The maximally pessimistic obesity and weight loss argument
Weight loss is hard, what actually works?
This is another cluster of facts that I end up referring to or commenting on all the time, so I thought I’d write this up somewhere where I can refer to the link or various anchors in the post.1
First, the bad news
Obesity is a one-way ratchet, it is socially contagious, and it is functionally permanent for the overwhelmingly vast majority of people.
Not just that, it gets worse.
Exercise alone won’t make you lose weight, neither will dieting, and the *only* thing that works is complete and permanent lifestyle change, rigorously calorie counting AND exercising, and consistent self discipline for the rest of your life.
Or, you know, you could avoid the whole mess and just pop a pill or do an injection now that the ‘tides exist.
Whew. Well, let’s get into it.
Obesity is socially contagious and a one-way ratchet:
From Christakis and Fowler - The spread of obesity in a large social network over 32 years (2007), n=12k:
If you have a friend who becomes obese, you’re 57% more likely to become obese yourself
If your sibling becomes obese, you’re 40% more likely
If your spouse becomes obese, you’re 37% more likely
Obesity is a one-way ratchet. On average, everyone gets fat no matter what over the long term, regardless of age, sex, race, income, or education:2
Here’s income and education. From Banas et al, United States Long-Term Trends in Adult BMI (1959–2018): Unraveling the Roots of the Obesity Epidemic (2024), n=47k:
More social contagion - it’s not just friends, it’s “ambient peers”
From Alex “Sandy” Pentland, whose book Social Physics I reviewed here, and whose group led the Social Evolution Study (among others). They took ongoing “big data” readings from students in a dorm over 500k hours, and found that for weight change, it “showed a very strong association with exposure to peers who gained weight, but not to those who lost weight,” confirming the one-way ratchet effect.
Worse, they found that “ambient peers” were a significant factor in weight gain, with exposure to those “behavior examples” more powerful than weight gain by friends, gender, age, or stress / happiness combined.
This is *terrible* news (I mean, I did say this was the maximally pessimistic picture, didn’t I?), because basically everyone is fat now. In America, about 75% of the population is overweight or obese, so pretty much everyone’s “ambient peers” are all influencing them fatwards.
Our only hope on this front is that the usual “18-20 year old undergrads” segment isn’t representative of the larger population, or to carefully curate our “ambient peer” environments, which sounds hard.
Obesity is permanent
So, more bad news. Once you gain weight, it never goes away, more or less. A full 98% of “dieters” fail and gain all the weight lost (and often even more) by 5 years after starting. Worse, this “cycling” of weight gain and loss puts you at greater risk of cardiovascular disease. Some doctors challenge this pessimistic view, and point out that if you use a definition of “losing at least 10% of body weight and maintaining this loss for at least 1 year,” it can get up to a whopping 20% of dieters succeeding!3
I’ll let the fact that the optimists are saying“literally 80% of people can’t lose even 10% and keep it off for a year” speak for itself.
People will follow diets for 24 months, and lose on average only 1.8kg. From Madigan et al, Effectiveness of weight management interventions for adults delivered in primary care: systematic review and meta-analysis of randomised controlled trials (2022), n=8k:
The mean difference between the intervention and comparator groups at 12 months was −2.3 kg (95% confidence interval −3.0 to −1.6 kg, I2=88%, P<0.001), favouring the intervention group. At ≥24 months (13 trials, n=5011) the mean difference in weight change was −1.8 kg (−2.8 to −0.8 kg, I2=88%, P<0.001) favouring the intervention.
If you survey the literature, you get such cheering soundbites as:
“Although most women who lost a clinically significant amount of weight regained most of it, they gained less weight over the entire 6 y period than their peers.”4
Or grimmer, in regards to weight loss ameliorating diabetes or metabolic syndrome:
“It is controversial whether lifestyle-induced weight loss (LIWL) intervention provides long-term benefit.”
