Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 39
4
Looking Across Domains
Key Points Noted in Presentations
• The National Collaborative on Childhood Obesity Research
(NCCOR) is a collaborative effort designed to improve the effi-
ciency, effectiveness, and application of research related to child-
hood obesity. It has developed new tools for obesity researchers:
(1) a Measures Registry—a searchable online registry of diet and
physical activity measures at the individual and environmental
levels; and (2) a Catalogue of Surveillance Systems—an online
directory of national, state, and local data resources available at
multiple levels.
• The body mass index (BMI) can be useful in screening to identify
those at either an elevated weight for height or a low weight for
height, but it is not a precise measure of adiposity and must be
interpreted with care. The significance of this ratio is affected by
variations in body structure across ethnic groups, growth and
development in children and adolescents, and loss of height with
aging.
• From one economist’s perspective, the argument that reducing
and preventing obesity will save money is not robust. Improve-
ments in measurement and analyses would strengthen the ability
to support detailed claims for most interventions. An argument
based on cost-effectiveness—taking into account the trade-offs
that come with any intervention and focusing on health benefits
rather than cost savings—may be more effective.
39
OCR for page 40
40 MEASURING PROGRESS IN OBESITY PREVENTION
In addition to measures of the built environment for physical activ-
ity and of the food and nutrition environment, there are measures, data
resources, and methods that relate to both of these environments and may
help researchers and policy makers assess progress in obesity prevention.
Robin McKinnon, health policy specialist in the Risk Factor Monitoring
and Methods Branch at the National Cancer Institute, described research
activities of the National Collaborative on Childhood Obesity Research.
Robert Malina, professor emeritus in the Department of Kinesiology and
Health Education at the University of Texas, Austin, and research professor
in the Department of Kinesiology at Tarleton State University, described
research related to growth and BMI. Roland Sturm, senior economist at
RAND and professor of policy analysis at the RAND Pardee Graduate
School, discussed economic perspectives on nutrition, physical activity, and
obesity interventions.
NEW TOOLS FOR CHILDHOOD OBESITY RESEARCH
Presenter: Robin McKinnon
The National Collaborative on Childhood Obesity Research is an effort
to enhance coordination of activities related to childhood obesity research.1
Four of the major funders in this area—the Centers for Disease Control
and Prevention (CDC), the National Institutes of Health (NIH), the Robert
Wood Johnson Foundation (RWJF), and the U.S. Department of Agricul-
ture (USDA)—joined forces to improve the efficiency, effectiveness, and
application of research related to childhood obesity, McKinnon explained.
Newly available NCCOR tools are designed to address a range of issues
facing researchers interested in childhood obesity, she noted. To make opti-
mal research decisions, researchers need to know, for example, what data
resources exist and how to gain access to them, how much they cost, and
which variables can be linked across data systems, as well as what measures
exist and which ones have been tested for validity and reliability.
NCCOR’s priorities include promoting the use of common measures
and methods across childhood obesity prevention efforts and research, and
encouraging the development of standard measures with which to describe
and evaluate interventions, especially projects that address policies and
environments, McKinnon explained. NCCOR has produced a new Mea-
sures Registry to facilitate access to available measures, identify gaps in
measures, and foster the development of new measures. This searchable,
1 For more information about NCCOR, see www.nccor.org (accessed August 26, 2011).
NCCOR’s Measures Registry is available at www.nccor.org/measures, the Catalogue of Sur-
veillance Systems at www.nccor.org/css.
OCR for page 41
41
LOOKING ACROSS DOMAINS
web-based tool includes nearly 750 measures in four domains: individual
dietary behavior, individual physical activity behavior, the food environ-
ment, and the physical activity environment. Questionnaires, diaries, logs,
electronic devices, and direct observations of people and environments are
among the types of measures included.
The Measures Registry provides such information as which domains
are measured; validity and reliability; settings, geographic areas, and popu-
lations measured; and protocols for use. Measures in the registry focus on
individual behaviors and the environments that may affect those behaviors.
Information in the registry is based on literature searches, a best-practice
review, lists of recommended measures prepared by experts in the areas cov-
ered, and a tailored abstraction tool. The registry allows users to search and
browse for measures, search for specific details related to those measures,
link to other measures and other studies that have been conducted using
a particular measure, and find measures that are in development and have
not yet been published. Researchers may also submit their own measures
for inclusion.
NCCOR also has developed the Catalogue of Surveillance Systems,
which is intended to increase awareness and use of existing obesity-related
data resources at the individual, household, organizational, community, and
macro/policy levels. Knowledge of the broad array of relevant surveillance
systems has been limited, given that there has been no easy way to locate
these systems. In addition, relatively little research has taken advantage of
possibilities for linking data across the spectrum of influences on behavior.
