Web-based physical activity interventions: a systematic review and meta-analysis of randomized controlled trials

p u b l i c h e a l t h 1 5 2 ( 2 0 1 7 ) 3 6 e4 6

Available online at www.sciencedirect.com

Public Health journal homepage: www.elsevier.com/puhe

Review Paper

Web-based physical activity interventions: a systematic review and meta-analysis of randomized controlled trials Leila Jahangiry a,b, Mahdieh Abbasalizad Farhangi c,d,e,*, Sakineh Shab-Bidar f, Fatemeh Rezaei g, T. Pashaei h a

Health Education and Health Promotion Department, Faculty of Health, Tabriz University of Medical Sciences, Azadi Street, Golgasht Street, Tabriz, Iran b Tabriz Health Services Management Research Center, Tabriz University of Medical Sciences, Tabriz, Iran c Drug Applied Research Center, Nutrition Research Center, Tabriz University of Medical Sciences, Tabriz, Iran d Nutrition Research Center, Tabriz University of Medical Sciences, Tabriz, Iran e Department of Community Nutrition, Faculty of Nutrition, Tabriz University of Medical Sciences, Tabriz, Iran f Department of Community Nutrition, School of Nutritional Sciences and Dietetics, Tehran University of Medical Sciences, Tehran, Iran g Department of Social Medicine, School of Medicine, Jahrom University of Medical Sciences, Jahrom, Iran h Health Education and Health Promotion Department, School of Public Health, Kurdistan University of Medical Sciences, Sanandaj, Iran

article info

abstract

Article history:

Objectives: It was estimated that approximately 60% of the world's population is classified as

Received 20 October 2016

inactive or insufficiently active. This meta-analysis investigated the effect of web-based

Received in revised form

interventions on different types of physical activity (PA) measurements in general popu-

11 April 2017

lation and potential moderating variables.

Accepted 2 June 2017

Study design: PubMed, CINAHL, EBSCOhost, PsycINFO, Scopus, Ovid, and ScienceDirect literature searches were conducted to identify studies investigating the effect of web-based interventions on PA.

Keywords:

Methods: Randomized controlled trials on PA changes reported in moderate to vigorous

Physical activity

intensity, walking, and step count in the intervention group in comparison with the control

Web-based intervention

group were pooled with a fixed-effects model separately.

Walking

Results: A total of 22 studies comprising 16,476 and 14,475 subjects in intervention and control

Meta-analyses

groups respectively were included. Web-based interventions had positive and significant effect on increasing PA. Of 14 trials reporting moderate to vigorous physical activity (MVPA), five showed a significant increase in the MVPA level after the intervention. There was significant heterogeneity between studies (P < 0.001 and I2 ¼ 67.8%). Of six trials that reported the number of steps by using the pedometer, three showed a significant increase for the step counts in intervention groups (P < 0.001 and I2 ¼ 93.3%), of 14 trials assessed PA level by

* Corresponding author. Health Education and Health Promotion Department, Faculty of Health, Tabriz University of Medical Sciences, Azadi Street, Golgasht Street, Tabriz, Iran. E-mail addresses: [email protected] (L. Jahangiry), [email protected] (M.A. Farhangi), [email protected] (S. Shab-Bidar), [email protected] (F. Rezaei), [email protected] (T. Pashaei). http://dx.doi.org/10.1016/j.puhe.2017.06.005 0033-3506/© 2017 The Royal Society for Public Health. Published by Elsevier Ltd. All rights reserved.

p u b l i c h e a l t h 1 5 2 ( 2 0 1 7 ) 3 6 e4 6

37

reporting walking minutes per week, four studies showed a significant increase in walking minutes. There was significant heterogeneity between studies (P < 0.001, I2 ¼ 68.1%). Overall, the effect of web-based interventions seemed to be influenced by the characteristics of mean age of participants, trial duration, and study quality (P < 0.05). Conclusion: The web-based PA interventions had a positive significant effect on increasing all the three types of PA among the general population. However, the effects appear to depend on the design of the study, age, and duration of studies. © 2017 The Royal Society for Public Health. Published by Elsevier Ltd. All rights reserved.

