Skip to main content
. 2021 May 14;23(5):e19688. doi: 10.2196/19688

Table 4.

Results of reviews on physical activity, ordered by type of control and further ordered by mode of delivery of intervention.

Review Relevant studies, n (total studies) Method of synthesis Interventions Outcomes Follow-up Summary of findings AMSTAR-2 rating
Mixed controls (active and nonactive)

Aalbers et al (2011) [47] 2 (10) Narrative synthesis of randomized and nonrandomized pre-post controlled trials Internet Total PAa and MVPAb 1.5-6.5 months 2 studies reported the effects of digital interventions, one was less effective on MVPA than a nonactive control and the other demonstrated a small positive effect on total PAb (P=.001). Critically low

Aneni et al (2014) [100] 9 (29) Narrative synthesis of RCTsc Internet PA measures 6-24 months No improvement was seen in virtually all the studies with PA outcome, only 11% (1/9) of studies demonstrated a significant intervention effect on PA. Critically low

Bottorf et al (2014) [26] (PA only) 8 (35) Narrative synthesis of all study types Internet Changes in PA; step count; self-reported walking; BMI; waist circumference; weight 3-12 months 63% (5/8) of studies demonstrated that PA significantly increased in the internet-based interventions, 2 studies showed a nonsignificant difference, and one showed that the effects were indeterminable. Critically low

Davies et al (2012) [29] (PA only) 34 (34) Meta-analysis of experimental design studies Internet PA 2-52 weeks The estimated overall mean effect of internet-delivered interventions on PA was Cohen d=0.14 (P<.001). Homogeneity tests from the fixed-effect analysis revealed significant heterogeneity across studies (Q=73.75; P<.001). The overall mean effect for sustained PA at least 6 months postintervention (n=11) resulted in a small but significant effect size Cohen d=0.11 (P<.01). Critically low

George et al (2012) [33] (PA only) 2 (14) Narrative synthesis of all study types Internet Step count; health status; BMI; weekly PA 2-8 months Increase in PA in 100% (2/2) of online interventions where participants were in competitive teams, including one that showed an increase in step count. Poor quality evidence. Critically low

Hou et al (2013) [108] 7 (38) Narrative synthesis of trials with comparison or control group Internet Level of physical activity 0-12 months 86% (6/7) of interventions were successful in the studies focusing primarily on PA. Critically low

Jahangiry et al (2017) [35] (PA only) 21 (22) Meta-analysis of controlled trials Internet MVPA; walking; step count (pedometer) 1-20 weeks 36% (5/14) of rials reporting MVPA, 50% (3/6) of trials reporting step count, and 29% (4/14) of studies reporting minutes walking showed significant increases. The interventions were influenced by the age of participants and trial length. Critically low

Lustria et al (2013) [110] 12 (40) Meta-analysis of experimental and quasiexperimental studies Internet Levels of PA 4 weeks-24 months The sample size–weighted mean effect size for studies on PA was not significant Cohen d=0.059 (k=12; 95% CI −0.02 to 0.14). Critically low

Maon et al (2012) [63] 13 (26) Meta-analysis and narrative synthesis of RCTs Internet PA levels, sedentary behavior, and MVPA 6 weeks-2 years 54% (7/13) of studies showed statistically significant effects on PA levels, such as increased walking or decreased sedentary behavior. However, a meta-analysis on 4 studies with extractable data for the outcome of moderate-to-vigorous weekly PA found a not statistically significant improvement: SMDd 0.15 (95% CI 20.06 to 0.35; P=.16)
Duration of studies and effects: 10% (3/30) of studies showed positive effects when outcomes were measured immediately after the end of the interventions. In total, 37% (11/30) of studies that lasted 3 months or less demonstrated positive outcomes; 43% (13/30) of studies with an intervention of 3-6 months showed positive results; and only 10% (3/30) interventions that lasted longer than 6 months were reported to have positive results.		 Critically low

Webb et al (2010) [113] 20 (85) Meta-analysis of RCTs Internet Level of PA 3-12 months Small effects on behavior were observed for interventions that targeted only PA (Cohen d+=0.24; k=20; 95% CI 0.09 to 0.38). Critically low