Most of these studies are RCT’s and/or meta-analyses, which is great from a study quality perspective. But it gets funny - somebody5 decided to look at weight loss in the Control groups, and found that they routinely have roughly half the weight loss as the “intervention” groups, likely due to the Hawthorne Effect:
Among studies including control group in waiting lists and combining standard care, advice and material, no heterogeneity was found (I2 = 0%, p = 0.589) and (I2 = 0%, p = 0.438); and the mean difference was − 0.84 kg (95% CI − 2.47, 0.80) and − 0.65 kg (95% CI − 1.03, − 0.27) respectively.
So not only is diet and exercise weight loss minimal, <2kg over 2 years, about half of it is fake to begin with!
Calorie counting is the One True Way™, with all else folly and waste
So basically, everyone wants to do something simple like “I’ll stop drinking Big Gulps, and I’ll start walking every morning.” Both those are good ideas, but they’re not going to do much for weight loss.
Herman Pontzer has done a lot of metabolic studies on hunter gatherers and westerners, and has found that for any reasonable amount of exercise, the body adapts to the incremental calorie burn.
Pontzer’s thesis and primary research result is that your body compensates for increased physical output over time, nerfing most exercise efforts by 2/3 in terms of excess caloric expenditure (ie if you exercise 360 cal extra a day, your body will eventually get that down to only needing 120 cal incremental daily in food). Worse, in the limit, your body will nerf your incremental efforts 100%, which is determined by looking at 5x active Hadza hunter gatherers vs sedentary moderns, and measuring that both (unexpectedly) have the same Fat Free Mass daily caloric expenditure.
This is good in one sense - that delta for exercisers is used to reduce inflammation and depression and anxiety and heart disease and all the other “diseases of civilization,” and that incremental “lost” 240-360 is directly eating into your ‘diseases of civilization budget’ and preventing them and keeping you healthy. Diseases of Civilization, so called, because hunter gatherers don’t get them. This includes a number of more-or-less surprising things, like Alzheimers, diabetes, anxiety, metabolic syndrome, depression, diaper rash, and much else.
But it’s bad in another sense, because it means that calorie counting is the ONLY thing that can work.
You can see my review of his book Burn here to learn more about this and see if you want to read it yourself.
Want more triangulation points than just Pontzer’s “constrained metabolism” model?
If you look at the Midwest Exercise Trial 1,6 overweight people were assigned to exercise or control groups, and the researchers monitored the exercise, so were sure it was done. The exercise groups worked up to 2,000 kcal of extra exercise per week (the equivalent of running 20 miles), for 16 months. At that level of expenditure, they should have lost 40 pounds. Instead, the men lost 10, and nearly all the weight loss occured in the first 9 months. The women lost nothing, they weighed exactly the same 16 months later while running ~20 miles a week.
In Midwest Trial 2, they upped the ante, and assigned some people to 3,000 kcal a week, or 30 miles of running per week equivalent. Average weight loss remained about 10 pounds, and there was no statistically significant difference between the 2kcal group and the 3kcal group. Worse, a full 46% of the participants had zero weight loss, while running ~30 miles per week.
“In other words, if you start a new exercise program tomorrow and stick to it religiously, you will likely weigh nearly the same in two years as you do right now.”
What actually works?
For the 2-20% of people who actually lose weight and keep it off, it requires drastic lifestyle changes across the board. Not only do you need to count calories rigorously, for the rest of your life, you ALSO need to exercise regularly, to prevent gaining it all back.
The National Weight Control Registry tracks those rare people who actually lose weight and keep it off:
National Weight Control Registry members have lost an average of 33 kg and maintained the loss for more than 5y. To maintain their weight loss, members report engaging in high levels of physical activity (1 h/d), eating a low-calorie, low-fat diet, eating breakfast regularly, self-monitoring weight, and maintaining a consistent eating pattern across weekdays and weekends. Moreover, weight loss maintenance may get easier over time; after individuals have successfully maintained their weight loss for 2–5 y, the chance of longer-term success greatly increases.
Don’t want to calorie count? These guys tried to replicate the protocals in the Midwest Exercise Trials with no changes to diet,7 to see if the high amounts of verified exercise would help keep weight off, and found that:
Weight change across 12 months was 1.1 ± 6.5 kg, 3.2 ± 5.7 kg, and 2.8 ± 6.9 kg in the 150, 225, and 300 min/wk groups, respectively.
