Community-Based Priorities for Improving Nutrition and Physical Activity in Childhood
Overweight among America's youth has prompted a large response from foundations, government, and private organizations to support programmatic interventions. The architecture for many of these programs was derived from “experts,” whereas the perspective of families, and communities—those most affected and most instrumental in altering behavior—is rarely the driving force. Shaping America's Youth (SAY) was established to assess programs that target nutrition and physical activity and to promote the necessary family and community input. In a 2004 report, SAY documented how community efforts are motivated, funded, structured, and evaluated. It identified discordance between that effort and the opinions of experts. To ensure that the voices of families and communities are integrated into such local and national policies and programs, SAY initiated a unique series of 5-day-long town meetings, input from which was independently statistically analyzed. Across a range of demographics, the results indicated that participants perceive the barriers and solutions similarly. There was broad agreement that the family has primary responsibility, starting with a need to focus on improved quality and duration of family time directed at nutrition and activity. Concurrently they identified needed actions from external sources, including clear and consistent nutrition information; ready access to healthy foods; and a built environment that promotes physical activity. Rather than one-dimensional or governmental solutions, they expressed a need for community-based partnerships integrating health care, education, environment, government, and business. Although this citizen-engagement process did not identify specific actions, it defined basic steps that communities must integrate into future approaches.
- childhood obesity
- child health
- citizen priorities
- consumer participation
- child nutrition
- physical activity
Overweight in childhood has become a major focus of public health efforts in the United States. Numerous studies and reports from experts have provided extensive assessment of the scope of what is recognized as an epidemic and have produced an array of recommendations on how to improve it. However, despite societal and health professional awareness and concern, there is little evidence of progress in reversing the trend of excess weight among youth.1 Although the most recent assessment by the Centers for Disease Control and Prevention reported that the prevalence of high BMI in children has remained stable for the past 10 years,2 it has not decreased. The fact remains that nearly one-third of children and adolescents in this country are at ≥85th percentile of BMI for age. In addition to the long-term consequences that excess weight in childhood imposes on the physical and mental health of the affected individuals, its impact on future health care costs, productivity, and longevity is profound and likely exceeds that of any health issue the United States has ever confronted.3
The marked increase in the prevalence of childhood obesity over the past 3 decades has occurred in parallel with the changing social structure of our nation.1,3 The increasing number of single-parent families and parents working outside the home; changes in dietary practices including what, when, and where we eat; the electronic revolution that has lessened children's physical activities; the time and budget constraints on our teachers and schools; the wider marketing and availability of foods in general and those of low nutritional value in particular; and changes in our communities, transportation, and the built environment that limit outdoor and leisure-time pursuits have all contributed to the epidemic of childhood obesity.
As a result of concerns in both the public and private sectors regarding the increase in childhood obesity, the Shaping America's Youth (SAY) initiative was created in 2003 with input from the Office of the US Surgeon General, the American Academy of Pediatrics, and several major corporations (Appendix). SAY was established to assess ongoing efforts to reduce excess weight in young people to identify the most effective means of prevention and intervention and of their implementation. SAY's objectives emerged from the belief that, at its roots, this health crisis is a problem of social norms and behavior that have led to the development of unhealthy nutritional and physical activity habits from the earliest years of life. Members of SAY postulated that to be successful, recommendations for lifestyle changes must incorporate the input of those who are most pivotal in instituting the changes: parents, families, and community members. SAY's review of the strategies of various expert panels revealed that the voices, opinions, and perspectives of these individuals were seldom considered. We therefore charted 2 initial programmatic components: (1) a nationwide survey of community-based funding and service programs that target childhood and adolescent nutrition and/or physical activity; and (2) a national citizen-engagement process based on the historical precedent of early-American town meetings and updated for the 21st century to gather a cross-section of citizen voices.
The program survey was undertaken as the first step to define America's effort at the community level (ie, a snapshot of how motivated organizations and individuals were targeting the problem of overweight youth).4 As a natural extension of the program survey, the town-meeting process was implemented as a means to give a voice to those at the grassroots level to determine how the individuals in families and in the community perceive the barriers to and solutions for improved nutrition and physical activity for their children.
Data from the Centers for Disease Control and Prevention strongly indicate that excessive weight is a health concern evident from infancy and well established by the time most children enter school.5 In contrast, evidence gathered in the SAY survey,4 which ascertained comprehensive data from 1100 programs, revealed that few were targeting the preschool years or the parents and caregivers who influence early childhood development. For example, lack of family involvement was identified as the largest barrier to success for most programs, yet only 50% of the programs surveyed directly involved the family of the affected child, and <10% specifically engaged families. Thus, the collective evidence available by 2005 suggested a critical disconnect between where the emphasis could do the most good and where it was actually being directed.