Buchholz et al (2013) [27] (PA only) 10 (10) Narrative synthesis of RCTs, quasiexperimental and, single groups Mobile (SMS text messaging) Self-reported frequency or pedometer-reported steps and level of PA 3-52 weeks Effect sizes across all studies were positive; the median effect size was 0.5 (medium) but heterogeneous. Sample sizes were small. Critically low

Elavsky (2018) [31] (PA only) 50 (52) Narrative synthesis of RCTs and pre-post studies Mobile PA and sedentary behavior <3 months 59% (17/29) of RCTs and 62% (13/21) of pre-post studies supported the effectiveness of mobile interventions to improve PA, and 9 (5 of 10 RCTs and all 4 pre-post) of 14 (64%) studies reduced sedentary behavior. Critically low

Lyzwinksi et al (2014) [60] 9 (14) Meta-analysis of RCTs Mobile Levels of PA 8 weeks-12 months Trials mostly found that PA levels increased in the intervention groups relative to the control groups. Critically low

Maher et al (2014) [61] 4 (10) Narrative synthesis of studies with comparator group (control or within subject) Mobile Levels of PA 8 weeks-24 months 25% (1/4) of studies demonstrated a significant change in PA Cohen d=0.84 (95% CI −0.49 to 1.19). Critically low

Muntaner et al (2015) [41] (PA only) 11 (11) Narrative synthesis of all study types Mobile PA; exercise 2-24 weeks 55% (6/11) of articles included in this review reported significant increases in PA levels. Critically low

O'Reilly et al (2013) [42] (PA only) 12 (22) Narrative synthesis of RCTs Mobile PA; sedentary behavior; BMI; blood lipids; blood pressure; QoLe; adverse effects Not reported 75% (9/12) of studies reported significant changes in PA or sedentary behavior. Critically low

Palmer et al (2018) [111] 15 (71) Meta-analysis of RCTs Mobile Level of PA 3 months Trials of PA interventions reporting outcomes at 3 months showed no benefits. Moderate

Schoeppe et al (2016) [68] 10 (27) Narrative synthesis of RCTs, randomized trials, controlled trials, and pre- and poststudies Mobile PA; sedentary behavior 1-24 weeks 59% (13/22) of studies reported significant improvements in levels of PA; 20% (1/5) of studies reported a significant change in sedentary behavior. Critically low

Elaheebocus et al (2018) [56] 25 (134) Narrative synthesis of RCTs Social media Body weight 6-48 months 76% (19/25) of studies using online social networks had positive results. Critically low

Mita et al (2016) [65] 11 (16) Meta-analysis of RCTs Social media PA; weight change 1-12 months For PA, significant mean difference 0.07; 95% CI −0.25 to 0.38; k=11. Moderate

Williams et al (2012) [75] 12 (22) Meta-analysis of RCTs Social media PA 3-24 months Meta-analysis showed no significant differences in changes in PA (SMD 0.13; 95% CI −0.04 to 0.30; k=12). Critically low

Willis et al (2017) [76] 3 (5) Narrative synthesis of all study types Social media Total PA 8 weeks-6 months Only one study reported significant changes in levels of PA, when the web-based social network intervention included an online support group. Critically low

Johnson (2017) [37] 10 (19) Narrative synthesis of RCTs Active gaming Behavioral and cognitive outcomes Not reported Findings were largely positive for behavioral impacts, specifically the impact of gamification for PA: 80% (8/10) positive and 20% (2/10) mixed. Critically low

Peng et al (2012) [43] (PA only) 4 (12) Narrative synthesis of all study designs Active gaming Heart rate; energy expenditure; and oxygen uptake 6-12 weeks Evidence does not support active video games as an effective tool to significantly increase PA or exercise attendance. Critically low

Street et al (2017) [46] (PA only) 9 (9) Narrative synthesis of studies with comparison or control groups Active gaming PA; maximum oxygen uptake; power; blood pressure; body mass; body weight; body fat; BMI; balance; speed; and strength 6-12 weeks Moderate-to-high exergaming participation was associated with statistically significant improvements in anthropometric outcomes but low participation was not associated with anthropometric changes. 38% (3/8) studies that investigated anthropometric outcomes, including BMI and body fat, found a statistically significant improvement, all 3 studies showed positive health outcomes associated with moderate-to-high participation in exergaming; 100% (3/3) of studies that reported on PA frequency reported higher frequency in the exergaming condition; however, a different 100% (3/3) of studies that reported on overall PA found no statistically significant increases. Critically low

