Use and Misuse of BMI Categories.

Body mass index (BMI) was introduced in the 19th century as a measure of weight relative to height. Before the late 20th century, overweight and obesity were not considered a population-wide health risk, but the advent of new weight loss drugs in the 1990s accelerated the medicalization of BMI. A BMI category labeled obesity was adopted in 1997 by a World Health Organization consultation and subsequently by the US government. Language in the National Coverage Determinations Manual stating that "obesity itself cannot be considered an illness" was removed in 2004, allowing reimbursement for weight loss treatments. In 2013, the American Medical Association declared obesity to be a disease. Yet the focus on BMI categories and on weight loss has yielded few health benefits and contributes to weight-related discrimination and other potential harms.

A Female Career in Research.

After a long career at the National Center for Health Statistics, I retired and joined the Stanford Prevention Research Center as an unpaid associate. I was once described by a former US Food and Drug Administration commissioner as "one of the great epidemiologists." The chair of the Harvard nutrition department, speaking on National Public Radio, once described my research as "rubbish." Both may be exaggerations. Here I address some of the events that led to these contrasting descriptions. I also address the extent to which the so-called Matilda effect may have influenced my career. Are women in science on an equal footing with men? The Matilda effect suggests not. Unlike the Matthew effect for scientists, whereby those of higher prestige accrue a disproportionate share of recognition and rewards, the Matilda effect proposes that women scientists are systematically undervalued and underrecognized. I could never get a faculty job and was often treated like an underling. Nonetheless I persevered to publish highly cited research on several high-profile and sometimes controversial topics. Though overt sexism in science and workplaces has diminished over the course of my career, progress toward eliminating unconscious bias has been slower. The Matthew and Matilda effects are still powerful forces that distort incentives and rewards in science.

Evaluation of a suggested novel method to adjust BMI calculated from self-reported weight and height for measurement error.

OBJECTIVE: In 2019, Ward et al. proposed a method to adjust BMI calculated from self-reported weight and height for bias relative to measured data. They did not evaluate the adjusted values relative to measured BMI values for the same individuals. METHODS: A large data set (n = 37,439) with both measured and self-reported weight and height was randomly divided into two groups. The proposed method was used to adjust the BMI values in one group to the measured data from the other group. The adjusted values were then compared with the measured values for the same individuals. RESULTS: Before adjustment, 24.9% were incorrectly classified relative to measured BMI categories, including 7.9% in too high a category; after adjustment, 24.3% were incorrectly classified, with 12.8% in too high a category. The variance of the difference was unchanged. The adjustments reduced some errors and introduced new errors. At an individual level, results were unpredictable. CONCLUSIONS: The suggested method has little effect on misclassification, can introduce new errors, and could magnify errors associated with factors, such as age, race, educational level, or other characteristics. State-level estimates and projections of obesity prevalence from values adjusted by this method may be incorrect.

The obesity wars and the education of a researcher: A personal account.

A naïve researcher published a scientific article in a respectable journal. She thought her article was straightforward and defensible. It used only publicly available data, and her findings were consistent with much of the literature on the topic. Her coauthors included two distinguished statisticians. To her surprise her publication was met with unusual attacks from some unexpected sources within the research community. These attacks were by and large not pursued through normal channels of scientific discussion. Her research became the target of an aggressive campaign that included insults, errors, misinformation, social media posts, behind-the-scenes gossip and maneuvers, and complaints to her employer. The goal appeared to be to undermine and discredit her work. The controversy was something deliberately manufactured, and the attacks primarily consisted of repeated assertions of preconceived opinions. She learned first-hand the antagonism that could be provoked by inconvenient scientific findings. Guidelines and recommendations should be based on objective and unbiased data. Development of public health policy and clinical recommendations is complex and needs to be evidence-based rather than belief-based. This can be challenging when a hot-button topic is involved.

The perils of using predicted values in place of observed covariates: an example of predicted values of body composition and mortality risk.

