Robin McKinnon, a health policy specialist in the Risk Factor Monitoring and Methods Branch at the National Cancer Institute, was asked to summarize the primary themes from the workshop as its final presentation. This chapter integrates her remarks with a brief overview of a few key points from the workshop presentations and discussions.
McKinnon reminded participants that the workshop’s purpose was twofold,
• to explore the ways in which measurement techniques, strategies, and data sources can impede or accelerate progress toward preventing obesity; and
• to explore what additional knowledge of environments and policies is needed to assess progress toward obesity prevention.
McKinnon noted that “people used the term ‘measures’ to mean many different things” in the course of the workshop—for example, as (1) methods of assessment, such as survey instruments or policy audit tools; (2) techniques, such as cost-effectiveness analysis; (3) strategies/indicators, such as number of supermarkets per population; and (4) data sources, such as surveillance systems or databases. She observed that greater clarity in the use and awareness of terminology could be helpful.
Many of the presenters offered illustrations of individual-level energy-balance behavior embedded within a nested set of influences, including the home, organizations, and the physical and policy environments. Presenters focused in detail on different aspects of the food environment and the built environment for physical activity and their influences on energy balance, but it is also necessary to consider the broader picture. What people consume appears to be influenced by a wide array of factors, including availability and convenience, food and nutrition knowledge, agricultural policies, economic incentives, marketing messages, family and cultural customs and preferences, and individual physiology. How and when people engage in physical activity is influenced by many of the same factors, and by the nature of the built environment where they live and work, their public transportation options, and more. Presenters made clear that the various influences on energy balance are important, although they may be difficult to isolate. Furthermore, several presenters emphasized that sectors that may appear to be unrelated to health actually may be relevant and important to efforts to reduce obesity rates.
Measurement strategies and techniques are a critical foundation for research, McKinnon observed, adding that a study may be well-designed and data rigorously analyzed, but if the basic measures, the assessment tools, are not valid and reliable, true associations between exposures and outcomes may not be understood. A relatively recent review of measures of the food and physical activity environments notes that although much progress has been made, further progress is needed (McKinnon et al., 2009). For example, it suggests that refinement of the measurements of environments in low-income and high-risk communities is needed, as are increased rates of validity and reliability testing and/or reporting. Progress since that report was published includes advancement and refinement in geographic analyses, as well as improvements in measures of the food and physical activity environments. The number of studies assessing the association between the food environment and health and dietary behaviors, in particular, has increased substantially in recent years. Nonetheless, there is still a relative paucity of measures with which to systematically measure policies and policy change.
The workshop presentations reflected both these issues and this progress. McKinnon observed that fairly good individual-level measures of diet and physical activity exist, and that environmental measures in these areas are developing that use, for example, surveys, geographic information systems (GIS), diaries, or universal product code (UPC) scanning.
Improvements for these measures include more validity testing and reporting (particularly with subpopulations of interest) and the reporting of both perceived and objective environmental measures. Some policy assessment methods exist, but they are often time-intensive, legislation language can be difficult to interpret, and enactment of a policy does not always equal implementation.
McKinnon recapped measurement techniques and methods presented at the workshop, and reiterated the importance of choosing a study design that focuses on answering the right questions. Measurement techniques included survey instruments, GIS, and diaries, as noted above. Methods included health impact assessment and economic methods. McKinnon provided additional context when summarizing the portion of the workshop covering those methods.
The impact of obesity on health and related costs is great. It has been estimated that 14 to 20 percent of cancer deaths are attributable to obesity (Calle et al., 2003), and the link between obesity and many other diseases, such as type II diabetes and cardiovascular disease, is well established. Obesity is estimated to be responsible for $147 billion in health care costs annually (Finkelstein et al., 2009). Researchers also have estimated that increased obesity rates are responsible for 27 percent of the rise in health care costs (Thorpe et al., 2004), having a greater impact than either smoking or problem alcohol consumption (Sturm, 2002). There are other costs of obesity to society as well, resulting from increased disability and absenteeism and reduced productivity (Finkelstein et al., 2005). It may be important to remind the public health community that there are other outcomes of interest besides health. Cost/benefit analysis can show that health, health care, and related costs are not the only outcomes to consider, and also that interventions may have unexpected associated costs and benefits. On the other hand, as discussed in Chapter 4, the claim that reducing obesity rates will reduce overall costs may be questioned. The public health community might benefit instead from analysis of value (in terms of health outcomes) for money spent that can be used to compare the effectiveness of one intervention versus another.
Partner organizations that reach a broad array of audiences and communities may be helpful in gathering data, as well as disseminating results. Data and policies from the worlds of transportation, urban planning, parks and recreation, and many other sources are important to obesity researchers.
