Whose Voice Counts? A Critical Examination of Discourses Surrounding the Body Mass Index

Abstract

The Body Mass Index (BMI) is almost universally used by health practitioners and the general public to determine how to classify a person's body in terms of relative weight. The author traces the story of the vocal actors who developed and championed the BMI as it moved from obscurity to occupy a central position in dominant discourse about body size and its relationship to health. She also highlights the voice of a fat advocate who is challenging not only the BMI, but the general dominant discourse concerning the relationship between body size, health, and personal value. The article serves as a call to all communication researchers working in health contexts to carefully consider how their work might reinforce or challenge this dominant discourse.

Keywords:

• body mass index
• health communication
• fat studies
• weight discourse
• weight stigma

The definition of disease is neither natural nor neutral, but is always a social construction (Lock & Nichter, 2002) that privileges some voices over others (Singer, 1995). Jordan (1997) argued that medicine represents a specialized system of knowledge that defines health experiences from an authoritative position. Lock and Nichter described western biomedicine's “obsession” (p. 26) with control and categorization that tries to fit lived experience into artificially clear-cut categories. This drive toward categorization not only creates unnecessary boundary conditions for specific health experiences, but for the body itself (Baer, Singer, & Susser, 2003). As such, biomedical discourse has the ability to define a “correct body” (Scheper-Hughes & Lock, 1987, p. 25) in terms of various attributes.

A striking example of biomedical discourse on the “correct body” is the widespread use of the Body Mass Index (BMI). The BMI is typically used as a metric for determining whether an individual's weight is in a “healthy” range, that is, whether a person has excess fat (adipose) tissue. Many scholars have voiced critiques of the BMI as a medical tool (e.g., Burkhauser & Cawley, 2008). However, the BMI is more than a medical tool. It is an element of medical discourse that classifies individuals according to weight; this classification can have chilling moral, ethical, and political ramifications. In this article I examine the BMI as an example of the privileged definitional space occupied by biomedical discourse, exposing its socially constructed nature, and challenging health scholars to consider how their work may reify or challenge this discourse. I explore the discursive journey of the BMI from its development, to its widespread use, to its contemporary criticisms, through the lens of Jordan's (1997) notion of authoritative knowledge.

AUTHORITATIVE KNOWLEDGE AND THE PRIVILEGED VOICE OF BIOMEDICINE

Jordan's (1997) notion of authoritative knowledge in the health context is a useful way to understand the current use of BMI. Jordan used a Foucauldian lens to argue that medicine, as a specialized form of knowledge and area of professional practice, requires ascribing power to certain types of knowledge and to the people who possess it. Lock and Nichter (2002) echoed the position that scientific medical knowledge is privileged and powerful, since many health claims outside biomedicine utilize scientific language and tools to establish credibility. Jordan argued that once a particular type of knowledge is seen as most legitimate, it becomes the most counted form of knowledge—regardless of its accuracy or correctness—and possessors of that knowledge have the power to be in charge of what counts as a fact. From Lock and Nichter's perspective, the possessors of such knowledge are also the only ones who may speak about the topic scientifically, and thus claim legitimacy in a biomedical system of health. The BMI offers a scientific and legitimate way for lay people and health professionals to speak authoritatively about weight and health.

Lock and Nichter (2002) argued that one of the most powerful ways to assert knowledge authority is to name, define, and categorize. The BMI allows researchers, medical professionals, insurance companies, government agencies, and even individuals to classify themselves and others according to weight. It creates discrete categories of people: underweight, normal weight, overweight, and obese. Thus, it is a powerful tool for categorizing not only a supposed physical condition, but also people. Singer (1995) discussed how biomedical definitions of disease have the power to guide government policies regarding health research, health care, and public health priorities. Indeed, classifying certain traits as risk factors prompts political attention to those health issues. The powerful medical-political discourse about health issues not only drives large-scale policies, but also has important implications for individuals (Scheper-Hughes & Lock, 1987). The BMI is a central aspect of contemporary biomedical discourse on physical health, and this article now turns to the key actors responsible for shaping that discourse.

