- Systematic review
- Open Access
Do mass media campaigns improve physical activity? a systematic review and meta-analysis
Archives of Public Healthvolume 71, Article number: 20 (2013)
Mass media campaigns are frequently used to influence the health behaviors of various populations. There are currently no quantitative meta-analyses of the effect of mass media campaigns on physical activity in adults.
We searched six electronic databases from their inception to August 2012 and selected prospective studies that evaluated the effect of mass media campaigns on physical activity in adults. We excluded studies that did not have a proper control group or did not report the uncertainties of the effect estimates. Two reviewers independently screened the title/abstracts and full articles. We used random-effects models to pool effect estimates across studies for 3 selected outcomes.
Nine prospective cohorts and before-after studies that followed-up 27,601 people over 8 weeks to 3 years met the inclusion criteria. Based on the pooled results from these studies, mass media campaigns had a significant effect on promoting moderate intensity walking (pooled relative risk (RR) from 3 studies=1.53, 95% Confidence Interval: 1.25 to 1.87), but did not help participants achieve sufficient levels of physical activity [4 studies pooled RR=1.02, 95% CI: 0.91 to 1.14)]. The apparent effect of media campaigns on reducing sedentary behavior (pooled RR=1.15, 95% CI: 1.03 to 1.30) was lost when a relatively low-quality study with large effects was excluded in a sensitivity analysis. In subgroup analyses, campaigns that promoted physical activity as a ‘social norm’ seemed to be more effective in reducing sedentary behavior.
Mass media campaigns may promote walking but may not reduce sedentary behavior or lead to achieving recommended levels of overall physical activity. Further research is warranted on different campaign types and in low- and middle- income countries.
The global burden of non-communicable diseases (NCDs) is rising rapidly due to ageing, poor dietary habits, harmful use of alcohol, obesity, smoking and insufficient physical activity . Globally, insufficient physical activity is the 10th leading risk factor for disease burden causing 3.2 million deaths and accounting for 31% of burden of disease from ischemic heart disease . However, a large proportion of the adults in both developed and developing countries fail to achieve the recommended levels of physical activity . Several systematic reviews have evaluated the effectiveness of different types of population- and individual-level interventions to improve physical activity [3–6]. However, four prior reviews that examined the effectiveness of mass media campaigns only included articles published in English, or over a short period of time [7, 8] and did not estimate the pooled effects due to heterogeneity in the effect measures reported in the original studies [7–10].
A quantitative estimate of the effect of mass media campaigns in promoting physical activity is required to compare the efficacy of alternative interventions, conduct cost-effectiveness analyses and make policy recommendations to prevent the epidemic of non-communicable diseases in developed and developing countries . The goal of this study was to conduct a systematic review and meta-analysis of the effect of mass media campaigns on physical activity.
We identified original research articles in Medline (through PubMed), EMBASE (Elsevier), Cochrane Library, CINAHL, PsycINFO and Web of Science, and reviewed Google Scholar for additional relevant papers. Database searches were concluded in August 2012. We searched PubMed using a number of Medical Subject Headings (MeSH) terms representing health promotion, mass media and physical activity (Table 1). We restricted the search to studies of adults (≥19 years old) but did not restrict by year of publication or language. We searched EMBASE using similar search terms and restrictions and checked the reference list of systematic reviews for additional articles. Two authors (A.A. and K.H.) independently screened the titles and abstracts for eligibility and subsequently examined the full texts of the articles. Any discrepancies were resolved in consultation with the third author (G.D.).
In both rounds of title/abstract and full text review, we excluded studies that did not evaluate the effect of mass media campaigns on physical activity. These included case-reports, cross-sectional studies, baseline-only or terminal-only surveys, and studies including only information on awareness, attitude or knowledge. Studies reporting mass media campaigns that were implemented simultaneously with other interventions were also excluded. If several articles reported results from the same study with the same effect measure, we included the more recent report unless an earlier report provided the relevant effect measures or had a higher quality. Studies that did not report measures of uncertainty for the effect estimates were excluded.
We extracted data on characteristics of the study population, including sample size, sampling method, eligibility criteria, mean or median follow-up time, response rate, age, sex and body mass index (BMI). We also extracted data on campaign characteristics including duration, scope, type of media, frequency and coverage; whether the campaigns were based on behavior change theory, prior research, or consultation with experts; use of risk message, shocking effects, social norm or celebrities. Coverage refers to the estimated proportion of the target population that is reported to have been reached by the media campaign. Point estimates of effect measures (odds ratios and difference measures) and their uncertainty intervals were extracted. If a study reported several effect measures in subgroups of participants defined at baseline (e.g., male versus female) we extracted and used the effect estimates separately for each subgroup. Data extraction was done by one investigator (A.A) and reviewed by another (G.D.).
