Adopting and implementing nutrition guidelines in recreational facilities: Public and private sector roles. A multiple case study

Data generation and analysis were completed concurrently to permit exploration of emerging themes and adjustment of data gathering instruments and procedures. Mixed methods were used for purposes of complementarity and triangulation, while maintaining an overall qualitative drive. When mixed methods are used for the purposes of triangulation and complementarity (ie. component designs), the different methods typically remain independent during data collection and analysis, and are integrated during interpretation [33]. Accordingly, each data source was first analysed independently by a single investigator as described below. A case study database was established to organize and document the chain of evidence, and thorough records of the data gathering and analytical process were also maintained [34, 35].

The multiple case study was conducted in the province of Alberta, Canada. Recreational facilities are ubiquitous throughout urban and rural Alberta. The available infrastructure varies among facilities, but typically includes swimming pools, ice arenas, soccer centres, curling rinks and/or gymnasiums. Approximately 80% of publicly funded recreational facilities in Alberta sell food as a means of generating additional revenue, most commonly through vending machines, however many facilities also contain publicly or privately operated concessions. Features of the recreational facilities included in this study are presented in Table 2. The adopter and semi-adopter cases were each limited to a single facility, however the non-adopter case included observations from four small recreational facilities where food services were managed by a single manager.

There were differences in the quality of items present within vending machines among facilities that had and had not adopted the ANGCY (Table 3). Compared to non-adopters, facilities that had adopted the ANGCY in vending machines had higher ANGCY implementation scores for food vending machines, greater availability of “choose most often” items, and their vending machine items contained fewer calories on average. Notably, all food vending machines contained few “choose most often” items, as few were available that could be sold in unrefrigerated vending machines. Nevertheless, the semi-adopter attempted to provide healthier choices in food vending machines by filling them primarily with “choose sometimes” items (77% of food items were “choose sometimes”). Beverage vending machines scored better than food machines on all measures. Comments from managers revealed why this was the case, as bottled water was a top selling item, and therefore it was in the financial interests of food vendors to place this healthy item in machines.

There were few clear differences in food environment scores between concessions that had and had not adopted the ANGCY (Table 4). The only concession that had adopted the ANGCY scored well in terms of the overall quality of its food environment, with high NEMS-R and ANGCY implementation scores, yet it provided a very limited proportion of healthy items. Most other concessions that had not adopted the ANGCY also scored highly in terms of the quality of their food environments, and had a similarly low proportion of “choose most often” items available, making them virtually indistinguishable from the adopter. An arena concession, a non-adopter with no availability of “choose most often” food items, stood out as having the poorest quality food environment, with consistently low scores on all measures.

Interviews and observations provided new dimensions for understanding quantitative findings regarding the quality of the food environment. Agreement was good overall in that managers generally recognized the need to improve the quality of the food environment within their facilities, however in some instances managers perceived they had more healthy options than they actually did. Observations revealed evidence of action on the part of adopter facilities, but also showed just how prominent unhealthy food was within these facilities that claimed to promote healthy lifestyles. Access to food was particularly high in the case of the semi-adopter, with concessions on both floors, and a large number of vending machines throughout the facility. These findings highlight the value of using multiple modes of data collection.

Facilities that had implemented the ANGCY perceived that food and beverage sales had fallen as a consequence. In the semi-adopter, the commission they collected on vending machine sales decreased by 16% from 2009 to 2010, and they anticipated a further decline of 14% in 2011. Similarly, in the full adopter facility, annual sales decreased by 17% in the concession from 2009 to 2010 and the vending machine operator estimated that revenues had declined by 20% since implementation began. Data capture systems were too limited to accurately depict the proportion of revenue declines attributable to ANGCY implementation, and which might have been due to other factors such as the economic recession or declines in facility usage. The inability to disentangle the impact of each of these factors was a barrier to greater implementation of the ANGCY, as managers assumed that increasing the proportion of “choose most often” items might further reduce profitability. Notably, annual revenues in two non-adopter concessions declined by 5% and 9%, respectively, over the same period, declines that the manager attributed to reduced facility usage. Table 5 presents industry’s perceptions of food service sales.