So the more exercise, the more weight you gain, if you don’t count calories.
Want to just count calories?
You’ll regain all the weight on “diet only,” per the above graph.
What do you have to do to actually lose weight and keep it off?
Well, this is what we know from the National Registry success cases:
Average 1hr / day of physical activity
Eat a low calorie, low fat diet - so you are counting both calories and macros
Eat breakfast
Self-monitor weight regularly
Maintain a consistent eating pattern across weekdays and weekends
If you’d like to learn more about exercise in terms of the amounts and types you should do and the benefits from them, including recommendations from Dan Lieberman, the Harvard authority on exercise, and the National College of Sports Medicine, read my review of Dan Lieberman’s Exercised, here, or the book itself.
It’s a tough road, but it’s the ONLY road we know about...
Or is it the only road to weight loss?
It depends how much you want to lose, but if it’s 15% or under, you’re in luck! The heavens have parted, hosannahs have been sung, and lo, the ‘tides8 have smiled upon humanity and granted pill-or-injection based weight loss to all!
They primarily work via mediating appetite and fulfillment. They have a ton of other benefits too - less addictive behaviors overall, less drinking, you might age better, and who knows what else. Read Scott Alexander’s post Why Does Ozempic Cure All Disease? on it for more.
But if you look up there at the lengths you’d have to go to WITHOUT the ‘tides, I think you can see that they’re a really big deal, and are probably the best first-line approach for anyone interested in weight loss.
Bariatric Surgery is another option
As seen above in the Genuine Impact chart, it’s the other success case that works for ordinary people, and it’s pretty much the only way to lose large amounts of weight if you’re not in the 2%.
The Longitudinal Assessment of Bariatric Surgery study9 was able to keep track of 83% of a 1500 person sample who had gastric bypass for 7 years, and found that 7 years later, they had maintained a mean weight loss of 38kg (83.6 lbs), or around 28% of body weight.
For gastric banding, 7 years after surgery mean weight loss was 18kg, or 15% of body weight. Both had lower dyslipidemia at the 7 year mark, but only gastric bypass patients had lower rates of diabetes and hypertension.
However, bariatric surgeries have a 17% complication rate, a 7% “needs resurgery” rate, and incremental mortality rates of ~.31% after thirty days. This is honestly probably fine. That’s 3.1 out of 1,000, and there’s 250k bariatric surgeries per year, so around 750 annual incremental deaths. The average one-year mortality rate for a 45 year old with a BMI of 45 is ~.9%, so not a ton of incremental risk from the surgery, and there’s some evidence that bariatric surgeries have beneficial effects on all cause mortality - the gastric bypass lowers diabetes rates, hazard ratios for cardiac mortality and myocardial infarction are 0.48—0.53 post surgery,10 and their 10 year cancer mortality rates are only 0.8% vs 1.4% in controls matched to characteristics who did not receive a surgery.11
However, a recent 2024 meta-analysis found high evidence of bias in previous studies, and concluded that once you adjust for bias, all cause mortality hazard ratios after bariatric surgery are more like .92 versus the reported .78,12 so there may be weaker effects than originally thought.
Still, if the ‘tides don’t work for you, and you’re pretty sure you’re not in the 2% of people with the willpower for “total and permanent lifestyle change,” bariatric surgery may be something worth considering.
Appendix: don’t let your kids get fat
More than 22% of kids today are *obese,* and a much bigger chunk are overweight. Knowing that this is permanent and one-way, this is a big problem - especially because kids can’t do ‘tides.
From M Simmonds, et al - Predicting adult obesity from childhood obesity: a systematic review and meta-analysis (2015)
Obese children and adolescents were around five times more likely to be obese in adulthood than those who were not obese. Around 55% of obese children go on to be obese in adolescence, around 80% of obese adolescents will still be obese in adulthood
Letting your kids get fat:
Puts them at significantly higher risk of depression and anxiety (OR 2.24-2.36)13
Puts them at significantly higher risk of metabolic syndrome (OR 3.4-5.8)14
Increases risk of having Type 2 Diabetes (OR 1.45-1.83)15
Puts them at 5x greater risk of adult obesity, and thus, cardiovascular disease, morbidity, and death when older (4M deaths due to obesity in 2015, and 70% of those due to CVD, and obesity rates are 37% higher today vs 2015)16
Obesity prevention needs to start young, and the younger the better, because once somebody is on a “high and gaining” BMI trajectory, they’re basically guaranteed to be fat.