We developed a set of quality criteria with binary scores to reflect (i) outcome ascertainment using pedometers or a validated questionnaire, (ii) objective reporting of exposure to mass media outlets, and (iii) large potential for confounding or selection bias. We assigned an additional two points for randomized trials with comparison groups and one point for observational studies with repeated measures.
The reported odds ratios and difference measures were converted to relative risks (RR) using RR = OR/((1-Po)+(Po*OR)) where OR is odds ratio and Po is prevalence of the outcome in the unexposed . We only pooled the results in a meta-analysis if three or more studies had reported the same effect measure which was the case for ‘reduction in sedentary lifestyle’ , ‘sufficient walking’ and ‘sufficient overall physical activity’. Sufficient walking was mostly defined as moderate intensity or brisk walking for at least 150 minutes per week. Sufficient physical activity was defined as aerobic and muscle-strengthening activity lasting at least 150 minutes per week in moderate intensity (3.0 to 6.0 Metabolic Equivalent Tasks) or 60 minutes per week if vigorous.
Pooled effect measures and 95% confidence intervals were estimated using random effects models , with Forest plots for graphical visualization. Random effects models do not assume that included studies are estimating the exact same parameters, allowing for some variability. Heterogeneity was formally assessed with the Q and I2 statistics. We used meta-regression to explore heterogeneity due to study design (cohort versus before-after studies), mean age and BMI of participants and proportion of men at baseline, year of study baseline, median duration of follow-up, type of media, coverage rate, approach of intervention, and study quality.
Publication bias was evaluated with Begg’s and Egger’s tests and influence analysis was conducted to ascertain the effects of omitting individual studies. We conducted a set of alternative analyses to evaluate the sensitivity of our results to the inclusion of studies that reported the effect of mass media campaign on a slightly different outcome measure or for a different duration of campaign or follow-up. A two-sided p value of 0.05 or less was considered as significant. All statistical analyses were conducted using Stata version 11 (College Station, TX).
We identified 723 unique articles from the database searches and after screening the abstracts, reviewed 55 full-text articles. Of these, nine studies were selected for the meta-analysis [14–22] (Figure 1). Selected studies were all conducted in high-income countries and between 1996 and 2008. The nine studies enrolled a total of 27,601 participants and used media campaigns that lasted anywhere between 8 weeks to 3 years. Two studies used a prospective cohort design [17, 18], five used a before-after design with comparison groups [14, 19–22] and the remaining two used a before-after design without a comparison group [15, 16]. The media campaigns were conducted on local, regional or national levels with coverage ranging from 11 to 90%. Some studies objectively reported the intensity of the mass media campaigns using ‘gross rating points’ or other similar measures [14, 19–22]. Quality scores varied from 2 to 4 out of 5 points, with a median of 3 (Table 2).
The selected studies had reported on eleven different measures of physical activity. Four studies reported effects on reducing sedentary behavior [14, 17–19]. The pooled relative risk from the six comparisons reported in these four studies was 1.15 (95% confidence interval (CI) 1.03 to 1.30). There was marked heterogeneity across these studies (I2=63%, p=0.018, see Figure 2 Panel A) and when a relatively low-quality study with 3 sub-groups reported was excluded, the pooled relative risk was not significantly different from the null (pooled RR 1.06, 95%CI: 0.95 to 1.17) (Figure 2 Panel B). Mean age at baseline was an important determinant of heterogeneity (p=0.054) in meta-regression analyses (Table 3). Each additional 10 years of mean age was associated with a 27% higher reduction in sedentary behavior. Media campaigns based on ‘social norm’  were more likely to lead to reduction in sedentary behavior (RR=1.33, 95% CI: 1.01 to 1.43) compared with those using celebrities or based on a 'risk message' (RR=1.05, 95% CI: 0.92 to 1.21).
Three studies [20–22] reported effects of media campaigns on achieving sufficient walking (Figure 3). When pooled, the results indicated that mass media campaigns increased the likelihood of achieving sufficient walking by 53% (RR=1.53, 95% CI: 1.25 to 1.87). There was no heterogeneity possibly because the studies were similar in design and implementation (I2 =0%, p=0.952).