THE STORY OF THE BMI

Quetelet and the Development of the BMI

To understand how the BMI came to occupy a prominent role in dominant biomedical discourse, one must begin with the statistician Adolphe Quetelet (1796–1874) and his mathematical formula for fitting weight distributions in adult populations to the normal curve. Eknoyan (2008) provided a detailed account of Quetelet's career and the development of the Quetelet index, later known as the BMI. As Eknoyan described, Quetelet was a statistician interested in documenting normal distributions of human characteristics across populations. Normal distributions rely on average values in order to make sense of all the values in the population; all values are fitted around the mean in order to create a normal distribution (Eknoyan, 2008) One of Quetelet's primary goals was to test the validity of utilizing a normal curve to describe the distribution of social phenomena (such as personality or aptitude), and later, weight. Thus, he was intrigued when weight distributions of populations seemed not to fit this model and worked to fit the distribution of human weight to the normal curve (Eknoyan, 2008)

Quetelet authored the book A Treatise on Man and the Development of His Aptitudes in 1835, in which he summarized his findings regarding the distribution of human characteristics in a population. He argued that probability mathematics could be used to examine trends in social issues such as intelligence, moral character, or weight. Eknoyan (2008) noted that this argument, and the evidence presented in his book, has led some experts to consider Quetelet the founder of social science (see Lazarsfeld, 1961, cited in Eknoyan, 2008). Thus, Quetelet has been an extremely influential voice in dominant scientific discourse. From this powerful voice came a mathematical formula that he believed could accurately capture his observation of the distribution of human weight in the Netherlands. This formula, which allows human weight distribution to fit a normal curve distribution, described his observation that “weight increases nearly as the square of the height” after growth periods in infancy and adolescence are finished (Quetelet, 1968/1842, cited in Eknoyan, 2008, p. 49). This is the basis of the current BMI formula, which is the proportion of weight in kilograms to squared height in meters.

Quetelet never considered himself a medical expert and did not create his formula as a means for assessing individual body composition or adiposity. It was simply a reflection of a pattern he had observed over the course of many studies of the Dutch population during the 19th century. Furthermore, for a long time, this formula for the proportion of weight to height was not well known outside of a small group of researchers familiar with his other work (Eknoyan, 2008).

Keys and the Scientific Endorsement of the BMI Over Other Indices

Almost a century later, Keys, Fidanza, Karovnen, Kimura, and Taylor (1972) examined Quetelet's index as part of a larger study that set out to determine which weight-to-height ratio could best reflect body fatness. By the time Keys et al. were investigating the relative accuracy and value of various weight-to-height ratios, the concern was no longer finding an index that best captured the distribution of weight in a population. Instead, there was considerable concern over the ability to accurately label a person as fat. Thus, in Keys et al.'s landmark study, the researchers were tasked with identifying a scientific measure that could support a body-labeling discourse wherein certain body types were normalized and others marginalized.

This concern for correct labeling of persons as fat or obese was partially propelled by the insurance industry's attempts to make inferences about fatness, based on relative weight, in order to determine relative health risks and to develop health insurance policies correspondent to those risks. Beginning as early as 1912, life insurance companies investigated the links between weight and mortality. By 1959, the Metropolitan Life Insurance Company (MetLife)had released a height and weight chart that was frequently used in the insurance industry to determine coverage costs for individuals, and was starting to gain some traction in the scientific community as well (Keys et al., 1972).

Within the scientific community, one prominent scientist interested in finding consistent ways to label individuals who fall outside average weight as obese was Ancel Keys. Keys openly voiced his personal ideas concerning the relative value of different body sizes; his opinions were documented in a 1961 issue of Time Magazine that depicted Keys on the cover. Keys was interviewed for an article entitled “Medicine: The Fat of the Land,” which detailed Keys' take on the rising average weight during that time. The article painted him as a foremost expert on the subject—an expert who expressed a deep dislike for excess fat and who repeatedly demonstrated a disdain for those who do not obtain or maintain a “healthy” weight. When asked about historic religious beliefs that cast excess weight as sinful, Keys responded: “Maybe if the idea got around again that obesity is immoral, the fat man would start to think” (p. 2). Later, the article included another telling quote from Keys: “People should know the facts … Then if they want to eat themselves to death, let them” (p. 8). Clearly Keys had strong opinions regarding the negative effects of excess weight gain caused by excessive eating; he also expressed—at this point in his career—a strong sense that weight is under a person's control and that a person's weight is somehow reflective of his or her moral character.