The SAY survey also revealed that the majority of ongoing programs lacked the funding, duration, and/or mechanisms necessary to document measurable outcomes, ensure sustainability, and promote collaborations. The limitations faced by most of the programs directed at the childhood obesity crisis are particularly concerning in light of the reported $3 to $5 billion that were expended on these programs in 2003–2005, a figure that is likely to be substantially higher in the future.
To address its goal of engaging the public in reversing childhood obesity, SAY conducted town meetings in 5 locations across the United States. Participants in the meetings were drawn from each local community, and every effort was made to include a representative cross-section of the area's population makeup. On the basis of the belief that families and communities have the greatest influence on and responsibility for children's development of healthy behaviors, we sought to engage these individuals in determining the barriers to better nutrition and physical activity and proposing and evaluating possible means of overcoming them. In this report we describe the findings from these town meetings. These findings add an essential element to America's discussion surrounding how we can most effectively meet this public health challenge from the viewpoint of those who are most instrumental in doing so.
Community-Perspective Data Collection Using a Town-Meeting Process
SAY issued a nationwide request for proposals to identify an organization with experience in citizen-engagement initiatives that generate usable data for setting public policy. After reviewing 7 proposals, the organization AmericaSpeaks was selected to partner with SAY to organize and conduct the town meetings. AmericaSpeaks is a nonpartisan, nonprofit organization founded in 1995 to involve citizens in the public decision-making process by creating an opportunity for them to give their civic leaders direct, substantive feedback on key issues.
The concept of involving the general public in health care–related decision-making is being increasingly promoted and practiced in various forms.6,–,8 Focus groups, town meetings, group interviews, and community-based participatory research are all formats for acquiring input from lay people on topics ranging from personal, local health care issues to national health care reform. Although the designs used in each of these formats vary widely, SAY opted to use the 21st Century Town Meeting model developed by the internationally recognized organization AmericaSpeaks based on its 15 years of experience and proven success in other public policy arenas.9,10
A primary goal of the town-meeting process was to ensure neutrality, fairness, and balance in all stages, including participant recruitment, meeting content and presentation, and data analysis. To this end, SAY and AmericaSpeaks committed to the following set of core objectives for the development of an effective citizen dialogue.
To ensure that the meeting outcomes represented the views of the entire community, the participants had to reflect the rich diversity of their region. In each city, a local coalition of organizations hosted the meeting and worked with SAY and AmericaSpeaks to tailor a site-specific recruitment strategy to meet demographic targets (for age, gender, race, etc).
Meeting participants were provided with (1) a comprehensive 30-page “participant guide” that included descriptions of the SAY initiative and town-meeting objectives and factual, objective information about the crisis of excess weight among American youth, and (2) a 4-page “issue overview” that provided a simplified depiction of the primary issues related to childhood obesity. These documents included both national and site-specific data (available as PDF files at www.shapingamericasyouth.org).
The town meetings were intended to ensure that every attendee played an active role in the deliberations and that every viewpoint was heard. The meeting agenda and presentation were designed to promote participation and allow all individuals to openly discuss their ideas and opinions.
Five town meetings were held in 2006 and 2007 in Memphis, Tennessee; Dallas, Texas; Philadelphia, Pennsylvania; Chicago, Illinois; and the state of Iowa. Their selection was based on the presence of a regional coalition or organization committed to the issue of childhood obesity, a commitment from regional leadership to support a town meeting, demographic diversity in the area, and the ability to meet an aggressive time line. After town meetings in 4 major metropolitan areas, the state of Iowa was chosen as the fifth meeting site to gather rural input, with Des Moines serving as the host city. Requests from a substantial number of additional cities were considered; however, budget and scheduling limitations necessitated limiting the initial round of town meetings to these 5 sites.
At each site, SAY and AmericaSpeaks worked with a local host organization and a local communications firm to develop and implement a comprehensive community-outreach campaign to promote participation in the SAY town meeting. The primary objective of the outreach efforts was to attract an audience that was demographically representative of the entire community in terms of age (≥16 years), gender, income, education, occupation, religion, and racial and ethnic composition. US census data for each site were used to establish outreach targets for each relevant demographic characteristic. The target for desired scale was established relative to the size of the local community and local assessments of how large an audience could be expected.
Outreach Strategy Development
The initial step in each city was to convene a local host committee. Each city had at least a core coalition of organizations working to address childhood obesity. These organizations were the Healthy Memphis Common Table, the Dallas Area Coalition for the Prevention of Childhood Obesity, the Childhood Origins of Disease of Adulthood (CODA) in Philadelphia, the Consortium to Lower Obesity in Chicago Children (CLOCC), and Iowans Fit for Life. The local coalitions served as the basis to form a more inclusive steering committee and host organization for the town meetings.