As of the time of the workshop, 77 surveillance systems, providing
national, state, and local data, were featured in the catalogue, which is
updated continually. Only systems that provide data collected in the United
States within the last 10 years and make raw data publicly available (so that
researchers can access and manipulate the data for themselves) are included.
Extensive reviews are part of the process of selecting systems for inclusion
and creating their web-based profiles. Using the catalogue, researchers
can search by topic, compare system attributes, and gain access to related
resources such as summary statistics or legislative databases. Figure 4-1
illustrates the breadth of the resources represented in the catalogue, super-
imposed on an ecological model of factors that influence obesity.
NCCOR has plans for sustaining and updating these two resources,
McKinnon explained, and for expanding awareness of what they have to
offer. She also noted some early feedback that can be used in assessing the
measures and data resources featured in the NCCOR tools. Although a
wide range of measures already is available, she observed, the degree to
which they have been tested for validity and reliability varies considerably.
Relatively few surveillance systems include measured height and weight
data (instead, most include self-reported data), and very few include objec-
OCR for page 42
42 MEASURING PROGRESS IN OBESITY PREVENTION
Macro/Policy
Price Data on
Medical
Groceries,
School and Community Expenditure
Transportation
Household
School
Community Nutrition,
Diet
Crime Activity
Individual Online
Statistics
Physical Activities,
Activity Behavior
Household
Travel
Federal Nutrition
Agricultural Assistance Program
School
Production Participation
Traffic Wellness
Fatalities Policies
State Snack
and Soda Tax
Policies
FIGURE 4-1 Example of the breadth of resources or data sets represented within
4-1.eps
the NCCOR Catalogue of Surveillance Systems, categorized by levels of a social
ecological model.
SOURCE: McKinnon, 2011.
tive measures of physical activity. Data from nutrient and food group
databases can be linked to individual surveys, but not to surveys of grocery
stores or markets. Also, surveillance of obesity-related policies is limited.
McKinnon concluded by noting that NCCOR is working to build aware-
ness of resources already available, to identify gaps and priorities for future
work, to spur innovation in research, and ultimately to reduce rates of
childhood obesity.
BODY MASS INDEX
Presenter: Robert Malina
BMI is the most frequently used measure of individuals’ weight status.
Malina described its use and a number of related issues. BMI—an index
of weight for height—was originally developed by Adolphe Quetelet, a
19th-century sociologist who calculated it in terms of weight in kilograms
divided by height in meters squared (it was called the Quetelet Index until
the early 1970s). Its primary use is as a screening tool to identify those at
OCR for page 43
43
LOOKING ACROSS DOMAINS
an elevated weight for height or a low weight for height. Current BMI stan-
dards for overweight and obesity in the United States are based on weight
for height distributions from certain health surveys over a period of time.
The World Health Organization (WHO) defines an adult as overweight
if his or her BMI is at or above 25 kg/m2 and obese if it is at or above
30 kg/m2. Children and adolescents in the United States are classified based
on growth charts developed by CDC, and are considered overweight if they
are between the 85th and 95th percentiles in weight for their age and obese
if they are at or above the 95th percentile.
Even these simple calculations, however, raise issues, Malina noted.
CDC did not include weight data from the 1988-1994 National Health
and Nutrition Examination Survey (NHANES)2 in determining overweight
and obesity criteria for U.S. growth charts because the gain in weight
across surveys was viewed as undesirable from a public health perspec-
tive, Malina explained, even though that survey was the first to reveal the
obesity epidemic.
In most countries, people use the International Obesity Task Force
(IOTF) criteria for overweight and obesity. These criteria were derived from
a data set based on six cross-sectional studies that includes U.S. data col-
lected in the 1970s and 1980s—before the obesity epidemic had emerged,
Malina noted. As Figures 4-2 and 4-3 show, however, use of these criteria
results in significantly different prevalence rates of overweight and particu-
larly of obesity for male (Figure 4-2) and female (Figure 4-3) children and
adolescents.
BMI growth rates vary significantly among children, Malina added.
At birth children are fairly consistent in size, but they follow their own
trajectories as they grow. Based on data for French children, Rolland-
Cachera and colleagues (1991) show (Figure 4-4) that there is a rapid rise
in BMI from birth to age 1 or 2, followed by a gradual decline into middle
childhood. This phase is followed by an “adiposity rebound,” as Rolland-
Cachera calls it, which may have important implications for children’s later
rates of overweight and obesity. Children who begin this rebound early—
before age 5 1/2—are more likely to have a higher BMI in young adulthood
than are those who begin the rebound after age 7, said Malina.
