA
significantly
• standardised FA protocol vs. no-FA (FC)
2. Exercise consultation (EC):
•standardised EC (30 min) vs. no-EC (CE)

PA: physical activity, FU: follow-up, BL: baseline, AHA: American Heart Association, M: male, F: female, BGD: between group difference, NS: not significant, TEE: total energy expenditure, EE: energy expenditure, ACSM: American College of
Sports Medicine, wk: week, LTPA: leisure time physical activity.

359
360
Table 4
Long-term effectiveness of physical activity interventions. Studies comparing intervention vs. alternative-intervention control

Author Year Intervention components Participants Baseline Follow-up Maintenance PA-maintenance Energy Expenditure Physical Fitness
Study Quality comparability of groups strategy
Cox et al. (2003) Intervention components for all groups: Sedentary healthy women FU: 18 months (CBM: 100%, Retention to exercise Change in energy expenditure VO2max (ml/kg/min), change at
(40–65 yrs). CBV: 64%, HBM: 63%, HBV: program at 18 months FU: (kJ/kg/day):: 18 months:

F. Müller-Riemenschneider et al. / Preventive Medicine 47 (2008) 354–368


60%)
High • stage matched information on heart rate CB-moderate: Booster CB: 81% (52/64) CB: 3.1 (0.8–5.5) Moderate: 2.0 (−1.9–5.9)%
measurement and activity log n = 31(48 ± 5)
• exercise preparation and safety CB-vigorous: HB: 61% (38/62) HB: 2.5 (−0.1–5.1) Vigorous: 2.8 (0.01–5.6)%
n = 33 (48 ± 5)
• 30 min walking and aerobics activity HB-moderate: BGD: both p b 0.05. BGD: NS BGD and from BL: NS
classes with gradual increase during initiation n = 32 (48 ± 6)
• moderate: 40–55% of Heart Rate reserve HB-vigorous: Significant only for CB-vigorous:
n = 30 (49 ± 6) 6.0 (1.3–10.7)
• vigorous: 65–80% of Heart Rate reserve Groups comparable:
Centre-based (CB): Yes
• 3 supervised sessions/wk over 6 months
Home-based (CB):
• 10 supervised training sessions over 5 wks
• followed by 3 unsupervised sessions/wk
(print material, video and audio support)
All participants were HB from month 6–18:
• 6 phone calls over 12 months
Duncan et al. All participants were assessed at baseline, 6 and Sedentary adults FU: 24 months Adherence: Change in VO2max (L/min):
(2005) 24 months. Participants in counselling conditions (30–69yrs). (69.5%, BGD: NS)
received: individual walking prescription
(training heart rate zones based on
treadmill tests)
High b moderate-intensity–low frequency (LowF): • HardI-HIF: Booster ModI: 65.8% (132/200) HardI-HIF: 0.10 ± 0.21, significant
45–55% of HRres 3–4d/wk n = 102 (age F/M: compared to BL, Advice and
47 ± 8/52 ± 10) ModI-LowF
bmoderate-intensity–high frequency (Mod–HiF: • ModI-HIF: n = 107 HardI: 57.8% (115/199) ModI-HIF: 0.07 ± 0.20, significant
45–55% of HRres 5-7d/wk (age F/M: 48 ± 8/49 ± 8) compared to BL
bhard intensity-low frequency (Hi-LowF): 65–75% • HI-LowF: n = 97 BGD p b 0.03 HardI-LowF: 0.06 ± 0.20,
of HRres 3-4d/wk (age F/M: 47 ± 8/52 ± 10) significant compared to BL
bhard-intensity-High frequency (HardI-HiF): • M-LowF: n = 93 ModI-LowF: 0.03 ± 0.17, NS
65–75% of HRres 5-7d/wk (age F/M: 47 ± 8/49 ± 8)
group counselling sessions (11 over first 6 months • C: n = 93 (age F/M: Physician advice: 0.04 ± 0.25, NS
and 6 over 47 ± 8/51 ± 11)
remaining 18 months)
heart rate monitor and activity logs Groups comparable:
PA-advice: • Yes
• 90 min quarterly group session, print materials
(AHA booklets)
Simons-Morton Measurement and reinforcement of intervention Healthy inactive adults (35– FU: 24 months (91.4% Targets (moderate:
et al. (2001) at BL, 6, 12 and 24 months. 75yrs) (PAR), 77.6% (VO2max) 30 min 5d/wk, vigorous:
30 min 3d/wk), (F):
High Physician advice (A): Ass: n = 293 (age F/M: Booster Co: 25.7% (28/109) VO2max (ml/min),
52 ± 9/52 ± 10, F: 51%)
assessment and exercise goal setting Co: n = 289 (age F/M: Ass: 9.9% (12/121) BGD at FU, (F):
52 ± 9/52 ± 9, F: 45%)
health educator provided print material and A: n = 292 (age F/M: A: 14.3% (17/119) Ass vs. A 80.7 ml/min
available for related questions 51 ± 10/50 ± 10, F: 46%) (8.1–153.2, p = 0.02)
Assistance (Ass): Groups comparable: Co vs. Ass p = 0.005, Co vs. A 73.9 ml/min
other comparisons NS. (0.9–147, p = 0.046)
additional counselling by health educator Yes Men (M): Ass vs. Co -6.7 ml/min
(40 min) (−80.9–67.5, p = 0.99)
phone contact (one week after baseline) Co: 18.5% (28/151) Men (M):
monthly newsletters and step counter Ass: 29.9% (46/154) • Ass vs. A 49.5 (−44.6–143.6,
p = 0.66)
exercise log and weekly interactive mails A: 16.4% (24/146) • Co vs. A 15.3 (−81.1–111.7, p = 0.99)
Counselling (Co): Ass vs. A p = 0.02, other • Ass vs. Co −34.2(−130.1–61.6,
comparisons NS p = 0.92
• initially additional biweekly than monthly
telephone counselling (at agreed frequencies