The host committees were primarily responsible for providing feedback on outreach strategy and messaging, recruiting participants, and developing strategies for acquiring input on local issues. To recruit the scale and diversity of participants we were seeking, a local communications firm was engaged to drive the participant recruitment.
The most effective strategy used by the local organizations was direct, face-to-face outreach at community events, conferences, health fairs, and similar gatherings. This strategy was complemented by telephone and e-mail outreach using available lists from members of the steering committees and other organizations that represent youth, parents, educators, community leaders, and service providers (Appendix). Special efforts were also made to recruit local policy-makers to participate.
Demographic information was acquired from participants at the time they registered to attend the meetings to determine if demographic diversity targets were being met and to adjust outreach strategies as needed to ensure representative participation. Weekly registration reports (increased to daily reports in the final 2 weeks before the event) and teleconferences were used to review current registrations and discuss recruitment and marketing tactics.
Local marketing of the event through both earned and paid media supported our recruitment activity. Examples of earned media include local television and radio interviews with SAY's national leaders and local steering committee members, event announcements in regional parenting magazines and newsletters, and blog posts on local news and community Web sites. Paid media included drive-time radio advertisements, announcements on public radio, and newspaper advertisements.
SAY and AmericaSpeaks developed a detailed agenda and day-long script that served as the template for all 5 meetings. These were used to track the themes, suggested actions, and recommendations derived from each meeting and to compare them between and across meetings. The meeting agenda provided adequate time periods for information-sharing, discussion, and feedback to achieve the desired outcomes and to ensure participants' continued engagement throughout the day (eg, exercise breaks, quizzes, raffles). The meetings concluded with a site-specific component in which the local coalition presented its own set of discussion and polling questions directly related to a local agenda, policy, or action plan. As part of the meeting process, the local coalition provided regional data for the participant guide as well as handouts and worksheets that focused on a local agenda.
For each meeting, the room was set in round tables that seated 10 participants each. A keypad polling device was provided to each participant for his or her anonymous and confidential input. A laptop computer was used at each table to instantly transmit the ideas generated at that table to a central computer. This format ensured that every idea entered into the computer became part of the permanent record for the meeting. Two moderators from AmericaSpeaks directed all 5 town meetings, which began with introductions of the guest speakers and visiting dignitaries, who varied according to meeting but collectively included state governors, city mayors, city and county council members, state and city health commissioners, and the US Surgeon General and state senators via videotape. The moderators introduced the key data and context surrounding each topic area, explained how the discussion would proceed, and reviewed recommendations after each round of questions.
A trained facilitator worked at each table to ensure that all individuals were engaged, had a meaningful opportunity to participate, and that the group remained on task. Facilitators were recruited both from local networks and from the AmericaSpeaks national network. In general, we selected people who had table-facilitation experience of at least 2 years in either his or her own professions or outside activities. All facilitators received 2 hours of training before the town meetings during which the meeting design was explained and their roles were reviewed. Facilitators were trained to stay neutral, to be inclusive, and to help the group look for synergy and common ground. Postmeeting feedback indicated that the vast majority of facilitators did either a good or excellent and well-balanced job at the tables.
After the moderators outlined a question for discussion, 10 to 30 minutes were allotted for the table discussion; collective responses from those at each table were entered into that table's computer by the group's designated secretary and submitted to the central servers. Collected comments from each table were categorized into common themes by the “theme team,” a group of local individuals who were selected for their backgrounds related to childhood weight, nutrition, physical activity, communications, and generation of public policy. The team reviewed the individual comments from the each of the tables, categorized each comment, and then reported the common ideas, or themes, back to the participants via jumbo video screens. After considering the categorical themes that had been identified, participants voted on their keypads for the theme they thought was most important. Polling results were reported instantly to the entire group on the video screens.
The participants' vote tally generated a prioritized list of barriers to and steps for improving childhood nutrition and physical activity within each of the family and community elements considered. Each keypad carried a unique identifier so that a participant's vote could be combined with the demographic information that he or she had entered at the beginning of the meeting. Thus, as discussed below, the demographic information identified with each keypad was combined with polling data from that keypad for statistical analysis of the data. That analysis permitted the within- and across-sites comparisons necessary to assess the broad applicability of the data.
At each town meeting, participants discussed the following topics sequentially.
What actions can families take to improve nutrition and increase activity among their children? One-third of the room focused on children prenatally and 0- to 5-year-olds, one-third focused on 6- to 11-year-olds, and one-third focused on 12- to 19-year-olds.
Barriers to Family Actions
What barriers impede actions that families might take? These barriers were discussed according to the 3 age groups listed above.