Introduction It was estimated that approximately 60% of the world's population is classified as inactive or insufficiently active.1 Physical activity (PA) has been associated with cardiovascular diseases and wide range of other chronic diseases including diabetes, cancer (colon and breast), hypertension, bone and joint diseases (osteoporosis and osteoarthritis), and depression.2 As a consequence, promoting improvements in PA behavior is of great importance to public health. To reach the large numbers of the inactive population at relatively low-cost, efficacious delivery of behavior change programs are needed.3 New communications by information and communication technologies such as the Internet and web technologies are the unique opportunity to formulate and implement effective strategies to change unhealthy behavior, e.g. PA in the general population.4 Several studies have investigated the effectiveness of webbased interventions to provide health behavior change among adults.5e7 T Webb et al.7 compared the effectiveness of webbased versus noneweb-based behavioral changes, and another study used both nutrition and PA of lifestyle dimensions to explore and evaluate web-based channels for changing the behavior.8 A meta-analysis study by Davies et al.9 demonstrated that the web-based interventions are effective in producing small and significant increases in the PA level. They reported overall mean effect size (d ¼ 0.14) of Internet-delivered interventions. However, the effect size cannot be reflected as the clinically meaningful change of PA levels. Although the accurate measurement of PA in the webbased intervention is very challenging due to the complex and multi-dimensional nature of the behavior, there are objective measurement approaches to quantify PA and type of exercise including; moderate to vigorous physical activity (MVPA), and step counts by pedometers or walking measures per week.10 This meta-analysis investigated the effect of web-based PA interventions based on the three main reported objective measurements (MVPA, step count by pedometer, and walking time) among the general population and potential moderating variables.

Materials and methods This systematic review was performed using a pre-specified protocol with the aim of reviewing the evidence that shows

the effectiveness of web-based interventions for improving the PA behavior. The statement of Preferred Reporting Items for Systematic Reviews and Meta-Analyses has been used to report the results of this study.11

Search strategy and study selection The data sources were PubMed, CINAHL, EBSCO, PsycINFO, Scopus, Ovid, and ScienceDirect from 2000 to 15 September 2015 including articles in English. Publications with the following search words in the titles, abstracts, or keywords of the original studies were included: Internet, computer, phone, smartphone, web-based, tele-health, social media, e-Health, Web, online, email, electronic mail, Internet, social networking combined with PA, exercise, leisure activity.

Selection criteria The published articles were searched for this meta-analysis. The following criteria were used for a publication to be included in the meta-analysis: a) published clinical trials on PA; and b) targeting adults aged over 18 years. The exclusion criteria were: a) noneinternet-based studies; b) systematic review, meta-analysis, qualitative, quasiexperimental, cohort, and cross-sectional study design; c) studies with combined interventions for lifestyle change or other interventions, trials without control groups; d) repeated studies; e) the results of the trials which were published in more than one article; and f) telephone-based interventions, studies on populations with chronic diseases diabetes, asthma, cancer, heart failure, multiple sclerosis.

Data extraction and quality assessment Two independent researchers (FR and TP) reviewed all identified studies for the relevant titles and abstracts, and then retrieved the full-texts for those that were potentially relevant. A screening form was used to select eligible articles. The quality control of the articles was performed independently by two authors (FR and TP) and any disagreement solved by the third author (LJ). Characteristics of studies were abstracted including the first author, publication year, country of origin, trial duration, number of participants in control and intervention groups, and measurements. Participants characteristics consisted of sex, mean age, baseline, and after intervention level of PA. Studies with two or more independent strata were considered

38

p u b l i c h e a l t h 1 5 2 ( 2 0 1 7 ) 3 6 e4 6

as separate studies. When the study reported two or more outcome measurements, separate estimates of each outcome was extracted for that study. The quality of studies was assessed by Jadad scales that assign scores for reported randomization, blinding, and withdrawals.12

the age of participants and mean of PA. Publication bias was tested using funnel plot and Egger's regression asymmetry test.14 All of the analyses were performed using STATA version 12.0 (Stata Corporation, College Station, TX, USA), and P < 0.05 was considered as level of significance.

Data synthesis and statistical analysis

Results Mean differences and standard deviation (SD) of PA level in baseline and follow-up measure of the study in intervention and control groups were considered. The studies that used more than one follow-up measure were considered as more than one study and if a confidence interval (CI) was reported in place of SD, we converted it to SD. Existence of heterogeneity was tested by Cochran's Q-test at P < 0.1 level of significance. The I2 test was also used to calculate percentage of heterogeneity.13 A fixed-effects model was used for estimating pooled effect sizes. To investigate the sources of heterogeneity, predefined subgroup analyses were performed using trial duration, baseline PA level, and the age of participants. Metaregression was performed using trial duration, baseline PA,

Search results and study selection The flowchart of the selection process in meta-analysis is shown in Fig. 1. Of 5831 studies, 4931 articles were excluded because of duplicated records. In all, 900 articles were included for the title and abstract screening. After screening titles and abstracts for duplicates, 795 articles were excluded because they did not meet the inclusion criteria. The remaining 105 articles were chosen for assessing full-text eligibility and in this step, 76 were excluded as they used web-based interventions for adolescents or children or used PA intervention with other lifestyle interventions. Then, 29

lack Fig. 1 e Study selection diagram.