Finally, four studies [14, 15, 17, 20] reported the effects on achieving sufficient overall physical activity (Figure 4). The pooled RR was 1.02 (95% CI: 0.91 to 1.14) with significant heterogeneity (I2=72%, p=0.012). None of the selected factors for meta-regression analysis were significant sources of heterogeneity across these four studies (Table 3).
Begg's and Egger's tests were suggestive of publication bias in the meta-analysis of the effect of mass media campaign on sedentary behavior (p=0.009). There was no evidence of publication bias in the other two meta-analyses: the p-value for the Egger’s test was 0.224 for sufficient walking and 0.597 for sufficient physical activity. Our influence analysis showed that the pooled effect estimates were not dominated by any of the individual studies on sufficient walking or sufficient overall physical activity (results not shown).
Other sensitivity analyses
When we included estimates from a study  that had a more stringent definition for sufficient walking (1 hour per day as opposed to 150 minutes per week) the pooled RR decreased from 1.53 (95% CI: 1.25 to 1.87) to 1.35 (95% CI: 1.05 to 1.72). There was substantial heterogeneity in the estimates (I2=68%, p=0.024) but none of the selected factors was a significant determinant of heterogeneity in meta-regression analyses. Furthermore, one of the three studies on sufficient walking  also reported the effects using a longer follow-up time (12 months compared with 3 months in other studies). We substituted the 12-month effect estimate from this same study for the 3-month effect in a sensitivity analysis. The pooled RR and heterogeneity did not change materially (RR=1.50, 95% CI: 1.20 to 1.88 (I2=0.0%, p=0.901)).
The pooled estimates for sufficient physical activity did not change materially when we excluded one study for which the duration of intervention was 8 weeks compared to the others with interventions lasting up to 24 months (RR = 1.06; 95% CI: 0.97 to 1.53). Heterogeneity remained substantial (I2=63.5%, p=0.065) but none of the selected factors was a significant determinant of heterogeneity.
Our meta-analyses of nine prospective studies found that mass media campaigns may improve sufficient walking but may not reduce sedentary lifestyles or encourage participants to achieve recommended levels of overall physical activity. We included moderate and high quality studies and our results for walking and sufficient physical activity were robust to the selected sensitivity analyses whereas the results for sedentary behavior were influenced by a single study and showed evidence of publication bias. Except for the ‘approach’ of the campaign and the mean age of participants that seemed to change the effect of campaigns on sedentary lifestyle, we found no other significant determinant of heterogeneity among the selected studies.
Our results have important implications for the design and implementation of NCD prevention programs that aim to improve physical activity. The pooled RRs suggest that a well-designed mass media campaign may increase the likelihood of achieving sufficient walking by 53% which is equivalent to about 80 minutes per week. A recent meta-analysis of prospective studies found that an additional 150 minutes of walking over 5 days led to a 19% reduction in risk of coronary heart disease . Applying this effect size to our results indicates a potential 11% reduction in risk of coronary heart disease following a well-designed mass media campaign.
Results from four previous systematic reviews of mass media campaigns and physical activity were mixed. The investigators qualitatively assessed the totality of evidence but did not conduct a meta-analysis. Two previous reviews concluded that mass media have either no effect or a very small effect on physical activity [8, 9] and another review suggested a significant effect on physical activity levels without specifying the effect size or the type of activities that were influenced . The original studies included in these four reviews reported different outcome measures and used widely different evaluation methods including sub-optimal designs such as post-campaign cross-sectional surveys [7, 8]. Systematic reviews of mass media and other health behaviors have faced a similar challenge. For instance, mass media interventions were found effective in encouraging their audience to quit smoking, but the effects were derived from heterogeneous studies of variable quality [24, 25]. Similarly, pooled analyses of mass media and diet found a beneficial effect but the pooled studies were widely different in design and quality [26–28].
Other interventions to promote physical activity have been systematically reviewed. Several prior meta-analyses have reported the effects of pedometers , internet-based interventions [4, 5] and telephone calls  on physical activity. The pooled effects were generally larger than those we observed for media campaigns, but similar to those reported for exercise referral schemes  and computer-tailored interventions [3, 30].
Strengths and limitations
We selected 9 moderate to high-quality studies and extracted comparable metrics of effect. When comparable metrics were not reported, we used the reported results to calculate a common metric for pooling. We explored the sources of heterogeneity across studies using meta-regression. However, our systematic review was still limited by the marked differences in the reported outcomes of the selected studies. We did not have sufficient power to detect differences across studies by study-level characteristics due to the small number of selected studies.