In what ways can stakeholder groups support families to positively affect children's behaviors? Stakeholders were defined as (1) educators and schools, (2) community programs and leaders, (3) members of the health care community, (4) legislators and policy-makers, and (5) corporations and the private sector. The room was divided into 5 equal sections for this discussion, with each section assigned 1 stakeholder group.
How can local stakeholder groups collaborate on issues of youth overweight to achieve a greater impact than they can achieve individually?
What actions can national leaders and organizations take to improve nutrition and increase activity among children? The room was divided again into the same 5 stakeholder groups as for the “supporting families” discussion.
Next Steps/Local Action
In the final stage of the meeting, participants discussed topics linked to either developing or providing feedback on a local agenda, policy framework, or action plan.
At the end of the meeting, interested participants were asked to make personal commitments as to what they would do to combat youth obesity. All participants were asked to respond to a series of questions evaluating the meeting. As they exited, each participant received a 4-page summary report that was prepared throughout the day that detailed the meeting content, participant demographics, and the predominant themes and results.
From the outset, we realized that a statistical analysis of a database of the magnitude we envisioned would present an unprecedented challenge. Specifically, the analysis team was confronted with developing an appropriate statistical strategy for organizing, summarizing, and examining the data to identify relationships among the tens of thousands of bits of data this scientific process collected. The analysis had to be done in a way that distinguished between potentially important and unimportant factors and trends while preserving the uniqueness of the sites and the points of view of different groups of participants.
To meet this challenge, we relied on statistical techniques that allowed us to (1) simultaneously examine and control multiple factors, (2) distinguish between factors that were found to be empirically important or extraneous, and (3) determine if an observed relationship was stronger than what was likely to arise by chance. We also relied on commonly used statistical methods such as multiple linear regression, logistic regression, and analysis of variance to reliably identify similarities and differences among sites and groups. In addition, we recognized that results had to be presented in a way that acknowledged that each site was its own descriptive study rather than an experiment with a control group and therefore observed that relationships were not necessarily causal.
The initial data analysis was conducted after the town meetings held in the first 4 sites and before the Iowa meeting. That initial analysis used data gathered from 543 Chicago, 467 Dallas, 880 Memphis, and 333 Philadelphia participants. These data included age, gender, family income, and racial/ethnic group. The database also indicated which strategies the participants felt would be most effective for improving nutrition and physical activity to reduce obesity in children. In addition, the data captured the language families and community entities used in their discussion or consideration of the issues surrounding childhood obesity. Two concurrent data analyses were conducted for reliability, and a deeper look at both suggested activities and ranked priorities: a quantitative data analysis of priority polling data based on categorized themes and a qualitative content analysis of raw data gathered from each tables' collective responses.
An initial summary multivariate analysis of the polling data was performed for the first 4 sites. Each participant provided substantial demographic information through his or her individual handheld keypads used for voting on topic priorities. Thus, it was possible to first determine if, within each city, the priorities varied on the basis of a given population characteristic. That analysis revealed that individual demographic characteristics did not predict the results of the polling within any city.
The relevant questions, similar across sites, covered 3 areas: participant demographics; perceptions of the nature of the childhood obesity problem and motivation to improve it; and evaluations of the perceived effectiveness of a selected set of stakeholder actions for combating childhood obesity. The evaluation questions addressed the following societal groups whose potential action was targeted:
families with children aged 0 to 5 years;
families with children aged 6 to 11 years;
families with children aged 12 to 19 years;
health care professionals;
business leaders; and
legislators and policy-makers.
In addition to these societal elements, the evaluation questions incorporated national initiatives.
To examine whether men and women within a site tended to have the same or different preferences for the strategies on which they voted, the analysis team constructed a table with 1 row for each of the suggested activities and 1 column for each group. The cells in this table showed the percentage of the group that voted for the activity. Thus, each column summed to 100%. There was a separate table for each of the 4 sites and the targeted societal groups, because the suggested actions at 1 site were not necessarily the same as those suggested at other sites or not expressed in exactly the same way.
For the first 4 cities, the analysis required the construction of 36 tables for gender and another 36 for race, for a total of 72 data matrices. The correlation coefficient between men and women was used as a measure of the degree to which they agreed with each other regarding the likely relative effectiveness of the different strategies. Coefficient α was used to obtain a comparable measure of agreement among racial groups. For both of these indices, the higher the value, the greater the consistency in the preferences of the groups being compared. Values over 0.85 indicate a high degree of agreement.