Wieland et al (2012) [74] 4 (18) Meta-analysis of RCTs, quasi-RCTs, cluster RCTs, and quasiexperimental studies Computer-delivered Steps per day and minutes walked continuously 4 weeks-30 months No studies demonstrated statistically significant effects on PA. Critically low

Afshin et al (2016) [99] 33 (224) Narrative synthesis of RCTs and quasi-experimental studies Various: internet and mobile Level of PA 1 week-5 years 88% (29/33) of studies reported significant improvement in PA; 83% (5/6) of phone interventions were effective, including 66% (2/3) of SMS text messaging interventions, 100% (2/2) of apps, and 100% (1/1) of automated voice response.

 Critically low

Carvalho de Menzes et al (2016) [53] 13 (18) Narrative synthesis of all study designs Various: email, telephone, websites Level of PA 1-36 months Most studies demonstrated statistically significant improvements in the level of PA. Critically low

Hakala et al (2017) [34] (PA only) 13 (23) Meta-analysis of RCTs Various: mobile, text messages, pedometers, wearables, email PA: self-reported or using an accelerometer or pedometer 3 weeks-24 months No differences were observed between the experimental and control groups (risk ratio 1.03; 95% CI 0.92 to 1.15; P=.57). Critically low

Muellmann et al (2018) [39] (PA only) 13 (20) Narrative synthesis of experimental designs and quasiexperimental studies SMS text messaging and internet PA and number of steps per day 4 weeks-24 months 75% (3/4) of studies using mobile phones demonstrated significant differences in the level of PA or steps per day (mixed controls). In 100% (9/9) of studies, internet interventions significantly increased PA compared with nonactive controls. Critically low

Muller and Khoo (2014) [40] (PA only) 4 (16) Narrative synthesis of RCTs and quasiexperimental studies Various: internet and mobile PA 1 week-18 months 75% (3/4) of studies reported significant improvements in PA; 25% (1/4) of studies reported nonsignificant decrease in PA. Critically low

Stephenson et al (2017) [45] (PA only) 15 (17) Meta-analysis of RCTs Various: mobile messaging, mobile apps, website, wearable technology Sedentary behavior 5 days-24 months Interventions using computer and mobile and wearable technologies can be effective in reducing sedentary behavior. Effectiveness appeared most prominent in the short-term and lessened over time. Meta-analysis of 88% (15/17) of RCTs suggested that computer, mobile, and wearable technology tools resulted in a mean reduction of −41.28 min/day of sitting time (95% CI −60.99 to −21.58; I2=77%). The pooled effects showed mean reductions at short (≤3 months), medium (>3 to 6 months), and long-term (>6 months) follow-up of −42.42 min/day, −37.23 min/day, and −1.65 min/day, respectively. Critically low
Nonactive controls

Jenkins et al (2009) [36] (PA only) 5 (22) Narrative synthesis of RCTs Internet PA 0-24 months Results were mixed; internet interventions can be effective, compared with control conditions, although poor compliance was an issue. 50% (2/4) studies reported an increase in PA compared with nonactive controls while 2 studies found no difference. Critically low

Bock et al (2014) [24] (PA only) 4 (50) Narrative synthesis of RCTs and quasiexperimental studies Internet or computer Weekly PA, proportion of sufficiently active persons; step counts 0 weeks-3 years Interventions had a nonsignificant, positive effect on PA (P=.88). Critically low

Krebs et al (2010) [109] 25 (76) Meta-analysis of RCTs Computer-delivered Minutes of PA 1-18 months The mean effect size was g=0.16 (95% CI 0.10 to 0.21). Low