Other evidence reinforces this point, Malina explained. Children who
are more active at ages 4 to 11 years experience the rebound at later ages
and are not as fat in early adolescence as less active children (Moore et
al., 2003). Thus, the “rebound” period may be an important window for
intervention, Malina suggested. Not only would it potentially be valuable
to delay the rebound, but there is also reason to believe that children’s basic
2 Formore information about NHANES, see http://www.cdc.gov/nchs/nhanes.htm (accessed
August 2011).
OCR for page 44
44 MEASURING PROGRESS IN OBESITY PREVENTION
30
28
IOTF
26
Obesity CDC
BMI, kg/m2
24
22
20
Overweight
IOTF
18
CDC
16
6 7 8 9 10 11 12 13 14 15 16 17 18
Age, years
FIGURE 4-2 Differences in overweight and obesity prevalence among male children
and adolescents using Centers for Disease Control and Prevention (CDC) criteria
New Figure 4-2
versus International Obesity Task Force (IOTF) criteria.
NOTE: Based on CDC percentiles for overweight and obesity and IOTF cut-offs
for overweight and obesity.
SOURCE: Malina, 2011.
movement skills are fairly set by age 6 or 7 years, and that obese youngsters
have deficiencies in movement coordination that limit their physical activity
opportunities as they age. Workshop participants suggested that the best
target for intervention might be even earlier if data indicating that very
young children who are overweight or obese have significantly increased
odds of being so as adults and experiencing the health problems that often
come with excess weight.
Greater variation in weight trajectory is evident as children enter ado-
lescence, Malina explained, because of the differential timing of growth
spurts. Generally, these spurts affect height first—the legs and then the
trunk—and then weight. Girls experience the spurts significantly earlier
than do boys, Malina added, but he noted that it is necessary to view data
collected across at least 5 or 6 consecutive years of growth to see the pat-
terns. Looking only at maturity indicators, such as stages of puberty, age at
menarche, or peak height velocity (the period of fastest growth) is a limited
OCR for page 45
45
LOOKING ACROSS DOMAINS
32
30
28
IOTF
26
CDC
BMI, kg/m2
Obesity
24
22
Overweight
20
IOTF
18
CDC
16
6 7 8 9 10 11 12 13 14 15 16 17 18
Age, years
FIGURE 4-3 Differences in overweight and obesity prevalence among female chil-
dren and adolescents using Centers New Figure 4-3Control and Prevention (CDC)
for Disease
criteria versus International Obesity Task Force (IOTF) criteria.
NOTE: Based on CDC percentiles for overweight and obesity and IOTF cut-offs
for overweight and obesity.
SOURCE: Malina, 2011.
approach, he noted, because these measures identify discrete milestones
or stages that are really part of a continuous process lasting several years.
Youngsters reach these stages at a range of ages, and chronological age mat-
ters as well. For example, observed Malina, “A post-menarcheal 11 year-
old girl is very different from a pre- or post-menarcheal 14 year-old girl.”
The significance of the timing of growth spurts is suggested by findings
from the Leuven Growth Study (Beunen et al., 1994). This mixed longi-
tudinal study of boys in Belgium revealed that those who attained their
peak height velocity at earlier ages had higher BMIs than those who did so
later, and that this difference lasted at least until age 30. Two other studies
(Guo et al., 1994; Rolland-Cachera et al., 1991) reinforced the connec-
tion between elevated BMI levels during childhood and adolescence and in
young adulthood. In particular, Malina noted, study data from Guo and
colleagues (1994) showed that youngsters at the higher BMI percentiles
(e.g., 75th, 85th, 95th) were most at risk for obesity at age 35.
OCR for page 46
46 MEASURING PROGRESS IN OBESITY PREVENTION
22
Boys
Girls
20
18
BMI, kg/m2
16
Adiposity
rebound
14
12
0 2 4 6 8 10 12 14 16 18 20
Age, years
FIGURE 4-4 The rapid rise in BMI from birth to age 1 or 2 years is followed by a
decline and then an “adiposity rebound” in middle childhood.
SOURCE: Adapted, with permission, New Figure 4-4
from R. M. Malina, C. Bouchard, and O.
Bar-Or, 2004, Growth, maturation, and physical activity, 2nd ed. (Champaign, IL:
Human Kinetics), 64. Data from M. F. Rolland-Cachera, T. J. Cole, M. Sempe, et
al., 1991, “Body mass index variations: Centiles from birth to 87 years,” European
Journal of Clinical Nutrition 45:13-21.