F. Müller-Riemenschneider et al. / Preventive Medicine 47 (2008) 354–368


during second year)
• weekly behavioural skill classes
Lamb et al. Advice (C): Sedentary adults FU: 12-month Targets (30 min 5d/wk): VO2max, (L/min)):
(2002) (40–70yrs) (C: 72%, I: 73%)
Good • one standardised PA group session in the I: n = 129 Booster I: 31% (40/129) C: 2.47 ± 0.39
primary care setting (physiotherapist led)
• print materials C: n = 131 C: 26% (34/131) I: 2.49 ± 0.46
• Advice and walks (I): mean age: 51 ± 8 (F: 51.2%) BGD: 6% (−5–16.4) Change to BL:
• lay led walks program Groups comparable: C: +0.08
• print material about local walking routes n.a. I: + 0.05
BGD: 0.03 (− 0.02–0.08)
Pahor et al. Physical activity intervention (I): Healthy sedentary FU: 12 months Frequency of moderate PA EE in moderate intensity PA Change in 400 m-walk speed
(2006) adults (70–89yrs) (I: 94%, C: 73%) (per wk) (kcal/wk) favoured Intervention group
Good Individual 45 min training, initially centre I: n = 213 (F: 69%) Booster I: 5.1 ± 4.0 • I: 1001 ± 1084
(3/wk) than home-based sessions
Monthly phone calls and optional continuing C: n = 211 (F: 69%) C: 3.5 ± 5.2 • C: 710 ± 978
centre-based sessions
Successful Aging intervention (C): Overall mean age: 77 BGD: p = 0.002 • BGD: p = 0.002
Health education and attention Groups comparable:
26 weekly than monthly group sessions Yes
5–10 min light stretching during sessions
Phone reminders
Rovniak et al. Both groups:
(2005)
Fair Baseline orientation session, walking Sedentary women FU: 12 months Walking (min/wk)
prescription (no HR), interactive e-mail (20–54yrs) (79%)
exchange (activity logs)
High theoretical fidelity to SCT (I): I: n = 30 Booster I: 51.7 ± 76.9
Brief skills training (modelling session) C: n = 31 C: 24.2 ± 39.0
Free stopwatch (feedback, goal setting) Mean age: 40 ± 9 BGD: NS
List of local walking routes Groups comparable:
Low theoretical fidelity (C): n.a.
As described above
Albright et al. All Participants: Healthy low-income women FU: 12 months Targets (30 min 5d/wk): EE post class to FU:
(2005) (79%)

(continued on next page)


(continued

361
362
Table 4 (continued)
Author Year Intervention components Participants Baseline Follow-up Maintenance PA-maintenance Energy Expenditure Physical Fitness
Study Quality comparability of groups strategy
Fair 8 h skill-building classes and feedback on I1: n = 37 (32 ± 11) Booster I1: 35% I1: − 1014 kcal/wk
laboratory parameters
Mail (I1): I2: n = 35 (33 ± 9) I2: 49% I1: + 0.6 (315 kcal/wk)
print material (AHA) Groups comparable: BGD: NS BGD and BL to FU: p b 0.05
monthly newsletters and pedometers n.a.
Mail + phone (I2):

F. Müller-Riemenschneider et al. / Preventive Medicine 47 (2008) 354–368


additional stage specific phone counselling
(14 sessions, 10–15 min) and activity logs
Castro et al. Following a 12 months home-based Sedentary healthy FU: month 12 to 24 Maintenance (Hi: 3/wk, Lo:
(2001) intervention (High (Hi) and Low (Lo) intensity) adults (50–65yrs) (n.a.) 5/wk):
randomisation to
Fair Mail (I1): • I1: Hi (n = 39), Lo (n = 33) Booster At 24 months FU:
monthly educational mailings (12 in total) • I2: Hi (n = 36), Lo (n = 32) Hi: 2.19 (73%) (55/75)
including self assessment
Mail + phone (I2): • Overall F: 43% Lo: 2.85 (57%) (35/65)
15 additional monthly counselling calls Groups comparable: BGD and 12 to 24 months:
p b 0.05
• Yes
Jimmy and Primary care based PA scheme. Sedentary primary care FU: 14 months Targets (moderate: 30 min
Martin (2005) patients. (I1: 84%, I2: 80%) 7d/wk, vigorous: 20 min 3/wk):
Fair Feedback (I1): I1: n = 92 (50 ± 18, F: 58%) None I1: 47% (36/77)
stage matched feedback to level of PA I2: n = 69 (47 ± 17, F: 57%) I2: 47% (26/55)
Advice plus (I2): Groups comparable: BGD: NS
additional stage matched print materials • n.a.
physician advice to increase PA
optional (18€): one to one counselling
Nies and All groups: Healthy sedentary women FU: 12 months PA (min walked/wk) Walk test (min/1mile) Overall
Partridge (2006) (30–60yrs) (overall: 81%) improvement BGD: NS
Fair • Baseline assessment, walk test and I1: n = 90 None I1: 69.8 ± 52.8
recommendations to walk 90 min/wk
Phone counselling (I1): I2: n = 80 I2: 66.4 ± 55.6
16 phone calls (15 min each) over 24 weeks C: n = 83 C: 67.5 ± 66.6
related to exercise benefits and efficacy, goal
setting, social support and relapse prevention
Brief phone call (I2): Groups comparable: BGD: NS
16 phone calls (2–5 min each) over 24 wks n.a.
inquiring about PA behaviour
Video education control (C):
20 min video at BL about benefits of PA