Table 1 e Summary of studies. Reference 17 20 23

18 16 30 35 15 31 36 28 27 34 22 21 26 32 25 24 37 19

Wijsman C A Vandelanotte C Soetens K C De Cocker K Watson A. Cadmus-Bertram L A Maher C Cavallo, D. N Dunton, G. F Hansen, A. W Carr, L. J Roesch, S. C Slootmaker, S. M Compernolle, S Spittaels, H Spittaels, H Sriramatr, S Friederichs, S. A Thorsteinsen, K. van Stralen, M. M Peels, D. A. Vroege, D. P

Year

2013 2012 2014 2012 2012 2015 2015 2012 2008 2012 2013 2010 2009 2015 2007 2007 2014 2015 2014 2011 2013 2014

Country

Netherland Australia Australia Belgium US US Australia US US Denmark US US US Belgium Belgium Belgium Thailand Netherland Norway Netherland Netherland Netherland

Sample size Intervention

Control

119 261 258 45 35 25 51 67 84 6055 32 429 51 132 174, 175 173, 129 55 987, 1129 14 652, 733 432 119

116 542 262 47 35 26 59 67 71 4509 34 413 51 123 177 132 55 1049 9 586 425 112

Duration (week)

Baseline measures

Interactive design

Recruitment

Tools

Outcomes report

Jadad score

12 1, 4 1, 4 12 3, 3, 3, 3 16 8, 20 12 4, 8.12 6 3, 6 6, 12 3, 8 4, 12 6 3 3 3, 6, 12 5, 9, 12 12 12 12

Face Online Online Face Online Face Online Online Online Face Online Face Face Face Online Online Face Online Online Online Online Online

Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes

Online Online Online Online Online Face Online Online Phone Online Online Face Face Face Online Online Face Online Online Online Online Online

Accelerometer e e Pedometer Pedometer Pedometer Pedometer e

MVPA Walk (min) PA (week/min), walk Walk, step Step Step Total PA, walk Total PA MVPA, walk PA (week/min), (week/min) Walk MVPA Step, total PA, walking MVPA, total PA MVPA Step MVPA, total PA PA week/min PA week/min Pa week/min MVPA

4 2 1 3 4 1 3 2 5 3 2 1 2 4 1 1 2 2 3 2 2 3

e e Pedometer e e Pedometer

e

p u b l i c h e a l t h 1 5 2 ( 2 0 1 7 ) 3 6 e4 6

33

Author

Abbreviations: MVPA, moderate to vigorous physical activity; PA, physical activity.

39

40

p u b l i c h e a l t h 1 5 2 ( 2 0 1 7 ) 3 6 e4 6

studies were retained for quantitative analyses. Of those, seven studies with no exact value in PA, lack of baseline value for PA for each group were excluded. Finally, 2215e36 studies were included for meta-analysis, which were all randomized controlled trials.

Study characteristics The characteristics of the included studies and participants are presented in Table 1. Publication year of the studies were from 2000 to 2015; and of the 22 studies, 20 were conducted on both males and females,17e36 and two studies were conducted only on females.15,16 Sample size of studies varied from 14 to 6055 with a total sample size of 16,476 and 14,475 subjects in intervention and control groups, respectively. All trials involved parallel designs, but only three of those were double blinded,15,25,30 and 14 did not clearly mention the blinding method.16,20e24,26e29,32,33,35,36 The mean ages of participants in intervention and control groups were 30.8 years (19e64.7 years), and 37.5 years (19e64.9 years), respectively. Trial duration was from 1 to 20 weeks with a mean of 7 weeks. All studies used the interactive or tailored design to deliver the interventions. Theories and models used to guide interventions including: theory of planned behavior and stage of change,15,20 social cognitive theory,26,36 self-monitoring,16,17,24,30

social support,35 self-determination theory and motivational interviewing,32 and self-efficacy.24,28 Quality scores were in the range of 1e5. The quality scores of nine studies were greater than and others were less than three,15,17e19,25,30,31,33,34 three.16,20e24,26e28,32,35e37 The studies were conducted in the United States (n ¼ 7),15,16,18,27,28,35,36 Netherlands (n ¼ 5),17,19,24,29,32 Australia (n ¼ 3),22,23,30 Belgium (n ¼ 3),21,33,34 Denmark (n ¼ 1),31 Norway (n ¼ 1),25 Thailand (n ¼ 1),26 and Canada (n ¼ 1).20 Fourteen studies have reported results based on measuring MVPA,15,17,21,22,27,32 six have reported number of steps that a person taken during PA and used a pedometer,16,18,26,30,33,34 14 have studies reported walking minutes per day or week,15,20,21,23,28,30,33,34 and one study has reported results based on measuring total kilo calorie; energy expeditor.35

Meta-analysis Fig. 2 provides a forest plot of 14 trials with mean differences (MDs) and their 95% CIs that are showing the effect of webbased interventions on MVPA. Of 14 trials which reported MVPA measure, five showed a significant increase in the MVPA level after intervention (MD ¼ 13.42 and P < 0.001). There was a significant heterogeneity among studies (test for heterogeneity; P < 0.001 and I2 ¼ 67.8%).