Comparison of sales of healthy and unhealthy items was only possible in two non-adopter concessions as others did not provide itemized sales data. In one, a pool café, sales of healthy options closely mirrored their availability, as 17% of menu items available, and 14% of items sold were “choose most often”. In the other, an arena concession, 11% of items available were “choose most often”, while 4% of items sold were “choose most often”. Of the top 15 food and beverage items sold in the pool café, only two were “choose most often” (water, juice), whereas none were “choose most often” in the arena concession. Observations made by researchers and managers in all facilities supported findings of low sales of “choose most often” items, and in particular it was noticed that students from nearby schools came to the full adopter facility at lunch to purchase the unhealthy items they could not purchase on school grounds.

We conclude by describing factors related to the implementation process. These factors were similar among adopters, however comparison to non-adopter facilities was not possible because they had not implemented the ANGCY. We are therefore unable to judge to what extent these factors may have been sufficient and/or necessary for ANGCY implementation.

Managers perceived adoption as a simple matter, whereas they described implementation as much more challenging: “Adopting was as easy as writing a policy and now the work begins with trying to find people who are able to, you know, develop programming around that and really implement it properly.” Adopters expressed frustration with the unfinished state of implementation and its apparent ineffectiveness. Healthy options had always been available, implementation of the ANGCY simply meant there were now more of them. Meanwhile, unhealthy foods continued to dominate the food landscape. Notably, it was the manager of the full adopter facility who expressed the strongest sentiment in this regard: “I really think that we’ve missed the mark with implementing… It’s one thing to have it in paper and contracts but it’s another thing to deliver it… It’s disheartening to see what doesn’t happen. Like, it’s not as simple as writing a policy and people picking it up. It just doesn’t work that way.” Managers recognized that assimilation of the ANGCY within the organization’s systems and structures would take time: “Am I jumping up and down saying we did it? No, because there’s more to do. But at this point in time, in the short-term, we can’t see that changing. Maybe long-term… There’s a lot more we can do in terms of integrating [nutrition] within our services and programs. So that will come with time.”

Implementation ultimately depended on the leadership and direction of facility managers, and therefore when time limitations prevented them from focussing their attention on the ANGCY no progress was achieved: “Other priorities haven’t let me focus any energy here in quite some time… It just sits on the back burner. And so I’d say I’m a huge barrier, if you’re looking at barriers and facilitators… Nothing is actively happening… and there is no plan to do differently or roll out anything new.” Because of these time limitations managers attempted to devolve most of the responsibility for implementation to food vendors. Managers trusted them to implement the ANGCY, and did not monitor their progress in a formal and systematic manner. Instead, they gauged the status of implementation based on their own periodic observations and anecdotal reports from customers: “Probably every couple of weeks as I’m walking through the halls, I take note of what’s in there… but it’s not a checklist. It’s just, what do I see, what do I observe. If there’s something that catches my eye that’s kind of off, we’ll address it.” This lack of monitoring likely contributed to the ambivalence surrounding ANGCY implementation, as no one could be sure whether their efforts had been worthwhile.

The analysis culminated in the development of 25 propositions, presented in Table 7.

Six to 18 months following completion of each case study all facilities confirmed that their adoption status was stable, and that no major nutrition-related changes had been made to their food services.

Collectively, findings suggest that the food environment in facilities that have adopted and implemented the ANGCY may not be superior to the food environment in facilities that have not adopted the ANGCY. Although adopters made changes to their food environment, these changes were not substantial and did not create truly healthy food environments. These results are consistent with the only other published study of the impact of government nutrition guidelines on the food environment in recreational facilities. That study documented a 19% improvement in facility environment assessment scores and in availability of “choose most” and “choose sometimes” items in vending machines which, although statistically significant, nevertheless meant that the average post-intervention facility environment score was only 59% and only 17% of vending items were “choose most” [15]. That substantial voluntary change appears to be so difficult to achieve in recreational facility food services is perhaps unsurprising in light of current funding models, which make facilities partially dependent on the sale of unhealthy foods.

The availability of “choose most often” items was low within all of the facilities and was not consistently higher in adopter compared to non-adopter facilities. This outcome was partially a reflection of the low availability of such items in the marketplace, as for example, there are few “choose most often” food items suitable for sale within unrefrigerated vending machines. Although < 20% of items overall were “choose most often” in all facilities, ANGCY implementation and NEMS-R scores were often high, suggesting that policy makers and researchers should reconsider what constitutes a healthy food environment. Specific, high targets for the proportion of items that must fall within the “choose most often” category (ideally 100%) would help to ensure that nutrition guidelines support healthy food environments.