The correlation analyses were run separately according to site, because the categorized themes of suggested activities varied somewhat across sites. To address this feature of the data and allow for more in-depth analyses, SAY staff unified the suggestions from the different sites into a common set applicable to all sites. The resulting cross-walk between this staff-developed common set of suggestions and the site-specific set was used to create a database that contained each participant's first choice of the possible actions listed in the common set for a given societal group. This database also contained the participant's demographic characteristics. This exercise resulted in identifying 51 actions common to at least 2 sites, 23 of which were common to all 4 sites.
Separate multivariate logistic regression models were computed for each of the 51 actions. The outcome variable in each model was whether the action was the participant's first choice (coded 1 or 0, respectively). The independent variables were race, site, gender, age, and income level. For race, “dummy” variables were used to classify a respondent as black, white, or Hispanic. The reference group was “all others” (most of whom did not indicate their race). Chicago was the reference-group site for all the models except when the action was not an option identified in the Chicago data set.
The size and algebraic sign of the coefficients in these models indicate the degree to which a participant's race, gender, age, income level, or site were related to whether participants did or did not prefer a given action. A statistically significant coefficient indicates that the groups differed more in their preferences than was likely to arise by chance alone. Because of missing-data problems, noted below, only 45 of the 51 models could be estimated.
After completion of the Des Moines meeting, the corresponding Iowa-specific data set was then merged with that of the first 4 sites. Rather than repeating the polling analysis, which had already statistically established the relative uniformity across numerous demographic variables in the first 4 sites, the analysis used the content analysis of all 5 sites to determine if the state of Iowa differed in its priorities from Memphis, Dallas, Philadelphia, and Chicago.
The multivariate analysis revealed that the barriers identified and the solutions proposed from 4 distinct venues were generalizable to the US population, a finding that supported the pooling of the ≥14 000 inputs entered from table discussions at the 5 sites in response to the comparable questions posed at each venue. The pooling of responses allowed for a formal content analysis of all the comments that were provided.
In analyzing and summarizing participants' responses, the analytical team first reviewed the responses and the summary themes for each question to develop an initial framework for coding the individual responses included in each of the meeting transcripts. Using this initial coding scheme, the team reviewed each response recorded in the meeting transcripts to assign it to the appropriate category. This process generally identified other common categories of responses, which then were incorporated into the scheme to continue the coding process. In the course of coding the responses to individual questions, a minority of responses were either not decipherable or did not seem to be related to the question that had been posed. In some instances a single response incorporated more than 1 idea and, therefore, could be coded into more than 1 category. In these cases, multiple codes were assigned to the same response. In summarizing the data, the analytical team calculated the total number of responses in each category divided by the total number of responses recorded to that question to define the percentage of responses in each category. This process allowed us to calculate the rank order of the responses fitting a defined category for a response to each question posed at each site.
The magnitude of this data set is such that only selected tables are included in this publication. The analysis of the full data set, tables generated, and all 14 000+ primary responses from each meeting table are available online at www.shapingamericasyouth.org.
Demographic characteristics and community roles of the participants at each site are shown in Tables 1 and 2. Although there was substantial variability for selected characteristics between sites, we were able to account for them by using an analytical approach. The demographics of the participants adequately reflected several of the known variables associated with excess weight in children. In general, the participants in each city were reasonably representative of the metropolitan area; for each case in which there was overrepresentation of a demographic group, it tended to favor factors associated with an increased risk for excess weight. Thus, the participants represented a good cross-section of that portion of the US population for whom solving this health crisis is particularly important. As noted in Table 2, there was significant representation from adolescents and young adults, aged 15 to 21, for whom the challenge of weight control has even greater immediacy. Their input was particularly valuable in considering the effectiveness of current school-based initiatives. Many of the participants identified with more than 1 role in the community; thus, the total percentages exceed 100.
Table 3 depicts the gender and race coefficients that were generated by the multivariate analysis of the polling data. Notable are the unusually high coefficients among the different gender and racial groups in all 9 targeted societal groups at each site. The major exception to this trend was at Philadelphia, where there were noticeable disagreements (much lower coefficients), especially among gender groups in several polls that targeted specific societal groups.
Table 4 shows the coefficients in the 45 models that could be estimated. For example, the first row of this table indicates that after holding the other variables in the model constant, Memphis participants were more likely than Chicago participants to select “Mothers should eat healthy and be physically active during pregnancy” as their top choice for families with 0- to 5-year-old children, whereas white participants (regardless of site) were statistically significantly less likely to select this option relative to the reference group of participants (ie, those who did not identify themselves as white, black, or Hispanic).
Slightly fewer than 50 of the nearly 400 coefficients in Table 3 were statistically significant, but ∼20 would be significant by chance alone. Thus, taken as a whole, the participants' background characteristics were generally not systematically related to their preferred actions. However, a disproportionately large number of the significant coefficients were related to site differences, whereas there was only 1 significant coefficient in the gender column and none in the income column. Viewed collectively, these findings indicate that location was generally more closely related to the participants' preferred actions than was their race, gender, or income.