[15] [21] [27] [22] [22] [22] [32] [32] [32] [17] [15] [15] [27] [32] [32]

Fig. 2 e Pooled effect size of web-based physical activity interventions based on moderate to vigorous physical activity measurement. WMD, weighted mean difference.

p u b l i c h e a l t h 1 5 2 ( 2 0 1 7 ) 3 6 e4 6

Fig. 3 shows a forest plot with MDs and their 95% CIs. Of six trials which reported the number of steps using the pedometer, three showed a significant increase of the step counts in the intervention groups, but one was declined during the study (MD ¼ 2185 and P < 0.001) with a significant heterogeneity (P < 0.001 and I2 ¼ 93.3%). Of 22 trials included in the meta-analysis, 14 assessed PA levels and reported walking minutes per week. Of these, four studies showed a significant increase in the walking minutes after interventions, but one study showed a significant decrease in the walking minutes in the control group. The forest plot with MDs in post-trial for walking minutes of the intervention and control groups are shown in Fig. 4. There were a significant heterogeneity among studies (P < 0.001, I2 ¼ 68.1%).

Subgroup and meta-regression analyses Table 2 shows the results of the MVPA level according to the subgroup and the meta-regression analyses of predefined criteria to explore the sources of heterogeneity. Overall, the effect of web-based interventions seems to be influenced by mean age of participants (<45 years), trial duration (6 weeks), and study quality. Subgroup analyses showed that MVPA level increased more significantly after interventions in participants aged less than 45 years (51.25 [95% CI ¼ 29.1, 72.6]) compared with those 45 years (13.4 [95% CI ¼ 12.9, 13.8]). However, the increase of MVPA level in both intervention and control groups, was significant. MVPA level increased when

41

trial duration was less than 6 weeks (55.3) compared with trial duration of more than 6 weeks (13.4). There was a greater increase in the MVPA level in participants with baseline MVPA level of >300 (38 [95% CI ¼ 16.6, 59.5]). Also, there was a greater increase in the MVPA level in studies with quality score of more than 2 (50.3 [95% CI ¼ 20.6, 80]) compared with those less than 2 (1.6 [95% CI ¼ 56 to 59.5]). Table 3 shows stratified and meta-regression analyses of the web-based PA interventions for measuring step counts. Subgroup analyses showed that step counts increased more significantly after the web-based interventions in participants aged less than 40 years (3666.5 [95% CI ¼ 3096.5, 4236.5]) compared with those  40 years (422.67 [95% CI ¼ 199, 1044.4]). Step counts increased when the trial duration was 6 weeks (3074.3 [95% CI ¼ 2551.9, 3542.7]) compared with trial duration >6 weeks (22.3 [95% CI ¼ 815, 770.4]). There was a greater increase in the step counts in participants with baseline step counts 7000 (2344.9 [95% CI ¼ 1905.3, 2784]). The univariate meta-regression analysis indicated that mean age, trial duration, baseline step, interactive design, and study quality were not the sources of heterogeneity. Table 4 shows the stratified and meta-regression analyses of web-based PA interventions for walking minutes per week. Subgroup analyses showed that walking minutes increased more significantly after web-based interventions in participants aged less than 40 years (0.22.1 [95% CI ¼ 3.5, 40.7]) compared with those 40 years (0.34 [95% CI ¼ 0.23, 0.44]). Walking minutes increased when trial duration was 6 weeks (0.36 [95% CI ¼ 0.22, 0.5]) compared with trial duration of less

[33] [34] [34] [26] [26] [16]

Fig. 3 e Pooled effect size of web-based physical activity interventions based on step measurement. WMD, weighted mean difference.

42

p u b l i c h e a l t h 1 5 2 ( 2 0 1 7 ) 3 6 e4 6

[30] [15] [34] [28] [30] [28] [15] [30] [28] [15] [20] [23] [34] [20] [23]

Fig. 4 e Pooled effect size of web-based physical activity interventions based on walking minutes per day or week.

than 6 weeks (0.31 [95% CI ¼ 0.16, 0.46]). There was a greater increase in the step count in participants with baseline walking minutes >100 (18.89 (95% CI ¼ 7.5, 3.2). The univariate meta-regression analysis suggested that baseline walking minutes was the source of heterogeneity.

Publication bias For studies that reported MVPA, and the Egger's regression analysis showed asymmetric of the plot (P < 0.001). The Egger's regression analysis did not show asymmetric of the plot (P ¼ 0.164) in studies that reported walking minutes and step counts, respectively (Fig. 5).

