-0.16 ROI for Weight Loss Interventions (Optimistic)

post by Ingram Fry · 2019-03-11T09:42:51.181Z · score: 1 (1 votes) · EA · GW · 6 comments

I have calculated an estimated ROI for weight loss interventions of -0.16

For my methodology I first assumed that weight loss programs are effective long term, which is highly questionable.

I then assumed long term weight loss has desirable health benefits, despite evidence that the opposite is true.(Sorenson et al, 2005 ; Lantz et al, 1998).

I then assumed that correlation is causally linked, again highly questionable.

I then assumed that a typical obese person costs similar to an average obese person which is very unlikely to be true.

I further assume no participants drop out which is observably not the case.

Despite all these assumptions the ROI is still negative.

When we consider the opportunity cost, focusing on obesity due to health concerns seems entirely ridiculous.

So even if you believe I am biased and somehow down playing the seriousness of obesity, I would have to under-estimated the CBR of weight loss interventions by 46.4 times in order for the CBR to be the same as focusing on salt, and 75 times to compare to aspirin therapy according to the Copenhagen Consensus.


First of all we have to ask our selves do weight loss interventions work at all. Often they do work initially but in the long run the weight is gained back, often participants end up heavier. I don't know of any randomized control weight loss studies that have long term follow ups.

I do however know of a study that Looked at thousands of overweight metabolically healthy individuals and followed up after 18 years. When they used the
people who stayed the same weight as a reference they found that the people who lost weight were more than twice as likely to die young. (Sorenson et al, 2005)

So even if long term weight loss is possible it may in fact be bad for your health.

But there are many studies that show a correlation between increased weight and higher mortality or medical costs. And there are also other studies that contradict
these studies. (Borhani, 1963; Hodge, 1996)

So the next question is can we reasonably assume that correlation is indicative of causation?

A typical cost benefit analysis for weight loss interventions will straight up assume that correlation and causality are essentially the same thing.

Sometimes they are essentially the same thing. When researchers noticed that HPV was correlated with Cervical cancer they made an HPV vaccine and then
Cervical cancer rates went down. Scientifically proving a causal link for a rare cancer, which sometimes occurs in women who don't have HPV, would be
statistically almost impossible and would take decades. Scientists simply assumed HPV was causing cervical cancer, and given that cervical cancer rates dropped after
the vaccines were introduced it seems they were right. It would have been unreasonable to expect scientific proof of causation before introducing the HPV Vaccine.

So it is always easy to cast doubt by demanding a higher standard of evidence, and this is not always productive. But in the case of weight and health there are
many reasons to believe that weight does not determine health. The only study I am aware of that controlled for socio-economic status found that very heavy people were less likely to die than normal weight people (Lantz et al, 1998) . The epidemiological evidence is all over the place. It is obvious that there are many confounding variables.

But for the sake of this analysis I am going to let this pass and simply assume that correlation is indicative of causation and that lowering peoples weight will
decrease mortality and medical costs.

The first study I found, published in the international journal of obesity, gives a good idea of the cost (Fuller et al, 2013). This study has a long list of conflicts of interest, they are were funded by drug companies and one member works for weight watchers, when the study is evaluating the effectiveness of weightwatchers.
So their incentive is to report that weight loss is beneficial and cost effective. Despite this they estimated the cost of weight loss through WeightWatchers
was $151-$231 american dollars per kg.

So what is the cost to society of a person being obese, again I just looked to the first article I found to answer this question (Yates & Murphy, 2006). This study doesn't declare any conflicts of interest, but the general bias seems to be towards exaggerating the costs while simultaneously exaggerating the benefits of weight loss.

But even if I take the claim that an average obese person cost $554 Australian dollars, or $399 American dollars, per year at face value, it doesn't seem worth it.
$399 us dollars is only going to buy about 2 kg of weight loss which isn't going to do anything.

But is the average cost of an obese person really useful at all?

Anecdotally I heard of a woman who was admitted to hospital, and the nurses had to rub lotion on the ulcerated sores between her fat rolls.
One of the nurses did the math to try and figure out how much it was costing to treat this woman and estimated it at something like $20,000 a day.