PA: physical activity, FU: follow-up, BL: baseline, HRres: heart rate reserve, AHA: American Heart Association, M: male, F: female, BGD: between group difference, NS: not significant, TEE: total energy expenditure, EE: energy expenditure, PAR:
physical activity recall questionnaire, wk: week.
F. Müller-Riemenschneider et al. / Preventive Medicine 47 (2008) 354–368 363

– duration of the intervention/follow-up at least 12 months Partridge 2006) and 10 studies (Cox et al., 2003; Duncan et al., 2005;
– target population: healthy adults N18 years of age Simons-Morton et al., 2001; Lamb et al., 2002; Pahor et al., 2006;
– publication language: English or German Rovniak et al., 2005; Albright et al., 2005; Castro et al., 2001; Jimmy
and Martin 2005; Nies and Partridge 2006) compared different
Exclusion criteria:
intervention strategies but did not include a no-intervention control
– physical activity interventions targeting populations with diag- group (see Tables 2–4). Overall, only 7 studies could be rated as high-
nosed diseases or selected on the basis of risk factor clustering quality studies (Cox et al., 2003; Duncan et al., 2005; Elley et al., 2003;
– level of physical activity or physical fitness not stated as outcome in Petrella et al., 2003; Simons-Morton et al., 2001; Stewart et al., 2001;
title or abstract Marcus et al., 2007a) and 6 further studies as good-quality (Dubbert et
– size of intervention groups: b10 participants al., 2002; Hillsdon et al., 2002; Lamb et al., 2002; Resnicow et al., 2005;
Napolitano et al., 2006; Pahor et al., 2006). Common methodological
Data extraction of all included studies was performed by one
limitations were differences in baseline characteristics between
researcher and cross-checked by a second researcher. Included studies
intervention groups, high attrition rates or inadequate presentation
were classified according to their control group into three categories:
of investigated outcome parameters. Detailed descriptions of study
1. no-intervention control: studies which clearly describe a no-
characteristics, intervention components and outcomes are presented
intervention control or a usual care group. 2. minimal-intervention
in Tables 2–4. Included studies reported a variety of outcome
control: studies which include control groups consisting of the
measures and even when reporting the same outcome (e.g. proportion
provision of standard information-(material) or single and clearly
of participants meeting recommended physical activity targets) they
circumscribed intervention sessions. 3. alternative-intervention con-
used different definitions. Due to limited methodological quality or a
trol: studies which compare different intervention strategies which
lack of appropriately reported outcome measures, only 9 studies were
were more comprehensive than those described above. The metho-
found suitable for meta-analysis.
dological quality of included studies was evaluated using standardised
quality checklists employed by the German Institute for Documenta-
Long-term effectiveness of physical activity interventions
tion and Information. Depending on their risk of bias studies were
then rated according to the grading system developed by the Scottish
Compared to no-intervention or minimal-intervention control 10
Intercollegiate Guidelines Network Group (SIGN) as high (minimal
out of 15 studies found at least some evidence for positive
risk of bias), good (moderate risk of bias) and fair (high risk of bias)
intervention effects (see Tables 2–3). Especially 4 high and good-
methodological quality (Harbour and Miller 2001).
quality studies (Elley et al., 2003; Stewart et al., 2001; Petrella et al.,
We estimated pooled effects using Review Manager 4.2 separately
2003; Resnicow et al., 2005) provided strong evidence for increases in
for studies presenting appropriate dichotomous and continuous
physical activity behaviour. They reported increases in weekly energy
outcome measures. We included studies rated with good and high
expenditure of up to 975 kcal and physical fitness of up to 11%
methodological quality to minimise heterogeneity. Dichotomous
compared to control groups. Similarly, none of the studies reported
outcome measures reflected the proportion of participants meeting
unfavourable effects of the interventions on the level of physical
recommended targets of physical activity as specified in the study and
activity. However, physical activity behaviour varied substantially
were presented as odds ratios and 95% confidence intervals. To
between studies. Thus, the proportion of intervention participants
estimate the pooled effect and 95% confidence intervals of continuous
meeting recommended physical activity targets or adhering to
outcome measures we calculated the Standardised Mean Difference
physical activity prescriptions ranged between 4.6% and to 81%.
(SMD), using measures of physical fitness, energy expenditure and
Estimated pooled effects were consistent with these findings and
time spent for physical activity. All analyses were conducted using a
found significant increases in the proportion of participants meeting
random effects model. Sensitivity analyses were conducted by
recommended physical activity targets. Respective odds ratios to meet
including methodologically less rigorous studies, excluding outliers
physical activity targets were 3.31 (1.99–5.52) and 1.52 (1.07–2.14)
and using fixed effect models.
compared to no-intervention and minimal-intervention control, respec-
tively (see Fig. 2). Fig. 3 present pooled estimates of continuous outcome
Results measures, which also indicated significant increases in self-reported
energy expenditure and physical fitness associated with the interven-
Our literature search identified 5508 references and we assessed tions. Results of meta-analyses were robust to sensitivity analyses.
the full-text of 156 documents. Twenty-five studies met all selection Table 4 lists 10 identified studies which compared different
criteria and were included in this review (see Fig. 1). intervention strategies but did not include a no- or minimal-
The main reasons for exclusion of studies from this review were: intervention control group. Of these, four studies compared interven-
tion strategies to physician/health care professional advice (Duncan et
• non randomised study design
al., 2005; Lamb et al., 2002; Simons-Morton et al., 2001; Nies and
• participants selected on the basis of underlying disease or risk
Partridge 2006). Three of these studies tested more than one
factors
intervention strategy (Duncan et al., 2005; Simons-Morton et al.,
• not relevant to the research question
2001; Nies and Partridge 2006). Whereas none of the four studies
• outcome not level of physical activity but mediators or possible
reported consistently positive effects across intervention groups, all
surrogates of physical activity
except one fair-quality study (Nies and Partridge 2006) found some
Twenty-five RCTs investigated 39 intervention strategies to evidence for the long-term effectiveness of interventions compared to
promote physical activity behaviour. In the majority of studies walking advice only (Duncan et al., 2005; Lamb et al., 2002; Simons-Morton et
was the main physical activity component targeted by the interven- al., 2001) (see Table 4). Estimated pooled effects were consistent with
tion. Nine studies compared physical activity interventions to a no- these findings and found increases in the proportion of participants
intervention control group (Elley et al., 2003; Stewart et al., 2001; meeting recommended targets and physical fitness (see Figs. 2 and 3).
Hillsdon et al., 2002; Napolitano et al., 2006; Aittasalo et al., 2004; Pooled effects were robust to sensitivity analyses. The six remaining
Burke et al., 2003; Campbell et al., 2002; Eiben and Lissner, 2006; studies (Cox et al., 2003; Pahor et al., 2006; Rovniak et al., 2005;
Yancey et al., 2006), 6 studies included a minimal-intervention control Albright et al., 2005; Castro et al., 2001; Jimmy and Martin 2005)
(Petrella et al., 2003; Marcus et al., 2007a; Dubbert et al., 2002; compared a variety of different intervention strategies to promote
Resnicow et al., 2005; Bock et al., 2001; Lowther et al., 2002; Nies and physical activity behaviour. Of these, only 1 study reported clearly
364 F. Müller-Riemenschneider et al. / Preventive Medicine 47 (2008) 354–368