Discussion In this meta-analysis of randomized controlled trials on webbased PA interventions, including 22 trials with 21,316 participants between 2000 and 2015, we found that web-based interventions increases PA by 13.4 for MVPA, 2185 for step counts, 0.17 min for walking per week. Significant and positive effects of web-based interventions were found in all three types of measurements for PA. These findings are in consistent with previous reviews that have recommended on capability of web-based programs to improve PA.9,38 Some of

the heterogeneity among studies accounted for by variation in the mean age, baseline PA level, trial duration, and study quality. One reason for the effect of using Internet on promoting PA might be that PA programs were based on various topics such as self-monitoring, goal setting, and receiving feedback to PA.17 Using the interactive designed technologies for e-Health interventions is a proper way to deliver evidence-based interventions.39,40 It is known that self-monitoring is one of the important aspects of interactive interventions.41 In the reviewed studies, self-monitoring has been reported most frequently.16,17,25,27,30,35,36 The self-monitoring of PA refers to the observing and recording of exercise patterns. Using the selfmonitoring as a tool has encouraged the people to increase the targeted behavior.42 Recording daily activity can be used as part of a self-monitoring program to increase PA.43 A few studies have reported the detailed information for self-monitoring of PA and criteria for how they measured self-monitoring or how they defined recordings is not clear. It seems that using the technology-based devices for interventions has provided an objective validation of the self-reported behaviors.44 Six studies in the meta-analysis that used pedometerwalking interventions16e18,30,33,34 showed a modest improvement in daily step counts by 2185.1.45 Consistent with our study, Richardson et al.,46 in a pedometer-based meta-analysis of PA have shown that the average daily step count varied

43

p u b l i c h e a l t h 1 5 2 ( 2 0 1 7 ) 3 6 e4 6

Fig. 5 e Funnel plots. WMD, weighted mean difference.

Table 2 e Stratified analysis of web-based intervention on physical activity for MVPA. Variables

Number of data

Total 4 Mean age 45 years 9 <45 years 4 Trial duration >6 weeks 6 6 weeks 8 Baseline MVPA per week >300 8 300 6 Interactive design Yes 8 No 6 Study quality >2 9 2 5

Mean differences (95% CI)

I2%

P-value for heterogeneity

Significance*

13.4 (12.9e13.8)

67.8

<0.0001

<0.0001

e

13.4 (12.9e13.8) 51.25 (29.1e72.6)

68.8 54.4

0.041 0.015

<0.0001 <0.0001

0.006 Reference

13.4 (12.9e13.9) 55.53 (31.5e79.6)

48 63.2

0.087 0.008

<0.0001 <0.0001

0.006 Reference

38 (16.6e59.5) 13.4 (12.9e13.8)

67.5 64.5

0.003 0.015

0.001 <0.0001

0.134 Reference

13.4 (12.9e13.8) 55.4 (30.7e80.1)

43.8 65.6

0.113 0.005

<0.0001 <0.0001

0.109 Reference

50.3 (20.6e80) 1.6 (56 to 59.5)

74.6 55

<0.0001 0.064

0.001 0.956

0.018 Reference

Meta-regression**

Abbreviations: CI, confidence interval; MVPA, moderate to vigorous physical activity. *P-values for significance of mean difference in physical activity between intervention and control groups. **P-values for difference in physical activity change across strata.

44

p u b l i c h e a l t h 1 5 2 ( 2 0 1 7 ) 3 6 e4 6

Table 3 e Stratified analysis of web-based intervention on physical activity for step measurement. Variables

Mean differences (95% CI)

I2%

P for heterogeneity

Significance*

9

2185.1 (1764.9e2605.2)

93.3

<0.0001

0.146

4 2

422.67 (199 to 1044.4) 3666.5 (3096.5e4236.5)

82.7 0

0.001 <0.0001

0.183 <0.0001

0.251 Reference

3 3

22.3 (815 to 770.4) 3074.3 (2551.9e3542.7)

85.9 89.5

0.001 <0.0001

0.95 <0.0001

0.277 Reference

4 3

2344.9 (1905.3e2784.7) 507 (917.4 to 1931.4)

94.2 0

<0.0001 <0.0001

<0.0001 <0.0001

0.194 Reference

5 2

2344.9 (1905.3e2784.6) 507 (917.4 to 1931.4)

94.2 0

<0.0001 <0.0001

<0.0001 0.485

0.194 Reference

6 3

402.8 (288.1 to 1093.8) 3230.4 (2701.2e3759.6)

88.5 88.1

<0.0001 <0.0001

0.253 <0.0001

0.879 Reference

Number of data

Total Mean age 40 <40 Trial duration >6 weeks 6 weeks Baseline Step count 7000 <7000 Interactive design Yes No Study quality >2 2

Meta-regression** e

Abbreviation: CI, confidence interval. *P-values for significance of mean difference in MVPA of physical activity between intervention and control groups. **P-values for difference in MVPA of physical activity across strata.