About a decade later, Keys et al. (1972) conducted their landmark study comparing various indices of relative weight. A number of indices using various weight-to-height ratios were in use in both the scientific and insurance communities around the time of this publication. And, at that point, there was growing concern over which weight-to-height ratio did the best job of identifying true fatness. There was also concern in the scientific community about the validity and usefulness of the ubiquitous height/weight charts provided by insurance companies, which were being used—even in the scientific literature—as points of comparison for evaluating the relative weight of individuals. The height/weight charts essentially provided a one-to-one ratio of weight to height, and there was some question as to whether they could consistently and accurately reflect relative fatness across different populations, or even the same populations at different times. The one-to-one weight-to-height ratio was and is problematic because “body form does not remain constant with increasing length” (Keys et al., p. 330), so it does not provide information about the relative fatness of a person.

Hence, Keys et al. (1972) tested the validity of various height-to-weight ratios by correlating measures on those ratios to measures from skinfold and body density tests. Interestingly, Keys et al. noted that comparing height-to-weight ratios with skinfold tests is difficult, since results from skinfold tests do not fit a normal curve. Indeed, Keys et al. transformed the skinfold data in order for it to fit a normal curve model so that correlations and regressions with weight indices could be calculated. Keys et al. observed the same issue that plagued Quetelet over a century earlier: variations in human body composition did not follow the assumptions of a normal curve. Yet each of these researchers pursued explanations of human phenomena that would more easily fit within the boundaries of scientific thought. That is, rather than abandoning, or at least questioning, the idea that all human phenomena are distributed normally in populations, these researchers moved forward in their quest to reify scientific discourse that relies on mathematical operations to explain complex human phenomena.

Incredibly, Keys et al.'s (1972) seminal study is often cited as the basis for using the BMI as a proxy measure for obesity, or at least relative weight, even though the study does not provide solid evidence for this conclusion. In Keys et al.'s own words:

Again the body mass index (W/H²) proves to be, if not fully satisfactory, at least as good as any other relative weight index as an indicator of relative obesity. Still … not more than half of the total variance of body fatness is accounted for by the regression of fatness on the body mass index. (p. 339, emphasis added)

This demonstrates that, even in the first study to recommend the use of BMI, there is an acknowledgment that it is far from a perfect indicator of body fatness. Keys et al. (1976) explained that they recommended the use of the BMI as a measure of body fat to be used in population-based studies, not because of its accuracy or validity, but because it is far simpler to use than body density tests of fatness (the most accurate) or skinfold tests of fatness (moderately accurate). For insurance agencies, public health officials, lay persons, or medical professionals, the computation of a weight-to-height ratio represented a quick heuristic for gauging relative fatness—even though it has never been considered highly accurate.

NIH, WHO, and the Political Endorsement of the BMI

The National Institutes of Health (NIH) and the World Health Organization (WHO) are strong, vocal proponents of the use of BMI (Mascie-Taylor & Goto, 2007). Both organizations have a strong voice in discourse concerning how the BMI should be interpreted and utilized on national and international levels to guide public policy and diagnose individuals. The position of NIH on the interpretation and use of the BMI has changed substantially in the past few decades. In the mid-1980s, NIH included BMI as one type of measure of fatness. In fact, at that point, they were still offering advice about the best use of the height and weight charts compiled by MetLife. Representatives from MetLife served on the planning committee for the NIH 1985 report on obesity, which embraced MetLife's definition of “overweight” as at least 20% above desirable weight. However, the NIH recommendations for the use of BMI were limited and tentative; they only differentiated between “desirable weight” and “overweight” (BMI = 27.2 for men, 26.9 for women), and cautioned against hard-and-fast categorizations of people into certain weight classes (National Institutes of Health [NIH], 1985).