This is just an anecdote, but it draws to attention the fact that nobody seems to have published an estimation of the distribution of the increased medical expenses.
Some hospitals are installing MRI machines that are intended for cattle because they have patients that cannot fit into a regular MRI.
This obviously costs a lot of money but has nothing to do with your typical obese person, who would easily fit into an MRI.

I heard of another lady who got turned down for government sponsored weight loss surgery because she wasn't fat enough, so she went and got as fat as she could so she would get accepted. Which she eventually did and then became thin. Weight loss surgery is expensive but not typical.

Lets say that the average citizen cost $1000 a year due to criminal activity. Does that mean I have created a $1000 of value if I stop someone from Jay Walking?
Obviously not because there is a big difference between the average cost of a citizen and the typical cost of a citizen.

The typical cost might be $0 and yet the average is $1000. If one percent of the population is costing $100,000 a year then the rest of the population could be costing $0 and you still get $1000 on average.

The same is likely true of obesity. A typical overweight or obese person in a typical year is likely to cost $0 in additional medical costs.

We have a continuum between two approaches here, one would be a community based approach focusing on a large and diffuse problem, or an "us and them" approach that discriminates against a small minority. We have all been taught that discrimination is bad so there is a tendency to prefer defining problems as though they are communal, even in cases where this is ridiculous.

For example I heard a publicly funded radio program in which they interviewed a criminal who blamed all of society for her many crimes.
The message of the program was about shifting the blame from the few repeat offenders to the entirety of society, which I completely reject.
The problem is that you have a more introspective, reasonable and responsive audience when talking to normal people, so they are easier to deal with.
But the resulting world view is simply inaccurate.

So the WeightWatchers study estimated it cost $151-$231 american dollars per kg of weight loss among typical obese people. But the cost of these typical
obese people is likely to be a lot less than average. There were no extraordinarily fat people who participated in this voluntary experiment.

But for the sake of this video I will just assume that typical and average are the same.

I will run the math on a hypothetical case study to see how things add up in the long run.

Lets imagine we have a woman who is 5"8' and has a BMI of 32 which means she needs to lose about 19 kg in order to have a BMI of exactly 25,
which would put her at the upper limit of what is considered normal weight.

19 times the cost per kg equals 2883-4410 U.S. Dollars. Assuming our participant does not drop out which many did. It takes around 6 years for her to loose enough weight to be at the high end of normal weight. But this is assuming a linear weight loss after the first year, where in fact the weight loss is likely to level off.
So lets say it take 8 years for this weight loss to occur. This puts the cost up to 3844-5880 over 8 years.

This might be worth it if the patient were permanently cured for life, but all longitudinal studies would seem to suggest this is not the case.
Lets say that weight maintenance cost half as much as weight loss. A weight watchers membership plus travel costs on average costs 896 USD per year.
Half of that is 448 USD. whereas staying fat was only going to cost 399 USD per person.

Lets say our participant lives another 40 years from the start of the intervention, 40 years minus the 8 spent in the intervention is 32. 32 times maintenance
costs is $14336. If we add that to the cost of the eight years you get about 18-20 thousand

The cost of remaining obese on the other hand would be $15,960. So there is a net loss of 2-4 thousand dollars for one person.

And this was using the Private care model which was found to be more efficient than standard care provided by the government.


Borhani, N. O. (1963). Report of a ten-year follow-up study of the San Francisco Longshoremen. Journal of Chronic Diseases, 16(12), 1251-1266. doi:10.1016/0021-9681(63)90067-5

Campos, P., Saguy, A., Ernsberger, P., Oliver, E., & Gaesser, G. (2005). The epidemiology of overweight and obesity: public health crisis or moral panic? International Journal of Epidemiology, 35(1), 55-60. doi:10.1093/ije/dyi254

Fuller, NR. Colagiuri, S. Et al. (2013). International Journal of Obesity.