from less than 2000 steps per day to more than 4000 steps per day. Walking 2000 steps for an average person is approximately equal to one mile. As an average, walking 1-mile at a 20- and 15- minute pace a day is in the range of the recommended level of PA guidelines. This increase in PA can be expected to result in health benefits.47 It should be noted that in web-based interventions, integrating pedometer-based PA program in an interactive web-based and tailored interventions will result in more accurate feedback with higher personal relevance.34 However, a pedometer-based PA intervention through the Internet appears to increase both pedometer-based and self-reported PA level during the interventions.48

This study showed that web-based PA programs have beneficial effects in those younger than 45 years old. These results suggest that although web-based interventions produce positive and significant effects on PA, a significant greater effect was observed among people <45 years old. One reason is that the users of internet are different from the general population. The most ardent internet users are the young people and the well-educated middle-aged people, who take most interest in online health resources and services.49,50 Given the potential breadth of delivery, the public health impact of producing small changes in PA across a population has the potential for positive changes . Several limitations must be addressed in this metaanalysis. First, the current meta-analysis could identify 22

Table 4 e Stratified analysis of web-based intervention on physical activity for walking. Variables

Number of data

Total 14 Mean age 40 10 <40 4 Trial duration >6 7 6 7 Baseline walking (minutes) >100 9 100 5 Interactive web design Yes 8 No 6 Study quality >2 7 2 7

Mean differences (95% CI)

I2%

P-value for heterogeneity

Significance*

0.173 (0.76e0.27)

68.1

<0.0001

0.001

0.34 (0.23e0.44) 22.1 (3.5e40.7)

80.7 55.2

0.017 0.001

<0.0001 0.020

0.091 Reference

0.31 (0.16e0.46) 0.36 (0.22e0.5)

77.1 58.3

<0.0001 0.025

<0.0001 <0.0001

0.941 Reference

18.89 (7.5e30.22) 0.17 (0.074e0.27)

62.3 55.9

0.007 0.059

0.001 0.001

0.016 Reference

14.2 (6e22.4) 0.33 (0.23e0.44)

73 0

0.001 0.567

0.001 <0.0001

0.170 Reference

12.7 (4.5e20.9) 0.33 (0.23e0.44)

39.9 72.8

0.125 0.001

0.002 <0.0001

0.199 Reference

Abbreviation: CI, confidence interval. * P-values for significance of mean difference for walking between intervention and control groups. ** P-values for difference in walking change across strata.

Meta-regression** e

p u b l i c h e a l t h 1 5 2 ( 2 0 1 7 ) 3 6 e4 6

eligible studies with duration of 1e20 weeks. Therefore, this analysis might not necessarily provided sufficient evidence for the long-term effect of the Internet-based interventions on PA. Second, the numbers of pedometer-based studies were fewer than 10 studies, and then we could not run the metaregression analyses. Third, because of the nature of webbased interventions, the practicality of using a doubleblinded method was impossible.

Conclusion The web-based PA interventions had beneficial effects on increasing all three types of PA measurements among the general population. However, the effects appear to be dependent on the design of the study, age of participants, and duration of studies. It seems that development of a technology-based PA programs should be continued for longterm periods to assess the outcomes of the interventions, as well as focusing on appropriate educational programs that are more attractive to encourage participants.

Author statements Acknowledgments The authors acknowledged the contributions of Tabriz University of Medical Sciences, Tabriz, Iran for providing facilities to the study.

Ethical approval None sought.

Funding None declared.

Competing interests None declared.

Authors' contribution LJ designed and analyzed the research data, SSh helped the analysis, FR and TP collected and extracted the data, and MAF wrote the manuscript draft.

references

1. Physical activity. WHO. Available at: http://wwwwhoint/ dietphysicalactivity/publications/facts/pa/en/indexhtml [Accessed 30 August 2016]. 2. Warburton DE, Nicol CW, Bredin SS. Health benefits of physical activity: the evidence. Can Med Assoc J 2006;174:801e9. 3. Glasgow RE, Emmons KM. How can we increase translation of research into practice? Types of evidence needed. Annu Rev Public Health 2007;28:413e33.

