Physical activity levels in the United States have remained relatively unchanged during the past 20 years (1). Physical inactivity is responsible for approximately 250,000 deaths annually in the United States (2) and is likely a major contributing factor in the epidemics of overweight, obesity, and diabetes (3). In 1995 and 1996, major public health reviews recognized the health benefits of moderate-intensity physical activity, such as walking 30 minutes daily (4-5). The recommendation of the Centers for Disease Control and Prevention (CDC), American College of Sports Medicine (ACSM), and the U.S. surgeon general is to engage in moderate-intensity physical activity at least 30 minutes per day 5 days per week. Individuals not meeting the recommendation are deemed insufficiently active and at greater risk for chronic disease (4-5).

The pervasiveness of physical inactivity represents a populationwide problem (6). Most health promotion efforts, focused on individual and small group programs, have little impact (7). By contrast, populationwide methods to increase moderate-intensity daily physical activity represent a primary prevention strategy to address the growing problem of energy imbalance as well as the increases in chronic disease morbidity that confront industrialized societies (5). Early populationwide cardiovascular interventions had limited capacity to promote physical activity (8-11). More recent campaigns, including CDC's national VERB campaign, have documented significant changes in physical activity (12-18).

Social marketing strategies may promote populationwide public health change (19). Through the use of targeted mass media and environmental supports, social marketing strategies can increase public awareness and set a public agenda for improving healthy lifestyles (19-22).

One social marketing initiative, Wheeling Walks, involved paid mass media, public relations, and community health activities, along with efforts focused on policy and environmental changes in the community of Wheeling, WVa (23-25). This model campaign targeted older adults with 8 weeks of media messages encouraging 30 minutes or more of daily moderate-intensity walking. Ninety percent of individuals surveyed in Wheeling reported exposure to the campaign through mass media, and the campaign resulted in a 14% increase in the number of participants who changed from nonactive to active walkers.

Although model programs may succeed in their own communities, they frequently fail in other communities. To determine the effectiveness of the Wheeling Walks campaign methodology in another community, BC Walks was implemented in Broome County, New York. This paper reports on the short-term evaluation of the BC Walks campaign.

There was a short-term impact of the BC Walks campaign in Broome County that was not seen in the comparison community. Process and impact data suggest that BC Walks faithfully replicated the Wheeling Walks intervention methodology in a larger community but that BC Walks achieved smaller effects.

Few studies have examined the repeatability of the additional walking questions asked in the BRFSS. In one Australian study using a representative population sample, the BRFSS walking question had acceptable repeatability, with an intraclass correlation of 0.45 (95% CI, 0.30–0.58). This was the second most reproducible of five survey-based walking estimates (29). Also, in this study, the additional walking questions on the BRFSS were more reproducible than the moderate- and vigorous-intensity physical activity questions in the BRFSS. Furthermore, the Australian estimate was also similar to one estimated for 106 American women, with an intraclass correlation for walking of 0.40 (95% CI, 0.23–0.55) (30). Moderate reliability and validity of the physical activity module of the BRFSS has been reported (31).

The increase in walking was smaller in Broome County than in Wheeling (23). No statistically significant difference was observed for categorical change from nonactive to active walker in the BC Walks campaign; the Wheeling Walks campaign resulted in a significant increase of 14% in the number of participants who changed from nonactive to active walker. Regression to the mean may explain the declines in walking behavior observed in the more active Groups 3 and 4.

Although data generally do not suggest weight management and reduction in obesity are associated with 30 minutes of moderate-intensity physical activity such as walking (5), the greater magnitude of behavior change for Wheeling Walks would suggest more postcampaign caloric expenditure. But even small populationwide changes in behavior have significant implications for reducing weight gain and the prevalence of obesity. In fact, the average American is gaining 0.5 to 1.0 kg per year (32). Increasing daily walking by approximately 5 minutes would account for 15 kcal, which would neutralize annual weight gain.