Also important, Malina observed, is that BMI is an indicator of weight
for height, not of fatness. The calculation does not discriminate between
lean and fat tissue, and the correlation between BMI and percentage fat var-
ies by individual. BMI is about equally correlated with fat-free mass (lean
tissue) and fat mass and percentage fat in youth (Malina and Katzmarzyk,
1999). For men, he added, the correlation between BMI and percentage
lean tissue is greater than the correlation between BMI and percentage fat
tissue (Romero-Corral et al., 2008). The reverse is true for women, which
probably reflects sex differences in the ratio of lean to fat tissue, Malina
explained.
Figure 4-5 illustrates the point: the upper right and lower left quadrants
on both graphs represent men and women for whom BMI and percentage
fat tissue are either both above or both below the target range. The upper
left and lower right quadrants on both graphs represent those for whom the
two measures do not match. Thus, Malina explained, “something is being
missed by the BMI.”
Another issue to consider is that people lose height as they age, a fact
that should be taken into account in the interpretation of BMI measures. As
data from Sorkin and colleagues (1999) and others show, the decrease usu-
OCR for page 47
47
LOOKING ACROSS DOMAINS
FIGURE 4-5 Age- and race-adjusted correlation between body mass index and
body fat percentage for men and women less than 60 years of age.
NOTE: The World Health Organization reference standard for obesity is body fat
>25 percent in men and >35 percent in women; the Centers for Disease Control
and Prevention defines obesity as BMI ≥30 kg/m2, as indicated by the arrows in
the figure.
SOURCE: Adapted by permission from Macmillan Publishers Ltd: [International
Journal of Obesity] Romero-Corral, A., V. K. Somers, J. Sierra-Johnson, R. J.
Thomas, M. L. Collazo-Clavell, J. Korinek, T. G. Allison, J. A. Batsis, F. H. Sert-
Kuniyoshi, and F. Lopez-Jimenez. “Accuracy of Body Mass Index in Diagnosing
Obesity in the Adult General Population.” International Journal of Obesity 32,
no. 6 (2008).
ally begins at age 30 to 35 and accelerates after age 55, with women losing
height more rapidly than men. Body composition also varies, on average,
among ethnic groups, Malina added. Asian populations, for example, tend
to have lower BMIs but higher percentages of body fat than other groups,
as well as more abdominal fat at lower BMI levels (Malina, 2005). Thus,
WHO has identified different overweight and obesity thresholds for these
populations (WHO, 2000). In general, people who are shorter in stature
and have shorter legs will have higher BMIs, and the distribution of fat in
the body also varies with ethnicity (Malina, 2005; Malina et al., 1995).
These points suggest that BMI may be a less sensitive indicator of
health concerns for some populations than for others, Malina observed.
A number of observers have suggested that other measures—such as waist
circumference, an indicator of abdominal fat (which has the greatest health
significance), or ratio of waist circumference to height (e.g., Ardern et al.,
2004; Cleeman, 2001; Hsieh et al., 2003)—may be better tools for iden-
tifying the presence of excess weight that may be a health concern. How-
ever, defining the thresholds for these measures is still complex because of
variation by ethnicity and other characteristics, Malina explained. A 0.50
waist-to-height ratio has been proposed as the cut-off between normal
and overweight individuals in Japan (Hsieh et al., 2003), he noted, but in
OCR for page 48
48 MEASURING PROGRESS IN OBESITY PREVENTION
another population, the Zapotec Indians of Oaxaca, Mexico (whom he has
studied), who are extremely small in stature, the mean waist-to-height ratio
is 0.53, and they “are by no means overweight.”
Malina also noted that recent studies of the correlations among differ-
ent measures have yielded mixed results (Flegal et al., 2009). BMI, waist
circumference, and waist-to-height ratio were more strongly correlated with
each other (correlations ranged from 0.85 to 0.97) than with measures
of percentage of body fat (correlations ranged from 0.65 to 0.87). From
another perspective, the data showed that none of the measures were excel-
lent at classifying individuals as normal, overweight, or obese. For example,
BMI correctly classified only 46 percent of men and 49 percent of women,
while waist circumference correctly classified 51 percent of men and 42
percent of women (Flegal et al., 2009).
In conclusion, Malina commented that BMI is a useful measure that is
easy to calculate and is good for showing trends among populations and
subgroups. Because it is affected by growth and pubertal maturation, by
aging, and by ethnicity it must be interpreted carefully. Research is needed
to monitor BMI rates longitudinally, in Malina’s view, “so we can find
out how the BMI changes and what factors affect individual change.” For
example, little is known about how sensitive BMI is to physical activity,
particularly in youngsters, but BMI could be an important tool for tracking
the role of physical activity in preventing unhealthy weight gain.
Several participants agreed that BMI and basic height and weight mea-
sures can be valuable tools for screening in schools to identify children and
families who could benefit from counseling about nutrition and physical
activity (several states currently use such screens, Malina noted). Malina
cautioned that the information needs to be interpreted and communicated
with care, not only because of the variation in healthy physical proportions
among ethnic groups, but also because the information can do more harm
than good if individuals draw the wrong conclusions—for example, devel-
oping anorexic behaviors because of concern about modest excess weight.