45

4. Farvolden P, Cunningham J, Selby P. Using e-health programs to overcome barriers to the effective treatment of mental health and addiction problems. J Technol Hum Serv 2009;27:5e22. 5. Norman GJ, Zabinski MF, Adams MA, Rosenberg DE, Yaroch AL, Atienza AA. A review of eHealth interventions for physical activity and dietary behavior change. Am J Prev Med 2007;33:336e45. e316. 6. Wantland DJ, Portillo CJ, Holzemer WL, Slaughter R, McGhee EM. The effectiveness of Web-based vs. non-Webbased interventions: a meta-analysis of behavioral change outcomes. J Med Internet Res 2004;6:e40. 7. Webb T, Joseph J, Yardley L, Michie S. Using the internet to promote health behavior change: a systematic review and meta-analysis of the impact of theoretical basis, use of behavior change techniques, and mode of delivery on efficacy. J Med Internet Res 2010;12:e4. 8. Kroeze W, Werkman A, Brug J. A systematic review of randomized trials on the effectiveness of computer-tailored education on physical activity and dietary behaviors. Ann Behav Med 2006;31(3):205e23. 9. Davies CA, Spence JC, Vandelanotte C, Caperchione CM, Mummery WK. Meta-analysis of internet-delivered interventions to increase physical activity levels. Int J Behav Nutr Phys Act 2012;9:52. 10. Hills AP, Mokhtar N, Byrne NM. Assessment of physical activity and energy expenditure: an overview of objective measures. Front Nutr 2014;16(1):5. 11. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Ann Intern Med 2009 Aug 18;151(4):264e9. 12. Jadad AR, Moore RA, Carroll D, Jenkinson C, Reynolds DJM, Gavaghan DJ, McQuay HJ. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin Trials 1996;17:1e12. 13. Cochran WG. The combination of estimates from different experiments. Biometrics 1954;10:101e29. 14. Egger M, Smith GD, Schneider M, Minder C. Bias in metaanalysis detected by a simple, graphical test. BMJ 1997;315:629e34. 15. Dunton GF, Robertson TP. A tailored Internet-plus-email intervention for increasing physical activity among ethnically-diverse women. Prev Med 2008;47:605e11. 16. Cadmus-Bertram LA, Marcus BH, Patterson RE, Parker BA, Morey BL. Randomized trial of a fitbit-based physical activity intervention for women. Am J Prev Med 2015;49:414e8. 17. Wijsman CA, Westendorp RG, Verhagen EA, Catt M, Slagboom PE, de Craen AJ, et al. Effects of a web-based intervention on physical activity and metabolism in older adults: randomized controlled trial. J Med Internet Res 2013;15:e233. 18. Watson A, Bickmore T, Cange A, Kulshreshtha A, Kvedar J. An internet-based virtual coach to promote physical activity adherence in overweight adults: randomized controlled trial. J Med Internet Res 2012;14:e1. 19. Vroege DP, Wijsman CA, Broekhuizen K, de Craen AJ, van Heemst D, van der Ouderaa FJ, et al. Dose-response effects of a Web-based physical activity program on body composition and metabolic health in inactive older adults: additional analyses of a randomized controlled trial. J Med Internet Res 2014;16:e265. 20. Vandelanotte C, Duncan MJ, Plotnikoff RC, Mummery WK. Do participants' preferences for mode of delivery (text, video, or both) influence the effectiveness of a Web-based physical activity intervention? J Med Internet Res 2012;14:e37. 21. Spittaels H, De Bourdeaudhuij I, Vandelanotte C. Evaluation of a website-delivered computer-tailored intervention for increasing physical activity in the general population. Prev Med 2007;44:209e17.

46

p u b l i c h e a l t h 1 5 2 ( 2 0 1 7 ) 3 6 e4 6

22. Spittaels H, De Bourdeaudhuij I, Brug J, Vandelanotte C. Effectiveness of an online computer-tailored physical activity intervention in a real-life setting. Health Educ Res 2007;22:385e96. 23. Soetens KC, Vandelanotte C, de Vries H, Mummery KW. Using online computer tailoring to promote physical activity: a randomized trial of text, video, and combined intervention delivery modes. J health Commun 2014;19:1377e92. 24. van Stralen MM, de Vries H, Mudde AN, Bolman C, Lechner L. The long-term efficacy of two computer-tailored physical activity interventions for older adults: main effects and mediators. Health Psychol: Off J Div Health Psychol Am Psychol Assoc 2011;30:442e52. 25. Thorsteinsen K, Vitterso J, Svendsen GB. Increasing physical activity efficiently: an experimental pilot study of a website and mobile phone intervention. Int J Telemed Appl 2014;2014:746232. 26. Sriramatr S, Berry TR, Spence JC. An Internet-based intervention for promoting and maintaining physical activity: a randomized controlled trial. Am J health Behav 2014;38:430e9. 27. Slootmaker SM, Chinapaw MJM, Schuit AJ, Seidell JC, Van Mechelen W. Feasibility and effectiveness of online physical activity advice based on a personal activity monitor: randomized controlled trial. J Med Internet Res 2009;11:e27. 28. Roesch SC, Norman GJ, Villodas F, Sallis JF, Patrick K. Intervention-mediated effects for adult physical activity: a latent growth curve analysis. Soc Sci Med 2010;71:494e501. 29. Peels DA, Hoogenveen RR, Feenstra TL, Golsteijn RH, Bolman C, Mudde AN, et al. Long-term health outcomes and cost-effectiveness of a computer-tailored physical activity intervention among people aged over fifty: modelling the results of a randomized controlled trial. BMC Public Health 2014;14:1099. 30. Maher C, Ferguson M, Vandelanotte C, Plotnikoff R, De Bourdeaudhuij I, Thomas S, et al. A web-based, social networking physical activity intervention for insufficiently active adults delivered via facebook app: randomized controlled trial. J Med Internet Res 2015;17:e174. 31. Hansen AW, Gronbaek M, Helge JW, Severin M, Curtis T, Tolstrup JS. Effect of a Web-based intervention to promote physical activity and improve health among physically inactive adults: a population-based randomized controlled trial. J Med Internet Res 2012;14:e145. 32. Friederichs SA, Oenema A, Bolman C, Lechner L. Long term effects of self-determination theory and motivational interviewing in a web-based physical activity intervention: randomized controlled trial. Int J Behav Nutr Phys Act 2015;12:101. 33. De Cocker K, Spittaels H, Cardon G, De Bourdeaudhuij I, Vandelanotte C. Web-based, computer-tailored, pedometerbased physical activity advice: development, dissemination through general practice, acceptability, and preliminary efficacy in a randomized controlled trial. J Med Internet Res 2012;14:e53. 34. Compernolle S, Vandelanotte C, Cardon G, De Bourdeaudhuij I, De Cocker K. Effectiveness of a web-based, computer-tailored, pedometer-based physical activity intervention for adults: a cluster randomized controlled trial. J Med Internet Res 2015;17:e38.