The Iowa site was added to the project in 2007. Some of the scripted queries at this site differed slightly from those used at the other sites. Although identical scripts were used at all sites regarding barriers and solutions for family action, Iowa participants addressed a new series of questions, “stakeholder actions in support of families,” that were identified in the analysis of the initial 4 sites. Under each general stakeholder action, participants were provided 3 to 4 specific, theoretical examples of an action that a stakeholder might take. Participants evaluated the examples or proposed alternative actions they deemed more likely to be effective in supporting the family unit. This modification in the day-long script for Iowa was substituted for the open-ended discussion of stakeholder actions used in the earlier town meetings. The discussion regarding “creating partnerships” was eliminated from the Iowa meeting to provide additional time for discussion of specific actions by stakeholder groups.
This adjustment in the second half of the Iowa town-meeting format and script was made to assess whether it was feasible to engage the participants over a more specific or detailed list of theoretical actions by stakeholders. This was done in anticipation of such a line of engagement being used in a new round of town meetings. Specifically, the putative second round of national town meetings would target the identification of specific action steps for the families and their communities to take that would be deemed responsive to the general action concepts identified in SAY's initial round of national town meetings. When the Des Moines script was identical to that of the other 4 sites, the data generated were incorporated in the overall analysis of “family actions and national actions.”
Table 5 summarizes the data regarding the number of suggestions per site about “barriers to family action” as they relate to nutrition and physical activity. In this table and Tables 6 through 9, “n” indicates the number of suggestions that were made at each site in response to each of the open-ended questions posed to the participants. In each cell of Table 5 the percentage of those suggestions that relate to the 4 most prevalent suggestions is recorded. What is evident in Table 5 and follows through in Tables 6 through 9 is the consistency of the rank order across the top 4 suggestions of the critical barriers that were identified at the 5 sites despite the diversity of participants that each town meeting included.
Table 6 provides similar data about the number of suggestions per site about “family actions” that can be made to improve nutrition and physical activity for each of the 4 target age groups. Once again this table shows the number of suggestions that were made, and the cells show the percentage of those suggestions that relate to the 4 most common types of suggestions. For example, Table 6 shows that, as a group, the Memphis participants made 238 suggestions regarding how families can improve nutrition for 0- to 5-year-old children, and 28% of those involved families providing healthy snacks and meals. Table 7 provides a comparable presentation of the data regarding “stakeholder support of family actions” as was presented in Tables 5 and 6. The top 4 actions for stakeholders to take are again provided. As in Tables 5 and 6, the consistency of the rank order of the percentage that a given suggestion was noted across the 5 sites is evident. This set of questions was not presented at the Iowa meeting.
Table 8 summarizes the national actions that stakeholders might take to improve nutrition and physical activity. The format is identical to that of Tables 5 through 7 except that in Memphis and Dallas the question was posed separately for nutrition and for physical activity but combined as 1 question for the other sites. Once again the rank order and percentages are remarkably consistent across the different regional sites and their respective participants.
Table 9 summarizes the “creating partnerships” recommendations from the first 4 sites. The Iowa state-wide town meeting was not queried on this point. More than the top 4 categories are presented, in part, to document that a number of national partnership concepts that have been frequently touted in the past had little or no support from these participants. Once again, however, there was good agreement on the top 4 partnerships and some support for the next 6, although minimal in most instances.
The data in Tables 5, 6, and 8 suggest that the Iowa participants demonstrated good agreement with the participants at the other sites regarding which strategies would work best. Specifically, the rank ordering of their 4 most common suggestions in each set tended to coincide with the rank ordering of the most popular suggestions in that set at the other sites. This finding was consistent with the findings of the initial data analysis that across various demographic and regional factors, the town meetings identified a common set of barriers to and actions for improving nutrition and physical activity among children and adolescents in America.
The data presented here are descriptive of the views expressed by town-meeting participants during their deliberations with their tablemates and through the votes they cast via their polling devices. These participants are best characterized as a convenience sample of teenagers and adults who are sufficiently interested in the problem of childhood obesity to devote a full day to discussing its resolution.
There may be subtle factors at work that could affect and homogenize the participants' suggestions, such as what the session leaders said about the problem before the participants discussed and voted or how the theme team's list of suggestions may have been influenced by the views of the table recorders and team members through whom all the suggestions were filtered, translated, and expressed. Moreover, there were no independent checks on the consistency of how table recorders or theme team members worded the participants' suggestions or how their personal views may have seeped into the items on which the participants ultimately voted.