We were also unable to evaluate the dose–response curve for mass media campaigns because most studies did not provide sufficient information to calculate a uniform metric for exposure to campaign (such as ‘gross rating points’). Measurement error in assessment of physical activity is possible in the original studies because few studies used validated questionnaires or objective measurements of activity. This study was also limited by its focus on adults and the findings among young people may differ from ours. Finally, the selected studies were all conducted in developed countries. Therefore, our results cannot be generalized to developing countries.
In summary, our results indicated that mass media campaigns may promote walking, but may not reduce sedentary behavior or lead to achieving sufficient physical activity. Further research is required to examine the effect of mass media campaigns on other measures and types of physical activity (such as time spent walking and overall time spent in physical activity). We suggest that investigators report intensity and frequency of mass media campaigns using standard metrics and measure physical activity objectively or using validated questionnaires. Similar evaluations are needed to examine the effect of mass media campaigns in low and middle-income countries and in different cultural milieus.
Lim SS, Vos T, Flaxman AD, Danaei G, Shibuya K, Adair-Rohani H, Amann M, Anderson HR, Andrews KG, Aryee M, et al: A comparative risk assessment of burden of disease and injury attributable to 67 risk factors and risk factor clusters in 21 regions, 1990—2010: a systematic analysis for the global burden of disease study 2010. Lancet. 2012, 380 (9859): 2224-2260. 10.1016/S0140-6736(12)61766-8.
Bull FC, Armstrong T, Dixon T, Ham S, Neiman A, Pratt M: Physical inactivity. Comparative quantification of health risks : global and regional burden of diseaseattributable to selected major risk factors. Edited by: Ezzati M, Alan D, Lopez AD, Rodgers A, Murray CJL. 2004, Geneva: World Health Organization, 730-881.
Heath GW, Parra DC, Sarmiento OL, Andersen LB, Owen N, Goenka S, Montes F, Brownson RC: Evidence-based intervention in physical activity: lessons from around the world. Lancet. 2012, 380 (9838): 272-281. 10.1016/S0140-6736(12)60816-2.
van den Berg MH, Schoones JW, Vliet Vlieland TP: Internet-based physical activity interventions: a systematic review of the literature. J Med Internet Res. 2007, 9 (3): e26-10.2196/jmir.9.3.e26.
Vandelanotte C, Spathonis KM, Eakin EG, Owen N: Website-delivered physical activity interventions a review of the literature. Am J Prev Med. 2007, 33 (1): 54-64. 10.1016/j.amepre.2007.02.041.
Eakin EG, Lawler SP, Vandelanotte C, Owen N: Telephone interventions for physical activity and dietary behavior change: a systematic review. Am J Prev Med. 2007, 32 (5): 419-434. 10.1016/j.amepre.2007.01.004.
Leavy JE, Bull FC, Rosenberg M, Bauman A: Physical activity mass media campaigns and their evaluation: a systematic review of the literature 2003–2010. Health Educ Res. 2011, 26 (6): 1060-1085. 10.1093/her/cyr069.
Marcus BH, Owen N, Forsyth LH, Cavill NA, Fridinger F: Physical activity interventions using mass media, print media, and information technology. Am J Prev Med. 1998, 15 (4): 362-378. 10.1016/S0749-3797(98)00079-8.
Cavill N, Bauman A: Changing the way people think about health-enhancing physical activity: do mass media campaigns have a role?. J Sports Sci. 2004, 22 (8): 771-790. 10.1080/02640410410001712467.
Brown DR, Soares J, Epping JM, Lankford TJ, Wallace JS, Hopkins D, Buchanan LR, Orleans CT: Stand-alone mass media campaigns to increase physical activity: a community guide updated review. Am J Prev Med. 2012, 43 (5): 551-561. 10.1016/j.amepre.2012.07.035.
Ezzati M, Riboli E: Can noncommunicable diseases be prevented? lessons from studies of populations. Science. 2012, 337 (6101): 1482-1487. 10.1126/science.1227001.
Zhang J, Yu KF: What's the relative risk? a method of correcting the odds ratio in cohort studies of common outcomes. JAMA. 1998, 280 (19): 1690-1691. 10.1001/jama.280.19.1690.
DerSimonian R, Laird N: Meta-analysis in clinical trials. Control Clin Trials. 1986, 7: 177-188. 10.1016/0197-2456(86)90046-2.