Bort-Roig et al (2014) [25] (PA only) 5 (26) Narrative synthesis of comparative and pre-postdesign Mobile PA (steps); energy expenditure; body weight and body fat; blood pressure and cholesterol; QoL 2 weeks-6 months 80% (4/5) of studies assessing PA intervention effects reported PA increases, with mean PA increases ranging from 800 to 1104 steps/day. Studies were small with differences in baseline characteristics. Critically low

Direito et al (2017) [30] (PA only) 17 (21) Meta-analysis of RCTs Mobile PA, MVPA, walking and sedentary behavior 1-52 weeks Not effective for MVPA outcomes, based only on adult studies SMD 0.14 (95% CI −0.10 to 0.37). For sedentary behavior outcomes, SMD −0.21 (95% CI −0.59 to 0.18). Critically low

Freak-Poli et al (2013) [32] (PA only) 4 (4) Narrative synthesis of RCTs and cluster RCTs Wearable technology PA; sedentary behavior; BMI; blood lipids; blood pressure; QoL; adverse effects 3-8 months Overall, there was insufficient evidence to assess the effectiveness of pedometer interventions in the workplace. 75% (3/4) of studies compared with a minimal control group, 33% (1/3) of studies observed an increase in PA under a pedometer program, but the other two did not find a significant difference. Moderate

An et al (2017) [49] 21 (22) Meta-analysis of RCTs, pre-post studies, and cohort studies Social media PA, sedentary behavior 3-102 weeks Interventions increased daily number of steps taken by 1530 (95% CI 82 to 2979). However, they were not associated with energy expenditure, total PA, or MVPA. Critically low

Tsoli et al (2018) [73] 3 (15) Meta-analysis of RCTs Interactive voice responses PA 6 weeks-12 months Interventions led to a small but statistically significant increase in PA (g=0.254; 95% CI 0.068 to 0.439; k=3; P=.007). Critically low
Active controls

Beishuizen et al (2016) [102] 5 (57) Meta-analysis of RCTs Internet Level of PA 4 weeks-3 months Interventions led to an increase in PA (SMD 0.25; 95% CI 0.10 to 0.39). Low

Covolo et al (2017) [104] 23 (40) Narrative synthesis of RCTs Mobile apps Daily steps, frequency, and intensity of PA 6 months-2 years 30% (7/23) of RCTs showed a significant increase in PA in the intervention group (measured in daily steps, frequency of PA, or level of intensity), 48% (11/23) of studies did not show a significant increase, and in 21% (5/23) studies, outcome measures were inconsistent in whether there was a significant difference between intervention and control. Critically low

Mateo et al (2015) [38] (PA only) 10 (11) Meta-analysis of controlled trials Mobile apps PA, MVPA, and steps 6 weeks-9 months Compared with the control group, use of a mobile phone app was associated with significant changes in body weight and BMI of −1.04 kg (95% CI −1.75 to −0.34; I2=41%) and −0.43 kg/m2 (95% CI −0.74 to −0.13; I2=50%), respectively (k=9); however, a nonsignificant difference in PA was observed between the intervention and comparison groups (SMD 0.40; 95% CI −0.07 to 0.87; I2=93%). Critically low

Song et al (2018) [44] (PA only) 6 (8) Narrative synthesis of all study types Mobile PA (frequency and step count); BMI; blood glucose 4 weeks-6 months Significant effects on frequency of PA in 80% (4/5) of studies (though the effect was reported to have disappeared after the 12-week follow-up), step count in 66% (2/3) of studies, BMI in 50% (2/4) of studies, and reduction in glucose in 100% (2/2) studies. Critically low

Cheatham et al (2018) [28] (PA only) 25 (25) Narrative synthesis of controlled clinical trials Wearable technology PA; BMI; weight; blood pressure; Resting Energy Expenditure; body composition; cardiovascular fitness; work productivity and absenteeism; waist circumference; blood parameters 3 weeks-24 months An activity tracker combined with a comprehensive weight loss program may provide superior short-term (≤6 months) results than a standard weight loss program in middle aged or older adults. 80% (20/25) of studies reported higher weight loss when an activity tracker was used with a weight loss intervention. Critically low

aPA: physical activity.

bMVPA: moderate-to-vigorous physical activity.

cRCT: randomized controlled trial.

dSMD: standardized mean difference.

eQoL: quality of life.