This study systematically applied a Diffusion of Innovations framework to better understand adoption and implementation of nutrition guidelines in recreational facilities in Alberta. We assumed causal complexity, that is, that there are multiple paths to adoption, and thus we used a set theoretic approach to discern three sets of factors [43]: 1) Factors that were common to all cases and were therefore not sufficient to compel or dissuade adoption and implementation of nutrition guidelines. 2) Factors that were unique to individual cases and not consistently associated with adoption and implementation. These factors may be influential in particular contexts. 3) Factors that distinguished adopters from the non-adopter, and were therefore sufficient and perhaps also necessary for adoption and implementation. The specific paths by which adoption and implementation may occur are not known, however, as many different combinations of these factors are possible [49]. In addition, it would be premature to discard factors within the first category as unimportant, as although they do not guarantee adoption and implementation, they may nevertheless prove essential in future studies [43]. Our analysis suggests that it is primarily factors within the third and perhaps also the second categories that determine whether or not adoption and implementation of nutrition guidelines will occur within a given context.

Although the specific adoption trajectories differed among cases, several important findings emerged. First, the keys to adoption and implementation relate to the manager. The manager is a reflective decision-maker whose beliefs, perceptions, and knowledge shape his decisions and actions. Adoption and implementation of nutrition guidelines in recreational facilities is more likely when the manager personally values healthy eating, has a broad scope of responsibility encompassing wellness, regards nutrition guidelines in a positive light, perceives a high tension for health-related change within his facility’s food environment, is willing to champion changes that contravene industry norms and that may be financially risky, perceives few competitive pressures, maintains good relations with industry, and is willing to partner with them to achieve desired outcomes.

The fact that adopters were willing to eschew industry norms and adopt the ANGCY despite potential negative repercussions marks them as innovators [50]. These individuals are critical to diffusion as they act as gatekeepers, importing new ideas into a system [26]. Managers, however, do not have free reign. Their decisions are made within a particular micro and macroenvironmental context that is a source of facilitators and barriers. Barriers, including poor managerial relations, financial constraints, limited capacity to implement nutrition guidelines, unfavourable power balances, and the provisions of food service contracts impeded action on the part of managers. Financial constraints in particular, were a strong and consistent barrier to adopting and implementing the ANGCY in all facilities, as sales reductions caused managers to question the degree to which the ANGCY would provide them with an advantage relative to their previous practices. We were unable to objectively verify whether offering healthier foods was profitable in this context, and evidence from other studies in recreational facilities [15, 24, 51] and schools is conflicting [52‐60] in this respect.

The challenge to balance support for affordable opportunities to be physically active with the need to promote healthy dietary behaviors is considerable in recreational facilities. Managers perceived that adopting the ANGCY in a choice-based format helped them to balance these competing priorities. Simply adding more healthy options to existing, largely unhealthy menus may not influence children’s dietary behaviors, however, as exhibited by students’ purchases in the full adopter facility. When given a choice, children tend to select unhealthy items [60‐67]. Parents too, at times may make poor nutritional choices for their children because powerful social factors, time [68] and informational constraints [69] can easily take precedence over longer-term, intangible health concerns. Providing individuals with both healthy and unhealthy options (ie. a choice-based format) and trusting them to choose the healthiest option in spite of environmental conditions that overwhelmingly promote the opposite is unlikely to curtail escalating obesity rates.

The second major finding that emerged from this study is that although managers played a major role in adopting and implementing nutrition guidelines, they could not accomplish these tasks alone. Intersectoral linkages and formal health promoting partnerships were essential. Multisectoral, health promoting partnerships have long been recognized as a fundamental ingredient in effective health promotion practice [70]. It is difficult to envision how effective solutions to obesity can be forged without active involvement from the corporations that control and shape the food supply [71]. In the context of implementation of voluntary nutrition guidelines, adopters recognized that they lacked capacity to implement the ANGCY and therefore requested assistance from industry, leveraging their existing collaborative relationships in a new, health promoting direction. Where health promoting public-private partnerships existed, adoption and implementation proceeded, whereas no action was taken in their absence. The sustainability of these partnerships is unclear, however.