By the late 1990s, both WHO and NIH had abandoned a cautionary approach to the interpretation and use of BMI, instead voicing their endorsement of hard-and-fast categories for classification of individuals as underweight (≤18.4), normal weight (18.5–24.9), overweight (25–29.9), obese class I (30–34.9), obese class II (35–39.9), and extremely obese (≥40). Note that the “normal weight” range had shrunk, which resulted in the classification of previously “normal” bodies as “overweight” with this new nomenclature. In addition, these reports eliminated the different BMI for men and women (NIH, 1998; World Health Organization [WHO], 1999), despite the fact that the BMI is not equally accurate for men and women due to their differing body compositions (Mascie-Taylor & Goto, 2007). Both reports also noted that there are variations in the applicability and validity of BMI as a measure of body composition across different groups of people, but still recommended the universal application of the BMI metric.

Perhaps the most striking shift in these organizations' recommendations for the BMI was the unequivocal endorsement of the BMI as an individual diagnostic tool rather than as a mathematical formula to be used in assessing population-based weight trends (as suggested by Keys et al., 1972). The authors of the NIH report (1998) specifically noted that “the BMI should be used to assess overweight and obesity and to monitor changes in body weight” (p. 14). A press release about the report stated that “the panel recommends BMI be determined in all adults” (National Heart, Lung, & Blood Institute, 1998, p. 1).

The current position of both the WHO and NIH is essentially unchanged from the seminal reports in the late 1990s. Given that both reports acknowledged that BMI might be differentially accurate or applicable for different populations, in 2004 the WHO began to systematically assess whether different BMI classifications may be needed for different ethnic groups (Mascie-Taylor & Goto, 2007). In particular, the WHO Expert Consultation (2004) reviewed the applicability of universal BMI cutoffs for Asian populations, but found too much variance in associations between certain BMI ranges and health outcomes to warrant the adoption of new cutoff points for that population. The authors of the report recommended maintaining current cutoff points and applying them universally (WHO, 2004). Currently, both the WHO (2010) and NIH (2010) recommend the use of universal cutoff points for determining weight classifications for all populations. Just a few months ago, Time Magazine reported on a study funded in part by the NIH and published in the New England Journal of Medicine that touted a scientifically determined ideal weight (Melnick, 2010). And, just as it did 50 years ago, Time Magazine used the authoritative voice of biomedicine to answer the question, “How fat is too fat?” (Melnick). The article tells the story of a marked difference between BMI categories that clearly identifies an ideal BMI category of 20.0–24.9. Readers are left with a sense that medical science has once again found evidence for its ideological stance on fatness.

However, close examination of the findings from that study (i.e., Gonzalez et al., 2010) reveals that the story is not quite so simple. Gonzalez et al. used the BMI category of 22.5–24.9 as the reference group because it exhibited the lowest rates of mortality. As with any study that reports comparative risk analyses, the mortality rate in this category is not identified explicitly, but inferred from comparison with other groups. Not surprisingly, the relative mortality risk of those in the 20.0–22.4 range was not significantly different from those in the comparison group. And the authors report, in their conclusions, a significant difference between the reference group and the 25.0–29.9 BMI group. However, in their initial analyses, that group was also broken into two sections: 25.0–27.4 and 27.5–29.9. And, in most comparisons in the study, the 25.0–27.4 group did not vary significantly from the 22.5–24.9 reference group. It also did not vary significantly from the 20.0–22.4 group. The significant jump in mortality rates occurred in the 27.5–29.9 group and thereafter.

Gonzalez et al. (2010) did not provide a detailed justification for how and why they broke down the BMI categories for analysis or how that affected their interpretation of the results. It is perplexing why they initially broke the BMI categories into smaller ranges for analysis, only to lump together the findings for the 25.0–29.9 range in their interpretation while keeping the 20.0–22.5 range separate from the 22.5–24.9 comparison group. When the findings are interpreted using the smaller BMI ranges and how they relate to the comparison group, one can see some support for the earlier interpretations of NIH (1985) and their guidelines for the use of BMI that indicated a BMI of around 27 is where risk increases significantly. But the authors used the BMI cutoffs identified in the 1998 NIH report, rather than the cutoffs used in the study, when presenting and discussing the results. This allowed the results to fit neatly within the framework laid out in the 1998 NIH report, but did not provide an entirely accurate representation of the data. Therefore, the presentation and discussion of the results arbitrarily and incorrectly reinforced the NIH-mandated cutoffs for the interpretation of BMI scores.