MAKING THE ECONOMIC CASE FOR OBESITY PREVENTION
Presenter: Roland Sturm
Making sustainable progress against obesity requires strong public sup-
port. Both economic and public health perspectives play an important role
in the policy process, and the interventions supported by both perspectives
are most likely to be effective and politically acceptable, Sturm explained.
Yet there is a gulf between these two research perspectives, and they often
appear to be at odds with each other. “Without understanding the eco-
nomic perspective,” Sturm suggested, “public health professionals will
OCR for page 49
49
LOOKING ACROSS DOMAINS
have limited influence and sometimes may appear out of touch with public
opinion.” Sturm suggested that the reputation of public health experts is
that their focus on health status and willingness to intervene if health could
be improved may ignore individual preferences and fail to take into account
other important goals that people have. When public health messages seem
to be too heavy-handed or out of touch with majority opinion, the messages
of industry lobbyists seem more reasonable. “They know how to play the
game,” he added, expressing his view that “they resonate with the public”
and are talented at communication.
From an economic point of view, health (and even more so a healthy
diet or physical activity) is just one of many competing goals that people
must reconcile in their daily lives. In contrast to a public health perspective
in which decisions are made by experts, an economic perspective tackles
questions with the assumption that people can and do make their own deci-
sions, Sturm said. In the United States, more than in most other countries,
the belief in consumer sovereignty has a strong influence and affects how
institutions work. Long-standing federal guidelines, for example, require
agencies to identify market failures (i.e., situations where market forces by
themselves cannot achieve desirable outcomes) before issuing regulations.
Moreover, considering benefits and costs, as well as their distribution,
is important in deciding what interventions might be valid and useful.
Thus, Sturm explained, interventions are most likely to be successful and
politically sustainable when public health and economic perspectives are
integrated.
Cost-effectiveness or cost-utility analysis is applicable to obesity pre-
vention, Sturm explained, because the goal is to improve health in an opti-
mal way, given constrained resources. In this sort of analysis, economists
prefer to quantify outcomes using natural units, which in this instance
might be cases of obesity or disease prevented, or life-years saved, perhaps
accounting also for years of disability saved. Since the field may not readily
produce those units, a more general measure is used, such as morbidity and
mortality rates. Economists also strive to account honestly for all costs of
an intervention, and Sturm stressed that when this is not done, the cred-
ibility of the analysis is undermined, and the result is likely to polarize the
debate further rather than broaden support for the intervention. Obesity
prevention efforts cost money, and making a strong case for such efforts
requires more than saying they improve health, Sturm observed. Nor should
we make promises that may not be fulfilled, he said. He shared his belief
that from an empirical point of view, the claim that money could be saved is
almost certainly wrong, saying that prevention improves health but “almost
never saves costs.” Of course, obesity is associated with higher health
care costs, and preventing obesity would avoid these excess expenditures,
which are about one-third higher among obese than nonobese individuals
OCR for page 50
50 MEASURING PROGRESS IN OBESITY PREVENTION
and even double among those with a BMI more than 40 (Andreyeva et
al., 2004; Sturm, 2002). Yet Sturm posed the question of whether there
would actually be cost savings after intervention costs were taken into
account. He noted that promises of such savings have been made in other
areas and reliably failed to materialize on a larger scale (Sturm, 2001). In
his view, however, an even larger conceptual flaw is involved in trying to
justify obesity prevention with an offset in medical costs. If reducing costs
were the true goal, he observed, eliminating doctors and hospitals—or
increasing smoking, which could lead to earlier deaths and thus save the
costs of Social Security pensions and Medicare—would be a surer way to
accomplish that goal.
Sturm summarized his points by explaining that cost-effective is not the
same as cost savings. Cost-effective means “good value for the resources
compared to alternative interventions or policies,” he said, reiterating his
belief that cost savings are unlikely to exist for many obesity interventions.