Fig. 2. Meta-analysis of the long-term effectiveness of physical activity interventions compared to no-intervention control, minimal-intervention control and health care staff advice
control, for studies reporting the proportion of participants meeting study specific physical activity targets.

positive intervention effects compared to an alternative intervention (Simons-Morton et al., 2001; Duncan et al., 2005; Cox et al., 2003;
(Pahor et al., 2006), whereas additional intervention effects in the Marcus et al., 2007a; Napolitano et al., 2006; Petrella et al., 2003;
remaining studies were less strong (see Table 4). Dubbert et al., 2002; Pahor et al., 2006). Although a moderate decline
of physical activity behaviour and physical fitness between early and
Sustainability of intervention effects late follow-up could frequently be observed, reported intervention
effects were mostly stable. Only one high-quality study (Petrella et al.,
Maximum study duration was 24 months. Nine methodologically 2003) reported a continuous increase of the intervention effect up to
reliable studies reported physical activity or cardiorespiratory fitness the 12 month follow-up. In contrast, two studies (Marcus et al., 2007a;
at different time points up to 24 months after intervention onset Napolitano et al., 2006) found no significant intervention effects at

Fig. 3. Meta-analysis of the long-term effectiveness of physical activity interventions compared to no-intervention control, for studies reporting continuous outcome measures.
F. Müller-Riemenschneider et al. / Preventive Medicine 47 (2008) 354–368 365