CRITIQUING THE BODY MASS INDEX

Tracing the history and development of the BMI, and more specifically the categories within the BMI, illuminates the very fluid, socially constructed nature of those categories and the powerful voices that have shaped their trajectory. It is a prime example of how authoritative knowledge (Jordan, 1997) is voiced, and how it is elevated above other ways of knowing not because of its accuracy or validity, but because it most readily serves the needs of those who dictate biomedical discourse and health policy (Jordan). And, as Lock and Nichter (2004) argued, there is nothing natural or neutral about the way we define disease; neither is there anything natural or neutral about the way we define body mass index or use the BMI as a tool to define what it means to be overweight and apply that term to individuals. Despite the limitations of such a category for conducting valid research and reaching tenable conclusions, the BMI remains a powerful tool for guiding research and classifying people.

Still, biomedical knowledge and discourse are not monolithic. Even within the medical field, there is widespread acknowledgment of the serious flaws in the index (for an overview, see Burkhauser & Cawley, 2008). Lewis et al. (2010) noted that the shifting categorization and classification of people as normal weight, overweight, and obese has led to a number of disputes and confusing interpretations of results in even the scientific literature on body composition and health outcomes. The overweight category is particularly ambiguous with respect to correctly identifying a fatty body composition and its subsequent health risks (Burkhauser & Cawley; Lewis et al.).

Beyond issues of the validity of the category for accurately assessing the amount of adipose tissue in a body (or a fatty body composition), the category of “overweight” has damaging social ramifications. Creating a category labeled “overweight” does at least four things. First, it reinforces the idea that there is such a thing as a “normal” weight. This idea goes back to Quetelet's early work with the subject, and his quest to fit human variations in body composition into a normal curve (Eknoyan, 2008). Though a normal curve does not provide any indication of what is normal either socially or medically, it does normalize data about population weight so that comparative weight categories—among them, a normal category—can be created. What occurs as a result of this framing is that weight can only be understood comparatively, and only some members of society have the means and authority to establish a method of comparison (Jordan, 1997; Lock & Nichter, 2002). Second, this discourse about normal weight propels the establishment of public policy aimed at moving the bulk of the population toward normal (e.g., Let's Move, 2010). Third, praising normal-size bodies has the (un)intended effect of stigmatizing large bodies. Indeed, large-bodied people report being stigmatized by medical professionals, educators, peers, and romantic partners (Crandall, 1994). Fourth, this discourse about normal weight privileges some voices—such as NIH, WHO, or the current administration—over others. In order to challenge such dominant discourse, I now highlights voices critiquing the BMI.

Wann and the Personal Effects of BMI Classifications

As Burkhauser and Cawley (2008) documented, within the medical community there is some debate as to the accuracy and usefulness of the BMI as a weight-classification tool. Indeed, even Ancel Keys (1989) concluded that fatness (as measured by the BMI or any other tool) does not have an independent, direct relationship to mortality; rather, extreme weight—on both ends—increases risks, while most weights across the spectrum show no relationship to mortality risk. Unfortunately, by the time Keys' work in the 1980s began to question the weight–health relationship, his voice was challenging rather than reinforcing dominant ideas—and was therefore not as widely recognized as it had been in the previous decades.

In recent years, researchers from fields such as psychology (Cogan, Smith, & Maine, 2009), public health (Lewis, et al., 2010), and nutrition (Bacon & Aphramor, 2011) have also voiced concern over the negative mental and emotional effects of the BMI and its relationship to body-size stigma. A voice less often highlighted in academic discourse is that of non-academic fat people sharing their experiences and fighting weight-based stigma.