Turning next to the sorts of evidence he believes would be useful in
developing an approach to obesity prevention, Sturm described a review of
behavioral physical activity interventions. Wu and colleagues (2011) identi-
fied 91 such interventions for which there was evidence of effectiveness and
for which cost information was available. They calculated cost-effectiveness
ratios for these interventions and found that the ratios ranged from $0.001
to $60 per MET3 hour. They found that the most cost-effective interven-
tions were also the least expensive per individual reached, but also had only
“tiny” absolute effects. The least cost-effective interventions were those that
provided some type of social support as part of an effort to promote behav-
ior change in individuals. If we could take those published estimates at face
value, Sturm said, they would provide exactly the information needed to
develop the most cost-effective obesity prevention interventions. Unfortu-
nately, while this is the direction research should pursue more intensively,
the research field is not yet at this point, Sturm noted. One indication is the
wide range of cost-effectiveness among interventions designed to improve
access to opportunities for physical activity or community-wide health
campaigns. To further illustrate his point, Sturm analyzed the quality of
136 studies of physical activity interventions cited in the appendix of the
systematic review by Wu and colleagues (2011) and compared the effects
reported. After controlling for such factors as whether the study was a
randomized trial, length of follow-up, and degree of attrition, he found that
3 MET = metabolic equivalent of the task. One MET hour is the equivalent of the energy
expended by the body during 1 hour of rest and is standardized based on body weight. Sev-
eral MET hours of exercise can be accumulated during 1 “real-time” hour; for example, 1
hour of moderate walking is approximately equal to 3 MET hours (http://www.johnshopkins
healthalerts.com/alerts/healthy_living/JohnsHopkinsHealthyLivingHealthAlert_1821-1.html;
http://mbsrunner.com/mtp/Forms/glossaryreg.aspx).
OCR for page 51
51
LOOKING ACROSS DOMAINS
the lower the quality of the evaluation, the lower was the cost-effectiveness
ratio (dollars per increase in activity) that was found. In other words, the
most positive results came from the weakest studies.
At present, Sturm noted, no single study provides sufficient information
to support cost-effectiveness analysis for many of the most common policy
suggestions, such as food taxes, advertising bans, or fast-food bans. An
alternative is to use modeling or simulation approaches (e.g., Levy et al.,
2011). However, these approaches are still in the developmental stage. The
best example to date is the Australia Assessing Cost-Effectiveness in Obe-
sity (ACE-Obesity) project, which has produced a number of publications
(e.g., Magnus et al., 2009; Moodie et al., 2009, 2010; Sacks et al., 2011).
ACE-Obesity uses a speculative measure of health outcomes, and more reli-
able estimates of effectiveness are needed for this work to inform policy,
Sturm said. Noting that the general approach is promising, he discussed a
few of that project’s findings. Traffic-light nutrition labeling (an approach
used in Australia and the United Kingdom, in which red, yellow, and green
color symbols indicate foods that are unhealthy, okay occasionally, and
healthy) and junk food taxes appeared to be highly cost-effective (Sacks et
al., 2011), although Sturm noted that some of the underlying assumptions
used in these studies are questionable. Removing television advertising for
unhealthy foods also appeared to be highly cost-effective among Australian
children (Magnus et al., 2009).
The picture is similar for cost-effectiveness analysis of the built envi-
ronment, Sturm explained. The methodology of the available studies varies
significantly, limiting the conclusions that can be drawn (Feng et al., 2010).
In Sturm’s view, the findings that are frequently cited in public debate have
been “cherry-picked and are likely to dramatically overstate possible out-
comes.” Further, the details of some interventions under discussion (e.g.,
increasing the number of supermarkets in underserved areas) are not suf-
ficient for costs to be calculated.
The rate of false published research results is high in new and competi-
tive research topics that are still “quite experimental,” in Sturm’s view. This
does not mean that this research area is inherently flawed, but rather that
more systematic research is needed. It is a well-documented phenomenon in
new fields that some initial results do not hold up when replicated, Sturm
said, citing as an example findings that there are genetic markers for obesity
and diabetes, which could not be replicated (Redden and Allison, 2003).
The key to improving policy recommendations is to accelerate this “shake
down” phase in the obesity prevention area through systematic replications,
he suggested.
“[Publication] biases may be amplified in the policy process,” Sturm
added. Policy makers have a natural tendency to use the most convenient
data. Cohen and colleagues (2010, p. 88) write, “it may be politically more
OCR for page 52
52 MEASURING PROGRESS IN OBESITY PREVENTION
expedient to promote an increase in consumption of healthy items rather
than a decrease in consumption of unhealthy items, but it may be far less
effective.”
Sturm offered several recommendations for identifying more efficient
and sustainable means of preventing and reducing obesity:
• Integrating economic perspectives with public health goals—spe-
cifically, taking into account the many trade-offs involved in such
decisions—will help to minimize ideological battles and build
broader support for interventions.
• Sturm urged policy makers and others to make an argument
based on health benefits and avoid using the argument that an
intervention will save money, stating his opinion that the latter
argument is “not credible, conceptually flawed, and likely to be
counterproductive.”
• Sturm also argued that cost-effectiveness estimates for obesity pre-
vention measures need further development and are most suit-
able at this stage for identifying broad classes of interventions
that are more or less cost-effective. It is difficult to conduct more
detailed comparisons of cost-effectiveness with currently avail-
able data. Modeling and simulation are more usable now, Sturm
added, although even these approaches are in the early stages of
development.