Missing data must also be considered. For example, ∼3% of the participants did not report their gender, race, age, or income. In addition, the number of participants who cast a vote varied across the different polls, with more attrition as the day wore on. Nevertheless, the generally high degree of agreement among participants regardless of site and background characteristics indicates that the views they expressed are applicable to the US population in general. Finally, the relative popularity of a suggestion may not be related to its effectiveness in reducing the childhood obesity problem. That could only be determined by outcome measures of any intervention that a community or element of a community might elect to institute.
Participants' Assessment of the Town-Meeting Process
At the conclusion of the meeting, participants were asked several questions to assess the quality of the meeting (Table 10) and whether they thought it would lead to change in the community. Their response to the latter was compared with a similar query to the audience at the beginning of the meeting. At the end, there was a consistent and substantial increase in the number of respondents who stated that they were more optimistic that there would be positive change in their community as a result of the meeting. When asked to rate the overall quality of the meetings, >92% at all sites graded the meeting as either “good” or “excellent.”
Citizens' Perception of the Childhood Obesity Crisis
Several important general insights emerged from this effort that should not be lost in the mass of data that were accumulated. First, through their responses participants indicated that they clearly understand that America's youth are at serious risk from poor nutrition and insufficient physical activity. Second, their responses maintained a positive tone, avoiding a simplistic focus on outside sources as the cause of the childhood weight crisis. Third, their recommendations indicated that they understood that solutions need to focus on the family and elements of the community in which they reside (ie, solutions rest largely with the family with support from the community to act successfully). Fourth, consistent with the previous point, they did not perceive that a universal approach conceived and implemented by government or some other larger element(s) of society was critical to reversal of this crisis. They recognized, however, that there were steps that could be taken at national or regional levels to facilitate the success of family and community efforts. Fifth, by their responses to the queries regarding families with children 0 to 5 years old, they acknowledged the long-term importance of the earliest years in the life of a child, including pregnancy and infancy, in developing habits that promote healthy weight.
Family and Community Priorities
The summary tables (Tables 5–9) of the content analysis demonstrate where the priorities for families and communities lie. Perhaps obvious, but nevertheless important to note, the participants view the family's role as primary and see the school setting, where their children spend the majority of their out-of-home time, as the secondary area for improvement. Two core issues run through the characterization of the barriers and solutions for the family to improve both the nutrition and physical activity of their children. One issue is time. Time pressures dominated the participants' characterization of the family's inability to ensure healthy eating and physical activity. The second issue is access, specifically access to effective nutrition information/education, affordable, nutritious food, and a built environment that promotes families' efforts to be physically active.
Much of what families expect from stakeholders appropriately reflects their recognition that they need to be more vigilant of what they eat and in what setting. The repeated concern about access to nutritional education throughout all years of a family's life connected with the recognition that they need to shop together and prepare together for their meals is a logical flow of priorities. Likewise, there is the common recognition that families need to be more physically active with both the educational and work settings supporting that effort. They want the built environment to enable safe family physical activity and the work and school settings to supplement the family's physical activities.
The participants' input consistently reinforced the notion that the family unit is at the center of improving both nutrition and physical activity in children. This finding validates the proposition that a greater effort to support the family in its challenge to improve childhood weight should be at the core of any community-based strategy. Recognition of family and community as the core of all efforts to reverse the excess weight trends in the nation's youth was emphasized in the recent Institute of Medicine's report on the prevention of childhood obesity.11 The SAY program survey, 2003–2004, documented that the family was not typically engaged in interventions directed at either nutrition and/or physical activity in the community.4 The critical importance of correcting that deficiency in the future is highlighted by the concern expressed by town-meeting participants 3 years later.
The “Wisdom of Crowds”
As noted at the outset, experts from a wide range of professional backgrounds and perspectives have attempted to address both the causes and corrections for overweight children. The collective views of these authorities have prompted the initiation of well-intentioned programs to reverse the increasing trend in childhood overweight. The scope, structure, and questionable efficacy of these efforts were documented in the summary report of the SAY 2004 survey.4 As that document suggested, the country's effort was largely driven by the media coverage of experts and expert groups. They were not grounded on the perspective of the community or family. In fact, the programs rarely involved the family unit or promoted structural changes in the family environment that might have an impact on nutrition or physical activity, and rarely were their outcomes or efficacy documented.
Table 11 highlights several areas in which the approaches to reversing childhood overweight differ between individuals at the grassroots level and experts at a national level. These are based, in part, on the priorities the participants identified for national actions compared with recommended actions of many reports from professional organizations. The comparison is also predicated on the tone and content of the priorities for family and the stakeholder actions the participants identified.