Bauman A, Armstrong T, Davies J, Owen N, Brown W, Bellew B, Vita P: Trends in physical activity participation and the impact of integrated campaigns among Australian adults, 1997–99. Aust N Z J Public Health. 2003, 27 (1): 76-79. 10.1111/j.1467-842X.2003.tb00384.x.
Bauman A, McLean G, Hurdle D, Walker S, Boyd J, van Aalst I, Carr H: Evaluation of the national 'Push Play' campaign in New Zealand–creating population awareness of physical activity. N Z Med J. 2003, 116 (1179): U535-
Craig CL, Tudor-Locke C, Bauman A: Twelve-month effects of Canada on the move: a population-wide campaign to promote pedometer use and walking. Health Educ Res. 2007, 22 (3): 406-413.
Hillsdon M, Cavill N, Nanchahal K, Diamond A, White IR: National level promotion of physical activity: results from England's ACTIVE for LIFE campaign. J Epidemiol Community Health. 2001, 55 (10): 755-761. 10.1136/jech.55.10.755.
Hopman-Rock M, Borghouts JAJ, Leurs MTW: Determinants of participation in a health education and exercise program on television. Prev Med. 2005, 41 (1): 232-239. 10.1016/j.ypmed.2004.11.014.
Lorentzen C, Ommundsen Y, Jenum AK, Holme I: The "romsas in motion" community intervention: program exposure and psychosocial mediated relationships to change in stages of change in physical activity. Int J Behav Nutr Phys Act. 2007, 4: 15-10.1186/1479-5868-4-15.
Reger-Nash B, Bauman A, Booth-Butterfield S, Cooper L, Smith H, Chey T, Simon KJ: Wheeling walks: evaluation of a media-based community intervention. Fam Community Health. 2005, 28 (1): 64-78. 10.1097/00003727-200501000-00009.
Reger-Nash B, Bauman A, Cooper L, Chey T, Simon KJ, Brann M, Leyden KM: WV walks: replication with expanded reach. J Phys Act Health. 2008, 5 (1): 19-27.
Reger-Nash B, Fell P, Spicer D, Fisher BD, Cooper L, Chey T, Bauman A: BC walks: replication of a communitywide physical activity campaign. Prev Chronic Dis. 2006, 3 (3): A90-
Zheng H, Orsini N, Amin J, Wolk A, Nguyen VT, Ehrlich F: Quantifying the dose–response of walking in reducing coronary heart disease risk: meta-analysis. Eur J Epidemiol. 2009, 24 (4): 181-192. 10.1007/s10654-009-9328-9.
Bala M, Strzeszynski L, Cahill K: Mass media interventions for smoking cessation in adults. Cochrane Database Syst Rev. 2008, 1: CD004704-
Brinn MP, Carson KV, Esterman AJ, Chang AB, Smith BJ: Mass media interventions for preventing smoking in young people. Cochrane Database Syst Rev. 2010, 11: CD001006-
Beaudoin CE, Fernandez C, Wall JL, Farley TA: Promoting healthy eating and physical activity short-term effects of a mass media campaign. Am J Prev Med. 2007, 32 (3): 217-223. 10.1016/j.amepre.2006.11.002.
Maddock J, Maglione C, Barnett JD, Cabot C, Jackson S, Reger-Nash B: Statewide implementation of the 1% or less campaign. Health Educ Behav. 2007, 34 (6): 953-963.
Verheijden MW, van Dommelen P, van Empelen P, Crone MR, Werkman AM, van Kesteren NM: Changes in self-reported energy balance behaviours and body mass index during a mass media campaign. Fam Pract. 2012, 29 (Suppl 1): i75-i81. 10.1093/fampra/cmr133.
Williams NH, Hendry M, France B, Lewis R, Wilkinson C: Effectiveness of exercise-referral schemes to promote physical activity in adults: systematic review. Br J Gen Pract. 2007, 57 (545): 979-986. 10.3399/096016407782604866.
Short CE, James EL, Plotnikoff RC, Girgis A: Efficacy of tailored-print interventions to promote physical activity: a systematic review of randomised trials. Int J Behav Nutr Phys Act. 2011, 8: 113-10.1186/1479-5868-8-113.
The authors declare that they have no competing interests.
GD designed the study. AA and KH conducted the search and screened the articles. AA extracted the data and conducted the analyses. GD and AA wrote the first draft of the manuscript. KH provided comments on the manuscript draft. GD oversaw the process and is the study guarantor. All authors read and approved the final manuscript.