BACKGROUND: Several studies have assessed the relation of body composition to health outcomes by using values of fat and lean mass that were not measured but instead were predicted from anthropometric variables such as weight and height. Little research has been done on how substituting predicted values for measured covariates might affect analytic results. OBJECTIVES: We aimed to explore statistical issues causing bias in analytical studies that use predicted rather than measured values of body composition. METHODS: We used data from 8014 adults ≥40 y old included in the 1999-2006 US NHANES. We evaluated the relations of predicted total body fat (TF) and predicted total body lean mass (TLM) with all-cause mortality. We then repeated the evaluation using measured body composition variables from DXA. Quintiles and restricted cubic splines allowed flexible modeling of the HRs in unadjusted and multivariable-adjusted Cox regression models. RESULTS: The patterns of associations between body composition and all-cause mortality depended on whether body composition was defined using predicted values or DXA measurements. The largest differences were observed in multivariable-adjusted models which mutually adjusted for both TF and TLM. For instance, compared with analyses based on DXA measurements, analyses using predicted values for males overestimated the HRs for TF in splines and in quintiles [HRs (95% CIs) for fourth and fifth quintiles compared with first quintile, DXA: 1.22 (0.88, 1.70) and 1.46 (0.99, 2.14); predicted: 1.86 (1.29, 2.67) and 3.24 (2.02, 5.21)]. CONCLUSIONS: It is important for researchers to be aware of the potential pitfalls and limitations inherent in the substitution of predicted values for measured covariates in order to draw proper conclusions from such studies.

Use and reporting of Bland-Altman analyses in studies of self-reported versus measured weight and height.

BACKGROUND/OBJECTIVES: Bland-Altman methods for assessing the agreement between two measures are highly cited. However, these methods may often not be used to assess agreement, and when used, they are not always presented or interpreted correctly. Our objective was to evaluate the use and the quality of reporting of Bland-Altman analyses in studies that compare self-reported with measured weight and height. METHODS: We evaluated the use of Bland-Altman methods in 394 published articles that compared self-reported and measured weight and height data for adolescents or adults. Six reporting criteria were developed: assessment of the normality of the distribution of differences, a complete and correctly labeled Bland-Altman plot displaying the mean difference and limits of agreement (LOA), numerical values and confidence intervals, standard errors, or standard deviations for mean difference, numerical values of LOA, confidence intervals for LOA, and prespecified criteria for acceptable LOA. RESULTS: Only 72/394 (18%) studies comparing self-reported with measured weight and height or BMI used some form of Bland-Altman analyses. No study using Bland-Altman analyses satisfied more than four of the six criteria. Of the 72 studies, 64 gave mean differences along with confidence intervals or standard deviations, 55 provided complete Bland-Altman plots that were appropriately labeled and described, 37 provided numerical values for LOA, 4 reported that they examined the normality of the distribution of differences, 3 provided confidence intervals for LOA, and 3 had prespecified criteria for agreement. CONCLUSIONS: Bland-Altman methods appear to be infrequently used in studies comparing measured with self-reported weight, height, or BMI, and key information is missing in many of those that do use Bland-Altman methods. Future directions would be defining acceptable LOA values and improving the reporting and application of Bland-Altman methods in studies of self-reported anthropometry.

Comparisons of Self-Reported and Measured Height and Weight, BMI, and Obesity Prevalence from National Surveys: 1999-2016.

OBJECTIVE: The aim of this study was to compare national estimates of self-reported and measured height and weight, BMI, and obesity prevalence among adults from US surveys. METHODS: Self-reported height and weight data came from the National Health and Nutrition Examination Survey (NHANES), the National Health Interview Survey, and the Behavioral Risk Factor Surveillance System for the years 1999 to 2016. Measured height and weight data were available from NHANES. BMI was calculated from height and weight; obesity was defined as BMI ≥ 30. RESULTS: In all three surveys, mean self-reported height was higher than mean measured height in NHANES for both men and women. Mean BMI from self-reported data was lower than mean BMI from measured data across all surveys. For women, mean self-reported weight, BMI, and obesity prevalence in the National Health Interview Survey and Behavioral Risk Factor Surveillance System were lower than self-report in NHANES. The distribution of BMI was narrower for self-reported than for measured data, leading to lower estimates of obesity prevalence. CONCLUSIONS: Self-reported height, weight, BMI, and obesity prevalence were not identical across the three surveys, particularly for women. Patterns of misreporting of height and weight and their effects on BMI and obesity prevalence are complex.

Flawed methods and inappropriate conclusions for health policy on overweight and obesity: the Global BMI Mortality Collaboration meta-analysis.

Guideline recommendations and health policy decisions rely on evidence from clinical and epidemiological studies. Adequate methodology and appropriate conclusions are essential to support healthcare and health policy decisions. An analysis of body mass index and mortality by the Global BMI Mortality Collaboration (GBMC) concluded that the association of excess body weight with higher mortality was similar worldwide and that overweight and obesity should be combated everywhere. To reach this conclusion, the GBMC used highly selected data, rather than a systematic approach. The GBMC initially chose individual participant data from 239 prospective studies with approximately 10.6 million participants. The GBMC then excluded over 60% of data and over 75% of fatal events by eliminating all cases with any reported disease at baseline or smoking history and all events within the first 5 years of follow-up. After applying these restrictions, the association of overweight with lower mortality was reversed and the association of obesity with higher mortality was increased. Given the major flaws in the selection process, in the adequacy of the data, in the data analysis, and in the interpretation, the GBMC conclusions should be viewed sceptically as a guide to action, either for clinical decisions or for public health in general. The flawed conclusion that overweight is uniformly associated with substantially increased risk of death and thus should be combated in any circumstances may lead not only to unjustified treatment efforts and potential harm in a wide range of clinical conditions but also to a tremendous waste of resources.