In addition to formal partnerships, informal linkages with schools were important. Adopters were motivated to seriously consider adopting the ANGCY when others within their social networks shared how they were using the ANGCY in schools, and encouraged them to do the same. Diffusion of the ANGCY therefore occurred within municipalities, from schools to recreational facilities, rather than among recreational facilities, as adoption of the ANGCY was too low for diffusion networks to become activated in this context [16]. In addition, industry’s willingness to collaborate with recreational facilities was partially determined by their pre-existing capacity to implement the ANGCY, developed through their school-based operations. Thus, efforts to improve the school food environment provided a supportive context and capacity to implement similar measures in recreational facilities.

Voluntary initiatives such as the ANGCY are of limited effectiveness in counteracting the pervasive influence of macro-level forces within the food system, as the environmental supports for voluntary action are poor. ANGCY uptake may therefore continue to falter under the current voluntary approach, and where it does occur, our findings suggest changes to the food environment may be relatively minor. Stronger government action is required to promote healthy dietary behaviors among children. Such action could include relatively less coercive (eg. incentives) or more coercive measures (eg. regulation). First and foremost, funding models should not be antithetical to recreational facilities’ wellness mandates. Facilities derived a small percentage of their overall revenues from food services and sponsorships, and thus it would not be costly to replace this revenue. Next, the ANGCY should be revised to include specific, measurable and robust recommendations. Other actions could include financial incentives (eg. tax breaks) for industry to develop products that meet the definition of “choose most often” and for those corporations that succeed in selling, not simply offering, a high proportion of “choose most often” items. Similar to pay-for-performance schemes in health care, governments could incorporate guideline-related outcomes as performance accountabilities for recreational facilities to continue to receive a portion of their public funding. Finally, governments could simply mandate that all recreational facilities adhere to the ANGCY, ideally in a restrictive format. Although some may argue that such measures interfere with the individual’s right to choose, many current policies already constrain food choice within recreational facilities (eg. funding models that make facilities partially dependent on food service revenues) and therefore such measures would merely counter existing obesogenic policies. These findings illustrate the tension that exists among individual rights, profitability and public health within market-based economies, and will assist policy makers to formulate policies that balance these competing interests.

This study was unique and had many strengths, including its in-depth nature and the range of cases studied. Mixed methods provided a more comprehensive understanding of the research questions than could have been achieved with a single approach. Multiple quantitative and qualitative perspectives of the food environment highlighted the many ways in which the food environment can be conceptualized, and showed that using a single tool is likely to yield incomplete and biased findings. We used a novel theoretical framework to discern factors that influenced uptake of the ANGCY, a model that may now provide a theoretical platform from which to investigate the uptake and operationalisation of a variety of obesity prevention policies. Finally, we contacted facilities 6–18 months following each case study to ascertain whether their adoption status had changed, and whether they had made any nutrition-related changes to food services.

ANGCY implementation scores were not consistently higher among adopters, and indeed, were high in some non-adopter facilities. This result was consistent with adopters’ perceptions that the food environment did not change substantially following ANGCY implementation. These results may also suggest a problem with the scoring system, as although the tool was judged to have good content validity, its construct validity may be poor. It is possible that the tool is not sensitive enough to differences between adopters and non-adopter facilities, as facilities could only receive a score of 0, 1 or 2 for each item. Others, however, have used similar scoring systems with good results [40]. Alternatively, the inability of the scoring system to distinguish adopter from non-adopter facilities may reflect problems within the ANGCY themselves, as informants felt several ANGCY recommendations were simply good business practice and likely to be practiced in all facilities. The guidelines also lack specific, measurable targets, which made it difficult to judge the degree to which facilities had implemented the recommendations. All of these factors likely contributed to the poor performance of the scoring system, however we believe the latter two were particularly influential.

We used multiple, mixed tools to assess food environment quality, however even these tools could not fully capture its many dimensions. We focussed mainly on physical aspects of the micro food environment, and did not extensively investigate its political, sociocultural and economic aspects [72], nor did we capture the subjective perceptions of patrons. Because we assessed the food environment at a single time point we could only infer change in food environment quality from managers’ comments and from comparison of adopter and non-adopter facilities. There is no universally agreed upon definition of a healthy food, and it is likely that a higher percentage of items would have been classified as healthy (ie. “choose most often”) using different standards, as ANGCY standards for sodium, in particular, are very stringent.

We make no claim that cases in this study are representative of all recreational facilities in Alberta. However, we have highlighted broad areas to target for change and provided as much detail as possible to allow the reader to evaluate the opportunities for generalization to other contexts.