12 month follow-up after observing statistical significant effects at 3 fitness compared to minimal-intervention control, Duncan et al.
and 6 months, respectively. In one of these studies Napolitano et al. (2005) only found increases in physical fitness compared to physician
(2006) reported that this lack of a significant intervention effect was advice when prescribing high-intensity and high-frequency exercise.
due to an unexpected increase of physical activity in participants of In contrast, no significant intervention effects were found in two
the alternative-intervention control groups. studies investigating fitness assessments without exercise prescrip-
tion strategies (Aittasalo et al., 2004; Lowther et al., 2002).
Intervention effectiveness in specific population groups
Tailored vs. standard intervention messages
Three studies, including 2 high-quality studies, investigated the
effectiveness of physical activity interventions exclusively among Six studies specifically compared tailored to standard materials,
older participants (60+ years) compared to no-intervention or phone calls, internet strategies or feedback protocols (Dubbert et al.,
minimal-intervention control (Dubbert et al., 2002; Petrella et al., 2002; Napolitano et al., 2006; Marcus et al., 2007a; Nies and Partridge
2003; Stewart et al., 2001). One additional study compared a physical 2006; Bock et al., 2001; Rovniak et al., 2005). Results were conflicting
activity intervention designed for older participants to an alternative- and did not provide strong evidence for favourable outcomes of
intervention control group (Pahor et al., 2006). All studies found tailored compared to standard interventions. One study (Resnicow et
significant increases in physical activity associated with the interven- al., 2005) found evidence for the effectiveness of culturally tailored
tion. In addition, Elley et al. (2003) included participants aged 40– strategies, however.
79 years (mean age 57 years) and also observed significant increases in
physical activity behaviour. Effectiveness of maintenance strategies
One study of high and 3 of good methodological quality
investigated intervention effectiveness compared to no- and mini- Tables 2–4 show that studies frequently used methods to reinforce
mal-intervention control exclusively in middle aged adults (b60 years the initial intervention strategy. While various studies compared
of age) (Marcus et al., 2007a; Napolitano et al., 2006; Hillsdon et al., different maintenance strategies, none of the included studies
2002; Resnicow et al., 2005). Three studies reported at least some specifically investigated the effectiveness of interventions with and
evidence for intervention effectiveness while one found no favourable without reinforcement strategy.
intervention effect compared to control group. Elley et al. (2003) With regard to these maintenance strategies, studies were defined
reported clear evidence for intervention effectiveness compared to as:
no-intervention control, study participation was not restricted to
– those including no maintenance strategy to reinforce the initial
middle aged adults, however. Compared to health care staff/physician
intervention.
advice three studies reported positive intervention effects in some
– those repeating (in the same or similar way) the initial interven-
groups (Duncan et al., 2005; Simons-Morton et al., 2001; Lamb et al.,
tion during the follow-up period.
2002). Two studies also included older adults, however.
– those using booster strategies during the follow-up period (e.g.
Resnicow et al. (2005), Albright et al. (2005), Lowther et al. (2002)
mailed reminder, phone call, e-mail, internet, or group sessions
and Yancey et al. (2006) targeted disadvantaged population groups.
and workshops) to reinforce the initial intervention.
While two studies found evidence for intervention effectiveness
compared to no- and minimal-intervention control (Resnicow et al., Four studies did not use any maintenance strategies, although in
2005; Yancey et al., 2006), the remaining 2 studies did not find some cases baseline interventions lasted for up to 6 months (Burke et
evidence of favourable intervention effects. These studies were not al., 2003; Jimmy and Martin 2005; Napolitano et al., 2006; Nies and
suitable for meta-analysis. Partridge 2006). Of these, 3 studies reported no differences in physical
activity behaviour at the end of follow-up between intervention and
Effectiveness of specific Intervention components control group, and only one fair-quality study (Bock et al., 2001) found
favourable intervention effects.
In the majority of studies, interventions consisted of a combination Five studies evaluated 7 repeat intervention strategies to increase
of different intervention components and only infrequently investi- maintenance of physical activity behaviour. Of these, one high-quality
gated specific intervention components (see Table 2–4). study (Petrella et al., 2003) reported strong evidence of intervention
effectiveness. Four interventions found no (Aittasalo et al., 2004;
Initial intervention intensity Campbell et al., 2002; Lowther et al., 2002) and two interventions found
some evidence of effectiveness (Bock et al., 2001; Lowther et al., 2002).
One fair-quality study specifically compared different intensities of The 16 remaining studies assessed 23 booster strategies (see Tables
the initial intervention strategy (Burke et al., 2003) and observed no 2–4). Two high-quality studies provided strong evidence of effective-
differences in intervention effectiveness. Cox et al. (2003) compared ness compared to no-intervention control (Duncan et al., 2005; Elley
strategies of centre-based and home-based interventions during the et al., 2003; Stewart et al., 2001), and 4 high-quality studies observed
