Articles by ECRs

Why do we still use the BMI?
A call for change

Georgie Burton discusses the (in)appropriateness of the Body Mass Index (BMI), focusing on its history, discrimination, and recommendations– if such measures should be used at all.

Before I begin, I wish to state that this article is not intended as health advice, nor an attack on any previous usage of the Body Mass Index (hereafter, BMI). It is however, intended to raise awareness of its origins and to encourage reflection on subsequent use. Having been ingrained in our everyday culture for so long, it can be easy to not think critically about it – this piece aims to change that.

History
BMI was first coined by Keys et al. (1972), drawing inspiration from Quetelet’s population statistics (Bogin & Varela-Silva, 2013) and based upon life insurance data available at the time. Consequently, they created a formula regarding an individual's body “with relatively little consideration of the evaluation of body composition, nutritional status or adiposity” (p. 340). Although the sample size for this study was a respectable 7,422, it only included men. Furthermore, despite having a range of ethnicities (American, Japanese, Finnish, Italian, and Bantu), the Bantu sample was 9-20 times smaller than the others. The samples were also not age matched. Despite this, the authors declared that it had “applicability to all populations at all times” (p.664) – a statement that if published today, would be rejected on the grounds of a lack of evidence. In an attempt at validity, they compared the BMI to the skinfold and water displacement tests (other measures of fat). Although skinfold thickness data was missing from the Minnesotan participants in the American sample, they claimed that BMI “proves to be, if not fully satisfactory, at least as good as the other proxy measurements” (p. 339, emphasis added).

In 1995, the World Health Organisation (WHO) released a report that advised BMI as best-practice, along with recommended categories. BMI has remained the ‘gold-standard’ for assessing body weight ever since (WHO, 1995). Whilst recommendation from the WHO for BMI in medical practice should be reliable and in good faith, it should be known that the approval process was sponsored by the companies Abbott and Rosche – both of whom manufacture weight loss drugs (Moynihan, 2006). Lowering the BMI obesity classification and in turn increasing the number of people being advised to lose weight indeed acts in their best interests, but potentially not the general public’s.

Inappropriate Use in the Real World
A stark limitation, as noted by the original authors, is a lack of distinction between the weight of muscle and fat. As a result, a study found that when using solely BMI, discrimination between athletes and non-athletes was at a rate of 52.4% (Kruschitz et al., 2013). That is, BMI could not reliably discriminate between increased muscle mass and increased fat. On a broader level, shorter people, by virtue of characteristics out of their control, are more likely to have a higher BMI regardless of their body fat percentage (Bogin & Varela-Silva, 2013). Furthermore, there is no recognition of the different types of fat nor the impact of water weight (Wu et al., 2024). If the individual menstruates, where are they in their cycle? The level of water retention can significantly fluctuate in each stage (Kanellakis et al., 2023). Are they bloated? Have they just eaten lunch? All impact the number on the scale and thus their BMI classification. Yet these variables are rarely considered.

It is of no surprise, given the skewed demographics of the original sample, that BMI continues to perpetuate racial disparities. For example, African Americans are significantly more likely to be misclassified by BMI (Burkhauser & Cawley, 2008). When considering just men, out of a sample of over 8,500, nearly a quarter (24.6%) were misrepresented by their BMI compared to their actual body fat percentage (Romero-Corral et al., 2008). This result was found in both underweight and overweight classification. Since the BMI was constructed for adults, application to children is inappropriate, and it is a poor predictor of their body fat levels (Vanderwall et al., 2017). But this has not stopped real-world usage. Some may remember WiiFit, a video game released in 2007 on the Nintendo Wii console encouraging weight loss through virtual activities. One aspect involved standing on a Wii balance board to have your weight taken, with no consideration or amendments for those under 18. Your avatar changed shape until it matched the BMI value calculated by the console. Classed either as underweight, overweight or obese, sound effects were played to instil a sense of disappointment (think ‘womp-womp-womp’ played on a trumpet). Anecdotally, this visceral reaction has had long-lasting impacts on people’s self-esteem; their beliefs; their relationship with food and their body. Although these shared experiences are prominent on X, Tik-Tok and other social media platforms, this is an area not yet explored with empirical, academic research.

Additionally, the inclusion of BMI in diagnostic criterion can mean health issues are missed. Take eating disorders for instance. Even now, the DSM-V has BMI as a diagnostic criterion. If an individual does not have a low enough BMI, they may not be given a diagnosis of anorexia or bulimia even if they have the cognitive and behavioural hallmarks (Toppino et al., 2022). In turn, people then lack access to support, hindering their recovery. I wish to emphasise that this is not a statistic that impacts an unknown other. I have known the real-world effect of such requirements and the harm it can cause. Given how prevalent (sub-)clinical eating disorders are within the population, there is a chance you do too.

Acknowledging its Usefulness
It is worth noting that, despite its deceptive origins and problematic applications, BMI does indeed have some level of usefulness (Gutin, 2017). It is both time and cost-efficient, especially when more apt technology like CT or DEXA scans are unavailable. Given the limited resources researchers face, especially as undergraduates or early-career researchers, BMI tends to be the best – and only – option. When used as a proxy indicator, its utility for considering complications relating to conditions like cardiovascular health is invaluable. Additionally, for assessing changes over time or correlations of health outcomes (especially within a population format), BMI is beneficial, demonstrating moderate to good specificity for the higher categories (Wells, 2014). Drawbacks none withstanding, when used in conjunction with other measures of obesity and/or health, BMI is more reliable than when in isolation (Chrysant & Chrysant, 2019).

Recommendations for Future Research
Ideally, BMI would not be used at all – why are we categorising individuals into arbitrary classifications using a system rooted in racism, misogyny, and capitalism? However, change does not occur overnight, and the collection of weight classification in body representation research is the ‘norm’. Therefore, if a measurement of weight categorisation is required, alternative methods may be more appropriate, such as the waist-to-hip ratio. Another option may be the Body Adiposity Index, which is similar to BMI but includes the addition of hip circumference (Bergman et al., 2011). This scale has been validated for both sexes, as well as for African- and Mexican-Americans. Further, when it comes to drawing conclusions from BMI data, researchers should be mindful of the way their results could be interpreted by the scientific community, the press, and general public. As BMI was created to be a proxy measure of weight classification, it should be explicitly reported as such.

To conclude, classifying body weight according to BMI categories has the potential to be misleading and detrimental. It is time researchers started considering using other measures. Discussion around the appropriateness of BMI has already begun to gain traction – but it is not enough. I hope this article encourages reflection on the use of BMI in research and its meaning.

Why do we still use the BMI? A call for change

Why do we still use the BMI?
A call for change