Several presenters noted the value of encouraging the view that questions about public health and obesity prevention in particular should be folded into policy thinking in a variety of areas and data collection across multiple disciplines and levels. Many speakers highlighted the importance of using both quantitative and qualitative data.
Comprehensive surveillance systems and databases are especially important because of the increasing focus on supportive environments and policies for improved diet and physical activity behavior. There are good examples of such systems in the areas of tobacco control and alcohol policy. Policy tracking databases exist, but no surveillance systems currently are in place with which to address all of the most important obesity questions. For example, no such comprehensive system exists for assessing the physical activity environment, as presenters pointed out, nor is there a national system for cataloguing local policies related to the food and physical environments.
Food marketing research produces a wealth of data that public health researchers may be able to use to understand the quantity and content of food and beverage advertisements to which people are exposed, their access to different food products, and other important questions. However, potentially helpful commercially available data sources are often costly, and thus researchers frequently are limited in their access to such data. Furthermore, communication between the food marketing and public health communities regarding these research data is not well established. One way to move forward in this area might be to encourage the commercial data sources to incorporate public health data and the gathering of those data within their systems.
With this quick snapshot of highlights from the workshop as a backdrop, McKinnon outlined her suggestions for moving forward. She reminded the audience that racial and ethnic minorities are at a higher risk for obesity, and as speakers had noted that marketers and food companies appear to target these groups, adapting measures to evaluate the impact of this marketing appears to be an important priority. Researchers need to find ways to capture the synergistic and cumulative effects of marketing that takes many forms and yet may target small segments of the population. Using qualitative as well as quantitative methods may be particularly helpful in developing measures for communities at highest risk.
McKinnon reiterated her support for Dr. Krebs-Smith’s observation regarding the importance of matching the measures and methods to the questions of interest and suggested some steps toward that end:
• Determine the exposures and outcomes of greatest interest using, for example, expert recommendations from the Centers for Disease Control and Prevention and the Institute of Medicine as starting points.
• Assess existing measurement techniques, measures (assessment methods), strategies, and data sources. The National Collaborative on Childhood Obesity Research’s Measures Registry and Catalogue of Surveillance Systems may be helpful resources in this regard.
• Identify gaps (such as the lack of a survey of public health policies, the lack of measures tailored to racial and ethnic minorities, and the lack of measures of consumers’ responses to marketing), and establish priorities. McKinnon presented a possible model for prioritizing future work, shown in Figure 8-1. She suggested that the focus should be on measures that are anticipated to have high impact but are relatively easy to implement. Measures with high impact and high implementation costs might also be a focus, but measures anticipated to have low impact should not have priority.
• Identify partners from beyond the public health sector, including the transportation and urban planning communities, and identify the necessary strategies for collaboration.
• Focus on and promote study designs that emphasize answering the right questions.
• Evaluate results, and disseminate them widely.
FIGURE 8-1 Possible model for setting priorities for filling gaps in measures.
McKinnon closed by saying, “Let’s not measure simply what’s easiest and most convenient. Let’s focus on the areas of greatest need and anticipated impact.”
Calle, E. E., C. Rodriguez, K. Walker-Thurmond, and M. J. Thun. 2003. Overweight, obesity, and mortality from cancer in a prospectively studied cohort of U.S. adults. New England Journal of Medicine 348(17):1625-1638.
Finkelstein, E., I. C. Fiebelkorn, and G. Wang. 2005. The costs of obesity among full-time employees. American Journal of Health Promotion 20(1):45-51.
Finkelstein, E. A., J. G. Trogdon, J. W. Cohen, and W. Dietz. 2009. Annual medical spending attributable to obesity: Payer and service-specific estimates. Health Affairs 28(5):w822-w831.
McKinnon, R. A. 2011. Workshop summary presentation. Presented at the Institute of Medicine Workshop on Measurement Strategies for Accelerating Progress in Obesity Prevention, March 23, Irvine, CA.
McKinnon, R. A., J. Reedy, S. L. Handy, and A. B. Rodgers. 2009. Measurement of the food and physical activity environments: Enhancing research relevant to policy on diet, physical activity, and weight. American Journal of Preventive Medicine 36(Suppl. 4):S81-S190.
Sturm, R. 2002. The effects of obesity, smoking, and drinking on medical problems and costs. Health Affairs 21(2):245-253.
Thorpe, K. E., C. S. Florence, D. H. Howard, and P. Joski. 2004. The impact of obesity on rising medical spending. Health Affairs 23(Suppl. 2):W4-480-W4-486.