initial six months of the study. Both groups were home-based some evidence of effectiveness compared to minimal- or alternative-
thereafter. While exercise retention rates were higher in the centre- interventions (Cox et al., 2003; Simons-Morton et al., 2001; Marcus et
based group, increases in energy expenditure or physical fitness did al., 2007a; Duncan et al., 2005). Among the remaining studies only one
not differ significantly. found no positive effect compared to control group (Hillsdon et al.,
2002) and three studies found no effect compared to alternative
Written exercise prescription booster interventions (Hillsdon et al., 2002; Resnicow et al., 2005;
Rovniak et al., 2005).
Three high-quality studies provided exercise prescriptions in
addition to other intervention strategies, e.g. exercise counselling, Discussion
planning and activity logs (Duncan et al., 2005; Elley et al., 2003;
Petrella et al., 2003). Two studies included exercise testing to write Our systematic review provides evidence for the effectiveness of
prescriptions according to target heart rate zones (Petrella et al., 2003; physical activity interventions to increase levels of physical activity
Duncan et al., 2005). While Elley et al. (2003) and Petrella et al. (2003) over 12 to 24 months. Compared to no-intervention control and to
found significant increases in physical activity behaviour and physical minimal-intervention, increases in self-reported weekly energy
366 F. Müller-Riemenschneider et al. / Preventive Medicine 47 (2008) 354–368

expenditure 12 months after intervention onset were almost of a recently published review of walking interventions which
1000 kcal. These gains in weekly energy expenditure are consistent reported “tailoring” to be one general component of effective
with current physical activity recommendations to achieve mean- interventions (Ogilvie et al., 2007). The evidence is limited to tailored
ingful health benefits (Warburton et al., 2006a; Warburton et al., exercise prescription or tailored counselling, however. There was no
2006b; Haskell et al., 2007). There was also evidence of substantial conclusive evidence that individually tailored intervention materials,
increases in physical fitness associated with physical activity inter- phone calls, internet strategies or feedback protocols are more
ventions. Immediate increases in physical activity or fitness were effective than standard approaches to maintain physical activity
frequently followed by a moderate decline of the intervention effects behaviour in the long-term.
over time. With regard to specific population groups, the evidence for The majority of studies used strategies to maintain increased levels
the long-term effectiveness of physical activity interventions seemed of physical activity in the long-term. These generally consisted of
to be somewhat stronger for older compared to middle aged adults. In repeat interventions or booster strategies in the format of mail, phone,
addition, there was some evidence that current comprehensive internet, group sessions or combinations. Although none of the
interventions can be more effective in the long-term to increase identified studies specifically investigated the additional benefit of
levels of physical activity behaviour than physician or health care staff maintenance strategies compared to none or to repeat intervention,
advice only (Fig. 4). However, these results where less conclusive and the overall evidence seemed more consistent for studies using booster
warrant further investigation. strategies.
Compared to previous systematic reviews the evidence for the Uncertainty remains, however, in respect to the most effective and
long-term effectiveness of physical activity interventions has thereby efficient delivery mode of booster messages. Most commonly print
increased considerably (Hillsdon et al., 2005; Lawlor and Hanratty material or phone reminders were used. Only one study (Marcus et al.,
2001; Petrella and Lattanzio 2002; Eakin et al., 2004; Marcus et al., 2007b) directly compared these delivery modalities. The authors
2006; Foster et al., 2005). However, this evidence of effectiveness is reported superior long-term effectiveness associated with delivery of
mainly attributable to three high-quality studies that used written print material. However, because this study was published after the
prescription of exercise as an intervention component in addition to end of our search strategy it did not meet the inclusion criteria of the
physician advice, counselling and provision of information material. present review and was not formally assessed. Three included studies
Elley et al. (2003) and Petrella et al. (2003) reported the most attempted to identify the additional benefit of phone interventions
substantial increases in physical activity compared to no-intervention (phone plus mail) over mail only (Albright et al., 2005; Castro et al.,
or minimal-intervention control, and Duncan et al. (2005) observed 2001; Resnicow et al., 2005). In addition Simons-Morton et al. (2001)
increases in physical fitness compared to a physician advice group at used regular additional phone counselling compared to an interactive
24 months follow-up. mail exchange. Overall, results were inconclusive and did not provide
Considering tailored strategies, intervention effectiveness seems to sufficient evidence for an additional effect of phone reminders. The
improve when messages and materials are culturally adapted to the mode of phone contract was further investigated by Dubbert et al.
needs of specific population groups. Our study also confirms findings (2002) and Hillsdon et al. (2002). They observed that automated