Gaesser, G. A., Tucker, W. J., Jarrett, C. L., & Angadi, S. S. (2015). Fitness versus fatness. Current Sports Medicine Reports, 14(4), 327-332. doi:10.1249/jsr.0000000000000170

Hodge, A. M., Dowse, G. K., Collins, V. R., & Zimmet, P. Z. (1996). Mortality in Micronesian Nauruans and Melanesian and Indian Fijians is not associated with obesity. American Journal of Epidemiology, 143(5), 442-445. Retrieved from https://academic.oup.com/aje

Lantz, P. M., House, J. S., Lepkowski, J. M., Williams, D. R., Mero, R. P., & Chen, J. (1998). Socioeconomic factors, health behaviors, and mortality. JAMA, 279(21), 1703. doi:10.1001/jama.279.21.1703

Sørensen, T. I., Rissanen, A., Korkeila, M., & Kaprio, J. (2005). Intention to lose weight, weight changes, and 18-y mortality in overweight individuals without co-morbidities. PLoS Medicine, 2(6), e171. doi:10.1371/journal.pmed.0020171

Yates, J. & Murphy, C. (2006). A cost benefit analysis of weight management strategies. Asia Pac J Clin Nutr.


Comments sorted by top scores.

comment by Denkenberger · 2019-03-14T05:23:31.296Z · score: 3 (2 votes) · EA · GW

Interesting analysis. You mean a Benefit to Cost Ratio (BCR) of 0.84. Then the Return On Investment (ROI) would be negative.

comment by Ingram Fry · 2019-03-15T03:49:56.230Z · score: 1 (1 votes) · EA · GW

Thanks. I was trained in health not economics. Edited the article accordingly.

comment by aarongertler · 2019-03-12T09:25:18.477Z · score: 3 (2 votes) · EA · GW

Thanks for writing this post, and for leaving a brief summary at the top!

In general, for posts like this that lay out an argument point by point, I'd strongly recommend adding section headers (highlight text in your editor and click the "T" button to create a header). This will give you a cool floating table of contents next to your post and make it easier to navigate (even for someone who reads the whole thing, top to bottom).

comment by PeterMcCluskey · 2019-03-11T19:25:00.385Z · score: 3 (2 votes) · EA · GW

I suspect you underestimate the cost of obesity.

But there's something seriously wrong with the cost of the typical weight loss approach, and your ROI estimate might be close to the right answer for that.

I believe it's possible to adopt a much better than average approach to weight loss, by focusing more on switching to healthier foods (based on the Satiety Index, or on high fiber content), and/or some form of intermittent fasting.

comment by Ingram Fry · 2019-03-11T23:00:53.191Z · score: 1 (1 votes) · EA · GW

Every study I have seen concludes that weight loss interventions are worth while, so I am contradicting these studies. However scrutinising these studies makes one feel that there was a pre-determined conclusion. The cost of weight loss seemed ligitimate in the study I used, (except that they didn’t account for drop outs,) but then they assume that the weight loss is permanent. They also use QALY’s to cost ratio to assess whether their program was worth it. In the book Effective Altruism the estimated loss of QALY’s for obesity is very low. But this is somewhat subjective and open to exaggeration. Other studies claim benefits from a program that costs $10 per person, which I am skeptical of. Often they count loss of productivity as an expense of obesity because obese people have more sick days. However studies show that weight satisfaction is better at predicting sick days than BMI so fat people who are satisfied with the weight don’t take extra sick days. So literally they are getting the cost of worrying about weight and using it as a reason why we should worry about weight.

comment by PeterMcCluskey · 2019-03-17T16:35:42.379Z · score: 1 (1 votes) · EA · GW

I'm a little confused by this reply. Did you think I was complaining that you over-estimated the costs of weight loss? Let me emphasize that I was complaining about the actual resources devoted to weight loss, not your estimates of it. I'll guess that you under-estimated those costs, by focusing on money spent, rather than trying to evaluate the psychological costs.

My main point is that we should focus more on getting people to switch from typical weight loss approaches to ones that are easier and more effective.

I'm unsure what to infer from your weight satisfaction evidence. It might mean that some people notice that obesity is harming them (via sleep apnea? romantic problems?) and that's what causes them to worry. Or it might mean they're just more responsive to peer pressure, and it's the peer pressure, not the obesity, that's harmful.