Follow me on

Why Treating Severe Obesity Has The Highest Potential For Health Cost Savings

sharma-obesity-dollarsAs the latest HQCA report on obesity in Alberta released this week, the substantial population burden of overweight and obesity (now affecting 6 in 10 Albertans) is a significant driver of health care costs in the province. In the US, this increased health care cost for adult obesity is estimated at around $3,508 per individual with a BMI greater than 30 for a total of well over US$ 300 billion per year.

However, as highlighted in a recent article by John Cawley and colleagues in PharmacoEconomics, health care costs are not equally distributed across all people living with increased body weight – rather, obesity related health care costs rise exponentially with increasing BMI levels (i.e. at the extremes of BMI).

Thus, the greatest health care savings for individual patients can be expected in those living with severe obesity.

To illustrate this, the researchers used data from the US Medical Expenditure Panel Survey from 2000-2010 (n=41,435), to calculate the potential annual savings in health care costs (in US $ in the US health care system), for various reductions in body weight in individuals with BMI levels ranging from 30 kg/m2 to 45 kg/m2.

Thus, for e.g. the annual cost savings with a 5% reduction in body weight for someone with a BMI of 30 kg/m2 amounted to a mere $69 per year.

This figure, however, increased exponentially for people with higher BMIs, increasing to $528, $2,137, and $10,030 in an individual with a BMI of 35, 40, and 45 kg/m2, respectively (these figures were somewhat higher, when the individual also has diabetes).

Thus, while treating obesity to achieve a 5% reduction in body weight in someone with a BMI of 30 kg/m2 may never be “cost-effective”, the same amount of weight loss in someone with more extreme obesity, would likely pay for itself or even lead to significant savings.

Because the impact of obesity on mental and physical health, life-expectancy and quality of life is also greatest at higher levels of BMI, one could also make a strong ethical argument for singling out these individuals for priority treatment in the health care system.

Obviously, as readers should be aware, BMI is at best a crude measure for health – a more precise assessment would have used more sophisticated staging systems like the Edmonton Obesity Staging System to calculate individual risk and benefits. However, we should remember that at a population level BMI does function moderately well as an indicator of obesity related risk (although not in individual patients).

This analysis has important consequences both for population and individuals approaches to obesity.

Although the population burden of obesity lies in the middle of the BMI bell curve, and shifting this ever so slightly can move a substantial number of people living with overweight or obesity to a BMI that lies below the current cut-offs, such a change may have little influence on the overall health care costs of obesity, as these live in the extremes.

Thus, using the above numbers in a crude back-of-the-envelope calculation, to save $1,000,0000 per year in health care costs, one would have to lower BMI by 5% in about 14,500 people living with a BMI of 30 kg/m2 compared to only 100 people with a BMI of 45 kg/m2 – a much more manageable problem.

This is why it is harder to make a cost-savings argument for addressing obesity at a population level rather than focussing on those living with more severe obesity, unless such population measures can also substantially help lower the BMI of the latter.

Unfortunately, current population trends show that while rates of overweight and mild obesity appear to be levelling off (thank perhaps in part to population health measures), severe obesity continues to increase at alarming rates.

This is why a greater focus on finding and delivering better treatments to those living with severe obesity, including those that can only offer modest reductions in BMI, has to be the main priority of any health care system seeking to reduce obesity related health care costs.

Edmonton, AB


  1. Dr. S: I think there is a phrase you casually throw around here that you may not even believe. It’s a common phrase that people (even in white coats) say it without thinking, like peanut butter and jelly, macaroni and cheese. In sentence 1 you talk about, “the substantial population burden of overweight and obesity.”

    Is there really a substantial burden specifically attributable to overweight? Even you argue a couple of posts ago that there is “controversy” over mortality and BMI (and I, of course, chimed in in support of Flegal et. al. who places the lowest mortality at a BMI of 26.7.) While there are increased morbidities associated with (not caused by) overweight, is it right to lump those two together? If the “obesity paradox” and a decreased mortality rate for overweight people is a high probability, does it make sense to say that overweight presents an increased burden? (I don’t know about you, but I’d rather have a 27.6 BMI and be ALIVE but with a comorbidity, say high blood pressure, than have a BMI of 18.5 in my casket.) It’s common practice to link “overweight and obesity” in terms of “burden,” but is it good medicine? I think there are special interests (popular diet programs, women’s magazines etc.) who LOVE it when respected doctors do link those two, but is it a disservice to us common folk? Overweight is not obesity in as much as obesity is not smoking. My humble opinion.

    Post a Reply
    • The word “burden” (as in “population burden”) is a technical term commonly used by epidemiologists to describe the magnitude of a given issue in a population. This does not actually mean that this issue is actually a “burden” (as used in normal parlance). I understand that the term can give cause to irritation.

      Post a Reply
  2. I actually believe that going to the root of the problem, which is not obesity but trauma will give much higher savings…

    Post a Reply

Submit a Comment

Your email address will not be published. Required fields are marked *