Fig. 4. Meta-analysis of the long-term effectiveness of physical activity interventions compared to minimal-intervention control and health care staff advice control, for studies
reporting continuous outcome measures.
F. Müller-Riemenschneider et al. / Preventive Medicine 47 (2008) 354–368 367

phone calls achieved similar maintenance rates as personal phone sidering the delivery mode of booster strategies, it seems that new
calls and that advice giving during counselling calls was less effective information technologies can be equally effective as print materials,
than brief negotiation. These strategies, however, require confirma- possibly at considerably lower costs. Due to the increased complexity
tion in more rigorous studies. New information technologies have of effective interventions, additional cost-effectiveness analyses of
frequently been used in recent years as a mode for the delivery of physical activity interventions are warranted to assess the feasibility
booster messages. Only Marcus et al. (2007a) specifically investigated of these strategies on a broad population basis. Further, there is a need
the internet as a delivery tool and reported that tailored internet for methodologically rigorous studies to investigate the sustainability
strategies were as effective as tailored print strategies in promoting of current physical activity interventions.
physical activity behaviour. They didn't compare this strategy to a no-
intervention control group, though. Conversely, a recent review by Conflict of interest statement
Vandelanotte et al. (2007) did not identify strong evidence for the All authors declare that they have no conflict of interest.
long-term effectiveness of internet strategies. Although it might be
possible to deliver internet messages at considerably lower costs than Acknowledgments
traditional print or phone messages, it has to be kept in mind,
however, that certain important population groups will inevitably be We would like to thank Dr. Klaus Linde for his helpful advice and
missed using this approach. constructive criticism with regard to execution and presentation of
Numerous other intervention components are nowadays fre- this Systematic Review and Meta-Analysis.
quently used to increase promotion of physical activity. Studies
often use and combine approaches such as goal setting, problem
solving, self-monitoring, exercise planning and incentives as well as References
provision of activity logs, pedometers and heart rate monitors. While
these strategies have not been specifically assessed in any of the Aittasalo, M, Miilunpalo, S, Suni, J, 2004. The effectiveness of physical activity
counseling in a work-site setting. A randomized, controlled trial. Patient Educ.
included studies, they have been used in different combinations in Couns. 55, 193–202.
studies providing strongest evidence for intervention effectiveness. Albright, CL, Pruitt, L, Castro, C, Gonzalez, A, Woo, S, King, AC, 2005. Modifying physical
Considering public health consequences, presented results require activity in a multiethnic sample of low-income women: one-year results from the
IMPACT (Increasing Motivation for Physical ACTivity) project. Ann. Behav. Med. 30,
further discussion. While interventions were effective to increase 191–200.
physical activity behaviour, comprehensive and well conducted Bock, BC, Marcus, BH, Pinto, BM, Forsyth, LH, 2001. Maintenance of physical activity
studies were necessary to achieve these effects. This has also been following an individualized motivationally tailored intervention. Ann. Behav. Med.
23, 79–87.
reported in the above mentioned review by Ogilvie et al. (2007). Even
Brownson, RC, Boehmer, TK, Luke, DA, 2005. Declining rates of physical activity in the
in the study by Elley et al. (2003), which reported most substantial United States: what are the contributors? Annu. Rev. Public Health 26, 421–443.
intervention effects, only 15% of the participants were actually Burke, V, Giangiulio, N, Gillam, HF, Beilin, LJ, Houghton, S, 2003. Physical activity and
meeting recommended targets at the end of the study period. nutrition programs for couples: a randomized controlled trial1365. J. Clin.
Epidemiol. 56, 421–432.
While reported gains in physical activity can be considered a success, Camacho, TC, Roberts, RE, Lazarus, NB, Kaplan, GA, Cohen, RD, 1991. Physical activity and
it still leaves a substantial proportion of the population not being depression: evidence from the Alameda County study. Am. J. Epidemiol. 134, 220–231.
sufficiently active. Furthermore, participants in the majority of studies Campbell, MK, Tessaro, I, DeVellis, B, et al., 2002. Effects of a tailored health promotion
program for female blue-collar workers: health works for women. Prev. Med. 34,
volunteered and were motivated to increase their level of physical 313–323.
activity. Generalisability is therefore limited. Future strategies to Castro, CM, King, AC, Brassington, GS, 2001. Telephone versus mail interventions for
reach the whole population and to motivate those not responding at maintenance of physical activity in older adults. Health Psychol. 20, 438–444.
Cox, KL, Burke, V, Gorely, TJ, Beilin, LJ, Puddey, IB, 2003. Controlled comparison of
present will therefore be necessary to achieve greater public health retention and adherence in home- vs center-initiated exercise interventions in
effects. women ages 40–65 years: The S.W.E.A.T. Study (Sedentary Women Exercise
Regarding the results of this systematic review some limitations Adherence Trial). Prev. Med. 36, 17–29.
Department of Culture Media and Sport & Strategy Unit, 2002. Game Plan: A Strategy for
should be noted. Firstly, although we conducted a highly sensitive and Delivering Government's Sport and Physical Activity Agenda. Strategy Unit, London.
comprehensive search strategy and handsearch, publication bias Dubbert, PM, Cooper, KM, Kirchner, KA, Meydrech, EF, Bilbrew, D, 2002. Effects of nurse
cannot be entirely excluded. Secondly, the assessment of single counseling on walking for exercise in elderly primary care patients. J. Gerontol. A.
Biol. Sci. Med. Sci. 57, M733–M740.
intervention components was limited. This however, was due to a lack
Duncan, GE, Anton, SD, Sydeman, SJ, et al., 2005. Prescribing exercise at varied levels of
of appropriate studies investigating these issues and should be intensity and frequency: a randomized trial. Archives of internal medicine 165,
included in future research. Thirdly, pooled effect estimates were 2362–2369.
frequently based on a limited number of studies. However, this Eakin, EG, Brown, WJ, Marshall, AL, Mummery, K, Larsen, E, 2004. Physical activity
promotion in primary care: bridging the gap between research and practice. Am. J.
approach was chosen to increase reliability of pooled effect estimates. Prev. Med. 27, 297–303.
Fourthly, the selection criteria exclude intervention strategies which Eiben, G, Lissner, L, 2006. Health hunters—an intervention to prevent overweight and
might less frequently be investigated in RCTs, such as environmental obesity in young high-risk women. International journal of obesity 30, 691–696.
Elley, CR, Kerse, N, Arroll, B, Robinson, E, 2003. Effectiveness of counselling patients on
changes, interventions targeting active transport or community- physical activity in general practice: cluster randomised controlled trial. British
interventions. The fact that we were not able to identify randomised medical journal 326, 793.
studies using these approaches should be considered when interpret- Foster, C, Hillsdon, M, Thorogood, M, 2005. Interventions for promoting physical
activity. In: Foster, C, Hillsdon, M, Thorogood, M (Eds.), Interventions for promoting
ing the results of our systematic review. Finally, because studies had a physical activity Cochrane Database of Systematic Reviews : Reviews 2005 Issue 1.
maximum duration of 24 months, the sustainability of observed John Wiley & Sons , Ltd, Chichester, UK. 10 1002 /14651858 CD003180 pub2.
effects remains unknown. Especially when considering the observed Giovannucci, E, Ascherio, A, Rimm, EB, Colditz, GA, Stampfer, MJ, Willett, WC, 1995.
Physical activity, obesity, and risk for colon cancer and adenoma in men. Ann.
decline of physical activity over time it could be argued that increases Intern. Med. 122, 327–334.
in physical activity will be difficult to maintain as long as environ- Harbour, R, Miller, J, 2001. A new system for grading recommendations in evidence
mental factors are not targeted concurrently. based guidelines. BMJ 323, 334–336.
Haskell, WL, Lee, IM, Pate, RR, et al., 2007. Physical activity and public health: updated
In conclusion, there is evidence for the long-term effectiveness of
recommendation for adults from the American College of Sports Medicine and the
physical activity interventions in healthy adults. Increases in physical American Heart Association. Circulation 116, 1081–1093.
activity can meet recommended targets necessary for meaningful Hillsdon, M, Foster, C, Thorogood, M, 2005. Interventions for promoting physical
health benefits. Comprehensive and high-quality interventions, using activity. Cochrane Database Syst. Rev. CD003180.
Hillsdon, M, Thorogood, M, White, I, Foster, C, 2002. Advising people to take more
additional exercise prescriptions and booster strategies achieved most exercise is ineffective: a randomized controlled trial of physical activity promotion
substantial long-term increases in physical activity behaviour. Con- in primary care. Int. J. Epidemiol. 31, 808–815.
368 F. Müller-Riemenschneider et al. / Preventive Medicine 47 (2008) 354–368