• It is important to avoid exaggerating the findings from early
research. Emerging areas of investigation commonly yield high
rates of false positives and overestimates of effects, and such results
are often not replicated in later studies. “This is a good, healthy
process,” Sturm explained. “We just have to watch out for it.”
• At the same time, replication is important, Sturm added. Promis-
ing results from early studies need to be tested with different data
sets. It can be difficult to secure funders for replication studies, but
policy makers need these possibly less visible, second-round studies
if they are to have reliable and credible data.
• Sturm urged the field to “keep a diversified portfolio.” At present,
the evidence is not strong enough to support dedicating major
investments in particular areas, and doing so would likely squeeze
out less ambitious but possibly more promising interventions.
Participants had a number of comments and questions about Sturm’s
perspective. One suggested that the logical implication of his view is that
if there is no clear evidence that reducing obesity rates saves money, there
must not be a problem, asking “Why don’t we all just be fat and not worry
about it?” Others shared this concern, noting, for example, that the diseases
OCR for page 53
53
LOOKING ACROSS DOMAINS
(such as heart disease and diabetes) that account for 80 percent of health
care costs in the United States are all obesity related. They questioned how
it then could be possible that investing resources in reducing obesity is not
cost-effective.
Sturm reiterated his basic argument that “we don’t have doctors to save
medical costs; we have doctors to improve health care.” Acknowledging
the importance of this point, a participant suggested that members of the
health community turn to cost arguments because they do not know how
to frame the potentially stronger arguments about the intrinsic value of
improving health outcomes. But she lamented that the economists’ perspec-
tive does not appear to offer a way to bolster the health argument. Another
participant agreed that it is difficult to make the case for interventions
because while they can be demonstrated to work, people tend to revert to
their former behaviors as soon as an intervention stops. Convincing people
to believe in interventions designed not only to reduce but also to prevent
obesity is thus very challenging. Another participant observed that the argu-
ment could be taken as more about messaging than about economics. Sturm
closed with the observation that “we want to maximize health benefits,”
but resources are limited. Ideally, cost-effectiveness analyses would support
optimal decisions, but “we need to make them better,” he said.
REFERENCES
Andreyeva, T., R. Sturm, and J. S. Ringel. 2004. Moderate and severe obesity have large dif-
ferences in health care costs. Obesity Research 12(12):1936-1943.
Ardern, C. I., I. Janssen, R. Ross, and P. T. Katzmarzyk. 2004. Development of health-
related waist circumference thresholds within BMI categories. Obesity Research 12(7):
1094-1103.
Beunen, G., R. M. Malina, J. Lefevre, A. L. Claessens, R. Renson, J. Simons, H. Maes, B.
Vanreusel, and R. Lysens. 1994. Size, fatness and relative fat distribution of males of
contrasting maturity status during adolescence and as adults. International Journal of
Obesity 18(10):670-678.
Cleeman, J. I. 2001. Executive summary of the third report of the National Cholesterol Edu-
cation Program (NCEP) expert panel on detection, evaluation, and treatment of high
blood cholesterol in adults (adult treatment panel III). Journal of the American Medical
Association 285(19):2486-2497.
Cohen, D. A., R. Sturm, M. Scott, T. A. Farley, and R. Blumenthal. 2010. Not enough fruit
and vegetables or too many cookies, candies, salty snacks, and soft drinks? Public Health
Reports 125(1):88-95.
Feng, J., T. A. Glass, F. C. Curriero, W. F. Stewart, and B. S. Schwartz. 2010. The built en-
vironment and obesity: A systematic review of the epidemiologic evidence. Health and
Place 16(2):175-190.
Flegal, K. M., J. A. Shepherd, A. C. Looker, B. I. Graubard, L. G. Borrud, C. L. Ogden, T. B.
Harris, J. E. Everhart, and N. Schenker. 2009. Comparisons of percentage body fat, body
mass index, waist circumference, and waist-stature ratio in adults. American Journal of
Clinical Nutrition 89(2):500-508.
OCR for page 54
54 MEASURING PROGRESS IN OBESITY PREVENTION
Guo, S. S., A. F. Roche, W. C. Chumlea, J. D. Gardner, and R. M. Siervogel. 1994. The pre-
dictive value of childhood body mass index values for overweight at age 35 y. American
Journal of Clinical Nutrition 59(4):810-819.
Hsieh, S. D., H. Yoshinaga, and T. Muto. 2003. Waist-to-height ratio, a simple and practi-
cal index for assessing central fat distribution and metabolic risk in Japanese men and
women. International Journal of Obesity 27(5):610-616.