35. Cavallo DN, Tate DF, Ries AV, Brown JD, DeVellis RF, Ammerman AS. A social media-based physical activity intervention: a randomized controlled trial. Am J Prev Med 2012;43:527e32. 36. Carr LJ, Dunsiger SI, Lewis B, Ciccolo JT, Hartman S, Bock B, et al. Randomized controlled trial testing an internet physical activity intervention for sedentary adults. Health Psychol: Off J Div Health Psychol Am Psychol Assoc 2013;32:328e36. 37. Peels DA, Bolman C, Golsteijn RH, de Vries H, Mudde AN, van Stralen MM, et al. Long-term efficacy of a printed or a Webbased tailored physical activity intervention among older adults. Int J Behav Nutr Phys Act 2013;10:104. 38. Webb T, Joseph J, Yardley L, Michie S. Using the internet to promote health behavior change: a systematic review and meta-analysis of the impact of theoretical basis, use of behavior change techniques, and mode of delivery. J Med Internet Res 2010;12:e4. 39. Jahangiry L, Montazeri A, Najafi M, Yaseri M, Farhangi MA. An interactive web-based intervention on nutritional status, physical activity and health-related quality of life in patient with metabolic syndrome: a randomized-controlled trial (The Red Ruby Study). Nutr Diabetes 2017;7:e240. 40. Farhangi MA, Jahangiry L, Mirinazhad M-M, Shojaeezade D, Montazeri A, Yaghoubi A. A web-based interactive lifestyle modification program improves lipid profile and serum adiponectin concentrations in patients with metabolic syndrome: the “Red Ruby” study. Int J Diabetes Dev Ctries 2017;37(1):21e30. 41. Linde JA, Jeffery RW, French SA, Pronk NP, Boyle RG. Selfweighing in weight gain prevention and weight loss trials. Ann Behav Med 2005;30:210e6. 42. Boutelle KN, Kirschenbaum DS. Further support for consistent self-monitoring as a vital component of successful weight control. Obes Res 1998;6:219e24. 43. Gleeson-Kreig JM. Self-monitoring of physical activity effects on self-efficacy and behavior in people with type 2 diabetes. Diabetes Educ 2006;32:69e77. 44. Burke LE, Wang J, Sevick MA. Self-monitoring in weight loss: a systematic review of the literature. J Am Diet Assoc 2011;111:92e102. 45. Bassett D, Strath SJ. Use of pedometers to assess physical activity. Phys Act Assess Health Relat Res 2002:163e77. 46. Richardson CR, Newton TL, Abraham JJ, Sen A, Jimbo M, Swartz AM. A meta-analysis of pedometer-based walking interventions and weight loss. Ann Fam Med 2008;6:69e77. 47. Tudor-Locke C, Bassett DR. How many steps/day are enough? Sports Med 2004;34:1e8. 48. Reinwand DA, Crutzen R, Elfeddali I, Schneider F, Schulz DN, Smit ES, et al. Impact of educational level on study attrition and evaluation of web-based computer-tailored interventions: results from seven randomized controlled trials. J Med Internet Res 2015;17:e228. 49. Andreassen HK, Bujnowska-Fedak MM, Chronaki CE, Dumitru RC, Pudule I, Santana S, et al. European citizens' use of E-health services: a study of seven countries. BMC Public Health 2007;7:1e7. 50. Jahangiry L, Shojaeizadeh D, Abbasalizad Farhangi M, Yaseri M, Mohammad K, Najafi M, et al. Interactive webbased lifestyle intervention and metabolic syndrome: findings from the Red Ruby (a randomized controlled trial). Trials 2015;16:1e10.