Hu, FB, Leitzmann, MF, Stampfer, MJ, Colditz, GA, Willett, WC, Rimm, EB, 2001a. Physical Pahor, M, Blair, SN, Espeland, M, et al., 2006. Effects of a physical activity intervention on
activity and television watching in relation to risk for type 2 diabetes mellitus in measures of physical performance: Results of the lifestyle interventions and
men. Arch. Intern. Med. 161, 1542–1548. independence for Elders Pilot (LIFE-P) study. J. Gerontol. A. Biol. Sci. Med. Sci. 61,
Hu, FB, Manson, JE, Stampfer, MJ, et al., 2001b. Diet, lifestyle, and the risk of type 2 1157–1165.
diabetes mellitus in women. N. Engl. J. Med. 345, 790–797. Petrella, RJ, Koval, JJ, Cunningham, DA, Paterson, DH, 2003. Can primary care doctors
Hu, FB, Stampfer, MJ, Colditz, GA, et al., 2000. Physical activity and risk of stroke in prescribe exercise to improve fitness? The Step Test Exercise Prescription (STEP)
women. JAMA 283, 2961–2967. project. Am. J. Prev. Med. 24, 316–322.
Jimmy, G, Martin, BW, 2005. Implementation and effectiveness of a primary care based Petrella, RJ, Lattanzio, CN, 2002. Does counseling help patients get active: systematic
physical activity counselling scheme. Patient Educ. Couns. 56, 323–331. review of the literature (Structured abstract). Canadian Family Physician 48, 72–80.
Lamb, SE, Bartlett, HP, Ashley, A, Bird, W, 2002. Can lay-led walking programmes Resnicow, K, Jackson, A, Blissett, D, et al., 2005. Results of the healthy body healthy spirit
increase physical activity in middle aged adults? A randomised controlled trial. trial. Health Psychol. 24, 339–348.
Journal of epidemiology and community health 56, 246–252. Rhodes, RE, Martin, AD, Taunton, JE, Rhodes, EC, Donnelly, M, Elliot, J, 1999. Factors
Lampert, T, Mensink, GB, Ziese, T, 2005. [Sport and health among adults in associated with exercise adherence among older adults. An individual perspective.
Germany]. Bundesgesundheitsblatt Gesundheitsforschung Gesundheitsschutz 48, Sports Med. 28, 397–411.
1357–1364. Robison, JI, Rogers, MA, 1994. Adherence to exercise programmes. Recommendations.
Lawlor, DA, Hanratty, B, 2001. The effect of physical activity advice given in routine Sports Med. 17, 39–52.
primary care consultations: a systematic review (Structured abstract). Journal of Rovniak, LS, Hovell, MF, Wojcik, JR, Winett, RA, Martinez-Donate, AP, 2005. Enhancing
Public Health Medicine 23, 219–226. theoretical fidelity: an e-mail-based walking program demonstration 192. Amer-
Lee, IM, Hsieh, CC, Paffenbarger Jr, RS, 1995. Exercise intensity and longevity in men. The ican Journal of Health Promotion 20 (2), 85–95.
Harvard Alumni Health Study. JAMA 273, 1179–1184. Sesso, HD, Paffenbarger Jr., RS, Lee, IM, 2000. Physical activity and coronary heart
Lowther, M, Mutrie, N, Scott, EM, 2002. Promoting physical activity in a socially and disease in men: the Harvard Alumni Health study. Circulation 102, 975–980.
economically deprived community: a 12 month randomized control trial of fitness Simons-Morton, DG, Blair, SN, King, AC, et al., 2001. Effects of physical activity
assessment and exercise consultation. J. Sports Sci. 20, 577–588. counseling in primary care: the Activity Counseling Trial: a randomized controlled
Marcus, BH, Lewis, BA, Williams, DM, et al., 2007a. A comparison of internet and print- trial. JAMA 286, 677–687.
based physical activity interventions. Arch. Intern. Med. 167, 944–949. Stewart, AL, Verboncoeur, CJ, McLellan, BY, et al., 2001. Physical activity outcomes of
Marcus, BH, Napolitano, MA, King, AC, et al., 2007b. Telephone versus print delivery of CHAMPS II: a physical activity promotion program for older adults. J. Gerontol. A.
an individualized motivationally tailored physical activity intervention: Project Biol. Sci. Med. Sci. 56, M465–M470.
STRIDE. Health Psychol. 26, 401–409. US Department of Health and Human Services, 2000. Healthy People 2010. With
Marcus, BH, Williams, DM, Dubbert, PM, et al., 2006. Physical activity intervention Understanding and Improving Health and Objectives for Improving Health. U.S.
studies: what we know and what we need to know: a scientific statement from the Government Printing Office, Washington,DC. 2 vols.
American Heart Association Council on Nutrition, Physical Activity, and Metabolism Vandelanotte, C, Spathonis, KM, Eakin, EG, Owen, N, 2007. Website-delivered physical
(Subcommittee on Physical Activity); Council on Cardiovascular Disease in the activity interventions a review of the literature. Am. J. Prev. Med. 33, 54–64.
Young; and the Interdisciplinary Working Group on Quality of Care and Outcomes Wannamethee, SG, Shaper, AG, 1999. Physical activity and the prevention of stroke. J.
Research. Circulation 114, 2739–2752. Cardiovasc. Risk 6, 213–216.
Napolitano, MA, Whiteley, JA, Papandonatos, G, et al., 2006. Outcomes from the Wannamethee, SG, Shaper, AG, 2001. Physical activity in the prevention of cardiovas-
women's wellness project: a community-focused physical activity trial for women. cular disease: an epidemiological perspective. Sports Med. 31, 101–114.
Prev. Med. 43, 447–453. Warburton, DE, Nicol, CW, Bredin, SS, 2006a. Health benefits of physical activity: the
Nies, MA, Partridge, T, 2006. Comparison of 3 interventions to increase walking in evidence. CMAJ 174, 801–809.
sedentary women. American journal of health behavior 30, 339–352. Warburton, DE, Nicol, CW, Bredin, SS, 2006b. Prescribing exercise as preventive therapy.
Ogilvie, D, Foster, CE, Rothnie, H, et al., 2007. Interventions to promote walking: CMAJ 174, 961–974.
systematic review. BMJ 334, 1204. Yancey, AK, McCarthy, WJ, Harrison, GG, Wong, WK, Siegel, JM, Leslie, J, 2006.
Owen, N, Bauman, A, 1992. The descriptive epidemiology of a sedentary lifestyle in adult Challenges in improving fitness: results of a community-based, randomized,
Australians. Int. J. Epidemiol. 21, 305–310. controlled lifestyle change intervention. Journal of women's health 15, 412–429.

You might also like