Feasibility of the Town-Meeting Process
Our town-meeting experience as described here demonstrates the feasibility of this process. With the assistance of AmericaSpeaks and local host organizations, SAY successfully recruited a large number of participants at multiple venues, sustained their commitment to a day-long interactive meeting, and acquired data that could be analyzed to provide critical, consistent input from the general population. Although covered in greater detail in “Methods,” it is beyond the scope of this article to provide extensive details of executing such a series of national town meetings.
Given the open-ended nature of the questions posed, it might have been anticipated that the differing demographics of the 5 sites would not have yielded consistency in responses as to the perceived priorities for action within families and communities. However, the consistency across the range of questions and sites is evident in the statistical analysis. Critical to that success was the application of relatively new technology, which permitted each participant to vote on the priorities for action that the table discussions generated. Future endeavors to use citizen engagement in resolving this public health crisis should benefit even more from advances in communication technology.
This experience of the SAY town-meeting process offers a compelling argument for the use of available technology to convene citizens in live, virtual (online), or a combination of the 2 formats to establish citizen-based priorities directed at solving national problems. Essential to our success for independently identifying grassroots' priorities, and future endeavors along these lines, was the commitment to and execution of an independent statistical analysis of the data generated from the town meetings. Only such a commitment will ensure that future public policy decision-making is based, at least equally, on verifiable citizens' priorities rather than solely on the opinions of the “experts,” as has typically been the case.
The data that SAY acquired from citizens in 5 US communities provide clear evidence of what families and their communities consider necessary to reverse the increasing trend in overweight in America's youth. Fundamentally, the town-meeting process identified uniform targets on which local, state, and national leadership should focus. This process has identified a number of actions that town-meeting participants agreed were fundamental to developing lifelong, healthy nutrition and physical activity habits in our children and youth.
From this research effort it is clear that Americans understand the significance and critical nature of the childhood weight crisis, and across a wide range of demographic variables, they similarly perceive the barriers and solutions. Families are not seeking one-dimensional approaches or government solutions alone. They want help from community-based partnerships that integrate the health, education, environment, government, nonprofit, and business sectors to assist them in making healthier choices and creating healthier lifestyles. Fundamentally, on the basis of the statistical analysis, the results of our town-meeting process identified uniform targets on which community, regional, and national leadership should focus. For families, this process has identified a number of actions that a cross-section of Americans agreed were fundamental to developing lifelong, healthy nutrition and physical activity habits in our children and youth. Many of these actions could be implemented by families immediately through changes in behavior they can control. Thus, disseminating and reinforcing these principles to families is an essential next step. However, for families to implement many others, the identified barriers need to be addressed. Other elements of the community either locally and/or nationally need to take action to remove or lessen those barriers.
Even with the basic delineation of the barriers and solutions the town-meeting data provide, each community will need to determine for itself how best to address its individual priorities. To that end, in the article “Future Directions: A Community-Based Approach”12 we offer a perspective on how best to build on the evidence this national town-meeting data has provided. In that article we outline steps intended to ensure that the basic priorities that the SAY town-meeting process identified are addressed at all levels of society. Failure to consider the “roadmap” these citizens have provided would be a lost opportunity for future well-intended efforts to reverse the crisis of excess weight in childhood.
- Accepted August 9, 2010.
- Address correspondence to David A. McCarron, MD, FACP, McCarron Group LLC, 120 NW Ninth Ave, Suite 206, Portland, OR 97209. E-mail:
Ms Richartz's current affiliation is Institute for Healthcare Improvement, Cambridge, MA.
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
- SAY =
- Shaping America's Youth
- 1.↵Trust for America's Health. F as in fat: how obesity policies are failing in America, 2008. Available at: http://healthyamericans.org/reports/obesity2008. Accessed January 22, 2010
- 3.↵Institute of Medicine. Preventing Childhood Obesity: Health in the Balance. National Academies Press: Washington, DC; 2005
- 4.↵Shaping America's Youth. Summary report September 2004: national survey and registry of programs addressing childhood physical inactivity and excess weight. Available at: http://shapingamericasyouth.org/summaryreportKMD_PDFhr.pdf?cid=101. Accessed January 22, 2010
- Farquhar SA,
- Parker EA,
- Schulz AJ,
- Israel BA
- 9.↵AmericaSpeaks–Engaging Citizens in Governance. Home page. Available at: http://americaspeaks.org. Accessed April 4, 2010
- Lukensmeyer CJ,
- Brigham S
- 11.↵Institute of Medicine. Strategies to prevent childhood obesity. Available at: www.rwjf.org/publichealth/product.jsp?id=47908. Accessed April 4, 2010
- Kessel SS,
- McCarron DA
- Copyright © 2010 by the American Academy of Pediatrics