In order to address the silenced voice of non-academic fat people within academic discourse on the BMI and body-weight stigma, a recent panel at the National Communication Association annual conference brought together health communication researchers and self-identified fat advocates. Fat advocates fight discrimination and stereotyping of people with large bodies, and often embrace Health at Every Size (Bacon, 2010) initiatives ¹ that encourage healthy behaviors (e.g., physical activity, nutritious eating) rather than weight loss. Marilyn Wann, a well-known author (Fat!So?: Because You Don't Have to Apologize for Your Size [1998]) and fat advocate, participated in the panel and provided some of the most striking comments.

Wann described the Let's Move campaign as a government-endorsed “genocide” of fat people. From Wann's perspective, the Let's Move campaign conveys that the United States, as a nation, wants to eliminate all fat people from the population—by changing their bodies into smaller bodies. Such a campaign reinforces the idea that a limited range of body size and shape is healthy and desirable, and suggests that everyone conform their bodies to this desirable shape. This echoes the “official” and “scientific” perspective on the issue by Gonzalez et al. (2010), and the NIH (2010). For Wann, campaigns that promote weight loss in order to achieve a normal body are threatening to her self-identity, since it stems in part from her physical identity as a fat woman. Returning to genocide as a rhetorical device, Wann et al. (2010) said, “If there were no more fat people in America tomorrow, I would no longer exist.”

In many ways, Wann's comments reflected Jordan's (1997) concept of authoritative knowledge, and her sense that medical and governmental authorities had been given free rein to define and control her body, and to define and control the way she related to her body. The ability to define body compositions as normal or abnormal is largely dependent on having a nomenclature like the BMI that allows authoritative voices to claim that a person's body is overweight or obese. Wann often derogated the use of the term “overweight,” saying, “over what weight? Compared to who?” (Wann et al., 2010). Such discursive turns also critique power positions and question who has the right to have an authoritative voice on this subject. This is an especially important concern for something like the BMI, which defines an individual's body via a biomedical voice while silencing the individual.

Wann not only voiced concern over her own experiences, but also over the experiences of school-aged children who are growing up with the (unintended) message of the Let's Move campaign that equates fat with bad, and by extension defines fat people as bad people—bad people that the government wants to eradicate. Wann specifically cited Michelle Obama's intimation that her daughters were approaching an unhealthy weight—according to BMI categories (Associated Press, 2010). When a child hears that her weight is unhealthy, and simultaneously hears that the nation at large desires to eradicate the presence of unhealthy weight in the population, she may feel as though she is an unwanted type of person (Wann et al., 2010). At the very least, such discourse can create a negative body image for children whose body does not conform to the norm. On a broader scale, it also implicitly endorses the valuation and moralization of different body types (Scheper-Hughes & Lock, 1987). Negative attitudes toward and discrimination against fat people are well documented (for a review, see Crandall, 1994) and this type of campaign and national discourse only serves to reinforce those types of attitudes.

CONCLUSIONS

At both the national policy and individual experiential levels, the continued use of BMI as a health indicator poses serious threats to the health of our bodies and the body politic. In this article I have traced the story of the BMI, demonstrating that it is a socially constructed instrument used to categorize people according to weight, rather than a value-neutral computation that accurately represents a particular health status either in a population or an individual. In interpreting this story, one notices the feedback loop between societal mores concerning the relative value or acceptability of certain body sizes and scientific literature that provides evidence for the relationship between certain body sizes and health outcomes.

The only way to break this feedback loop is for researchers and fat advocates to strengthen their presence and voice in this debate. As Lock and Nichter (2002) argued, it will be important for those critiquing the dominant paradigm to appear scientific; thus, quality research should back up claims to discontinue using the BMI as the primary measure of fatness, or to at least revise the current BMI classifications. This is an especially salient concern for this context as both the NIH and WHO are committed to evidence-based research, and will therefore not change their policies without data to challenge them. The road to size equality, and value-neutral measures of body composition—should they be found useful—will be long and arduous, but worth traversing.

Notes

1. Health at Every Size initiatives have also begun to receive academic attention and are beginning to be used as a framework for public health interventions (Bacon, 2010; Bacon & Aphramor, 2011).