Excess Deaths Associated With Underweight, Overweight, and Obesity: An Evaluation of Potential Bias.

As the prevalence of obesity has increased over time in the United States (1,2), concern over the association between body weight and excess mortality also increased. In 2005, an analysis of estimated excess deaths, relative to the normal weight category (body mass index [BMI] 18.5-24.9), that were associated with underweight (BMI less than 18.5), overweight (BMI 25.0-29.9), and obesity (BMI greater than or equal to 30) in U.S. adults in 2000 was published (3). Both underweight and obesity, particularly higher levels of obesity, were associated with increased mortality relative to the normal weight category. Obesity was estimated to be associated with 111,909 excess deaths (95% confidence interval [CI]: 53,754 to 170,064) in 2000 relative to the normal weight category, and underweight with 33,746 excess deaths (95% CI: 15,726 to 51,766). Overweight was associated with reduced mortality (-86,094 deaths; 95% CI: -161,223 to -10,966). This report evaluates several potential sources of bias in that analysis.

The Obesity Paradox: A Misleading Term That Should Be Abandoned.

The term "obesity paradox" is a figure of speech, not a scientific term. The term has no precise definition and has been used to describe numerous observations that have little in common other than the finding of an association of obesity with a favorable outcome. The terminology has led to misunderstandings among researchers and the public alike. It's time for authors and editors to abandon the use of this term. Simply labeling counterintuitive findings as the "obesity paradox" adds no value. Unexpected findings should not be viewed negatively; such findings can lead to new knowledge, better treatments, and scientific advances.

Bias in Hazard Ratios Arising From Misclassification According to Self-Reported Weight and Height in Observational Studies of Body Mass Index and Mortality.

Misclassification of body mass index (BMI) categories arising from self-reported weight and height can bias hazard ratios in studies of BMI and mortality. We examined the effects on hazard ratios of such misclassification using national US survey data for 1976 through 2010 that had both measured and self-reported weight and height along with mortality follow-up for 48,763 adults and a subset of 17,405 healthy never-smokers. BMI was categorized as <22.5 (low), 22.5-24.9 (referent), 25.0-29.9 (overweight), 30.0-34.9 (class I obesity), and ≥35.0 (class II-III obesity). Misreporting at higher BMI categories tended to bias hazard ratios upwards for those categories, but that effect was augmented, counterbalanced, or even reversed by misreporting in other BMI categories, in particular those that affected the reference category. For example, among healthy male never-smokers, misclassifications affecting the overweight and the reference categories changed the hazard ratio for overweight from 0.85 with measured data to 1.24 with self-reported data. Both the magnitude and direction of bias varied according to the underlying hazard ratios in measured data, showing that findings on bias from one study should not be extrapolated to a study with different underlying hazard ratios. Because of misclassification effects, self-reported weight and height cannot reliably indicate the lowest-risk BMI category.

A meta-analysis but not a systematic review: an evaluation of the Global BMI Mortality Collaboration.

Meta-analyses of individual participant data (MIPDs) offer many advantages and are considered the highest level of evidence. However, MIPDs can be seriously compromised when they are not solidly founded upon a systematic review. These data-intensive collaborative projects may be led by experts who already have deep knowledge of the literature in the field and of the results of published studies and how these results vary based on different analytical approaches. If investigators tailor the searches, eligibility criteria, and analysis plan of the MIPD, they run the risk of reaching foregone conclusions. We exemplify this potential bias in a MIPD on the association of body mass index with mortality conducted by a collaboration of outstanding and extremely knowledgeable investigators. Contrary to a previous meta-analysis of group data that used a systematic review approach, the MIPD did not seem to use a formal search: it considered 239 studies, of which the senior author was previously aware of at least 238, and it violated its own listed eligibility criteria to include those studies and exclude other studies. It also preferred an analysis plan that was also known to give a specific direction of effects in already published results of most of the included evidence. MIPDs where results of constituent studies are already largely known need safeguards to their validity. These may include careful systematic searches, adherence to the Preferred Reporting Items for Systematic Review and Meta-Analyses of individual participant data guidelines, and exploration of the robustness of results with different analyses. They should also avoid selective emphasis on foregone conclusions based on previously known results with specific analytical choices.