Levy, D. T., P. L. Mabry, Y. C. Wang, S. Gortmaker, T. T. K. Huang, T. Marsh, M. Moodie,
and B. Swinburn. 2011. Simulation models of obesity: A review of the literature and
implications for research and policy. Obesity Reviews 12(5):378-394.
Magnus, A., M. M. Haby, R. Carter, and B. Swinburn. 2009. The cost-effectiveness of remov-
ing television advertising of high-fat and/or high-sugar food and beverages to Australian
children. International Journal of Obesity 33(10):1094-1102.
Malina, R. M. 2005. Variation in body composition associated with sex and ethnicity. In Hu-
man body composition. 2nd ed., edited by S. B. Heymsfield, T. G. Lohman, Z. M. Wang,
and S. B. Going. Champaign, IL: Human Kinetics. Pp. 271-298, 473-476.
Malina, R. M. 2011. Body mass index: For better or worse. Presented at the Institute of
Medicine Workshop on Measurement Strategies for Accelerating Progress in Obesity
Prevention, March 23, Irvine, CA.
Malina, R. M., and P. T. Katzmarzyk. 1999. Validity of the body mass index as an indica-
tor of the risk and presence of overweight in adolescents. American Journal of Clinical
Nutrition 70(Suppl.):131S-136S.
Malina, R. M., Y. C. Huang, and K. H. Brown. 1995. Subcutaneous adipose tissue dis-
tribution in adolescent girls of four ethnic groups. International Journal of Obesity
19(11):793-797.
Malina, R. M., C. Bouchard, and O. Bar-Or. 2004. Growth, maturation, and physical activity.
2nd ed. Champaign, IL: Human Kinetics.
McKinnon, R. A. 2011. Measures registry & catalogue of surveillance systems. Presented at
the Institute of Medicine Workshop on Measurement Strategies for Accelerating Progress
in Obesity Prevention, March 23, Irvine, CA.
Moodie, M., M. Haby, L. Galvin, B. Swinburn, and R. Carter. 2009. Cost-effectiveness of ac-
tive transport for primary school children—Walking School Bus program. International
Journal of Behavioral Nutrition and Physical Activity 6:63.
Moodie, M. L., R. C. Carter, B. A. Swinburn, and M. M. Haby. 2010. The cost-effectiveness
of Australia’s active after-school communities program. Obesity 18(8):1585-1592.
Moore, L. L., D. Gao, M. L. Bradlee, L. A. Cupples, A. Sundarajan-Ramamurti, M. H.
Proctor, M. Y. Hood, M. R. Singer, and R. C. Ellison. 2003. Does early physical activity
predict body fat change throughout childhood? Preventive Medicine 37(1):10-17.
Redden, D. T., and D. B. Allison. 2003. Nonreplication in genetic association studies of obesity
and diabetes research. Journal of Nutrition 133(11):3323-3326.
Rolland-Cachera, M. F., T. J. Cole, M. Sempe, J. Tichet, C. Rossignol, and A. Charraud.
1991. Body mass index variations: Centiles from birth to 87 years. European Journal of
Clinical Nutrition 45(1):13-21.
Romero-Corral, A., V. K. Somers, J. Sierra-Johnson, R. J. Thomas, M. L. Collazo-Clavell, J.
Korinek, T. G. Allison, J. A. Batsis, F. H. Sert-Kuniyoshi, and F. Lopez-Jimenez. 2008.
Accuracy of body mass index in diagnosing obesity in the adult general population.
International Journal of Obesity 32(6):959-966.
Sacks, G., J. L. Veerman, M. Moodie, and B. Swinburn. 2011. “Traffic-light” nutrition la-
belling and “junk-food” tax: A modelled comparison of cost-effectiveness for obesity
prevention. International Journal of Obesity 35(7):1001-1009.
OCR for page 55
55
LOOKING ACROSS DOMAINS
Sorkin, J. D., D. C. Muller, and R. Andres. 1999. Longitudinal change in the heights of
men and women: Consequential effects on body mass index. Epidemiologic Reviews
21(2):247-260.
Sturm, R. 2001. The myth of medical cost offset. Psychiatric Services 52(6):738-740.
Sturm, R. 2002. The effects of obesity, smoking, and drinking on medical problems and costs.
Health Affairs 21(2):245-253.
WHO (World Health Organization). 2000. Obesity: Preventing and managing the global
pandemic. Technical report series no. 894. Geneva, Switzerland: WHO.
Wu, S., D. Cohen, Y. Shi, M. Pearson, and R. Sturm. 2011. Economic analysis of physical
activity interventions. American Journal of Preventive Medicine 40(2):149-158.
OCR for page 56