Preintervention to postintervention awareness of the BC Walks campaign among Broome County telephone respondents increased from 36% to 78%. No change of awareness was observed in the comparison community, indicating that the Broome County community was aware of the campaign message. BC Walks purchased the same amount of television and radio gross rating points as Wheeling Walks. Awareness of BC Walks was nearly as high as awareness of Wheeling Walks, with 78% of Broome County respondents reporting exposure to BC Walks and 90% of Wheeling respondents reporting exposure to Wheeling Walks.

BC Walks clearly created interest among media gatekeepers and generated 168 news stories and promotions on the campaign message during the 8 weeks of the campaign. During the first 8 weeks of Wheeling Walks, 280 television, radio, and print news stories were generated (25). A further comparison of campaign awareness based on type of media (Table 5) shows that although percentage of respondents reporting overall exposure was lower for BC Walks (78%) compared with Wheeling Walks (90%), the percentages are similar in both communities for reported exposure to television and radio. An exception is exposure to local news coverage: 81% of respondents in Wheeling indicated that they saw or heard news stories about the campaign; only 43% of the respondents in Broome County reported such exposure.

Several factors may have contributed to the diminished campaign impact of BC Walks compared with Wheeling Walks. Although the overall messages were the same in both campaigns (walk 30 minutes or more daily), a mass media telethon event held in Broome County in conjunction with the campaign kickoff to generate initial enthusiasm for walking encouraged callers to pledge to walk at least 10 minutes per day. The event was successful, with 10,800 community members making a pledge. The 10-minute message, however, may have created ambiguity. Although the 10-minute pledge was not inconsistent with the paid advertising campaign, which encouraged sedentary people to begin with 10 minutes, increase to 20 minutes, and then to walk 30 minutes daily, the 10,800 pledge respondents may have heard only the 10-minute daily message. Campaign implementation staff must ensure that all messages are designed and delivered to promote an established norm.

The fewer newspaper stories may also have lessened the impact of the campaign. Print news media may have a stronger impact on local community behavior than other media. The Broome County television market is more than double (890,890,573 viewers) that of Wheeling (418,170). Smaller media markets are easier to penetrate because they have fewer competing activities (20). In addition, the Wheeling Walks intervention staff spent 2 years planning implementation, which included a 12-week participatory process that increased social capital within the community (24). The participatory planning process resulted in the formation of community task forces to address barriers and resources associated with the development of Wheeling Walks. The Wheeling task forces were involved in conducting community focus groups as part of the formative research to develop campaign television, radio, and print advertisements. The participatory planning group developed into a community advisory commission (24). The mayor of Wheeling sanctioned the Walkable Wheeling Task Force, which he charged with providing semiannual progress reports on improvements to the walkability of the physical environment. The Wheeling community developed a great sense of ownership and engagement in the intervention. Although BC Walks had a communitywide commitment to increase physical activity and the regional metropolitan planning organization was actively involved, the process in Broome County was not as formalized at the local level as it was in Wheeling. (There was no participatory planning process, task force, community advisory commission, or environmental task force sanctioned by the mayor.) It is always a challenge to replicate programs and have the same efficacy achieved by the program originators.

Broome County was chosen as the intervention site because United Health Services Hospitals had established involvement with state and local stakeholders to promote physical activity in the region. Having only one intervention community and no random assignment limits the generalizability of the results. However, most of the large communitywide interventions reviewed for this study had similar methodologic limitations (8-11).

The BC Walks study, as in Welch Walks (a program in a West Virginia community even smaller than Wheeling [27]) and Wheeling Walks, demonstrated an increase in walking among participants (the purpose of the intervention), but it did not demonstrate an increase in moderate or vigorous activity. The positive change in walking behavior provides some evidence for the impact of targeted campaigns in smaller communities.

Future research should examine the potential effective mediators of change in social marketing campaigns. Additional trials should then empirically test the efficacy of interventions targeting the mediators. Public health intervention funding is limited and needs to be spent judiciously. Simultaneous or subsequent replication of small campaigns in similar communities will permit the empirical determination of relative cost-effectiveness of changing physical activity behavior and policy and environmental supports within communities.