From Buscot, et al (2017):
“Efforts to alter BMI trajectories for adult obesity should ideally commence before age 6 years. The natural resolution of high BMI starts in adolescence for males and early adulthood for females, suggesting a critical window for secondary prevention.”
And unsurprisingly, because weight gain is functionally permanent, any interventions are quite weak and ineffective, ranging from one twentieth to less than half of a stdev:
“On the one hand, therapeutic interventions for childhood obesity have shown only very limited success to date, with achievable weight loss of 0.05–0.42 BMD-SDS units (standard deviation of the body mass index) over 12–24 months”17
Every parent seems to be in a furious Red Queen’s Race to get their kids into the very best pre-schools and schools so their precious bundles can get into Harvard - did you know that if you simply got them to do at least 1hr of physical activity a day, just that single intervention alone would put them in the top quartile of all kids in the US, and that physical activity can “improve cognition, executive function, attention and academic performance?”
And what do you do to prevent your kids from getting fat?
If you’re serious, no junk or fast food, less screen time, and making them move.
“At the most basic level, childhood OW/OB emerges from consuming more calories than expended, resulting in excess weight gain and an excess body fat. Caloric imbalance is the result of, and can be further exacerbated by, a range of obesogenic behaviors. That is, behaviors that are highly correlated with excess weight gain. The most common obesogenic behaviors are high consumption of sugar sweetened beverages and low-nutrient, high saturated fat foods, low levels of physical activity and high levels of sedentary behaviors, and shortened sleep duration (e.g., Sisson et al 2016).”
The interventions:
Make sure your kids get ~1hr of physical activity EVERY DAY18
“Physically active children have lower rates of obesity in childhood, and a lower risk of obesity and chronic medical conditions in adulthood. In addition, higher levels of physical activity improve cognition, executive function, attention and academic performance.”
Make and serve real food, avoid buying fast food because it’s cheap and easy
Don’t let them eat candy, junk food, fast food, or soda19
Limit screen time20
Practice good sleep hygiene - time box the kids’ internet router or IP’s, have them turn their phones into you at night, keep their rooms cool and dark at night, and similar interventions
Do you want my “easy mode” recommendation for the “movement and screen time” things above?
GET A TREADMILL DESK
Yes, kids want to waste their lives staring at screens, so make a deal with them - they can spend X amount of time on whatever screens they want, as long as they’re on a treadmill desk, slowly moving, as ~2M years of hominin evolution intended. Video games? Sure. Streaming stuff? Why not? Tik Tok? I mean, that’s up to you. But as long as they’re moving, I think you can BOTH be happier, and can make sure they’re moving and getting the physical and mental benefits of moving every day.
Also, get a treadmill desk for yourself! Exercise is hard because adherence is hard - but do you know what’s easy? Slowly walking on a treadmill, in your own house, wearing whatever you want, while YOU’RE getting screen time, whether working or recreational.
The treadmill desks I’ve bought are UNDOUBTEDLY the single highest “unit of value in life per dollar spent” things I’ve ever owned in my entire life.
And if you’re like me and are always thinking “eh, I can do a smidge more than last time, why not?” and hit a single up-button on either speed or elevation, over time it can actually burn significant calories too.
I just found out recently I’d *inadvertently* been burning an extra 700-800 calories per day, while walking at an 8-10% incline for a few more hours. I only found out because I was hungry all the time and looking skinnier after about a week of it, so I wore my Polar heart rate monitor for a day to see where the energy drain was (you can’t trust “machine calories,” they’re all lies, but you can trust heart rate monitor calories if it has your age and weight). If you too would like to be able to accidentally burn an extra 800 calories a day, I highly recommend treadmill desks.
Substack has anchors! I just found this out. Any Heading generates an anchor, but you can only access the anchor link after it has been published, by hovering over the heading section in question.
Todd and Myrskylä - Projection of US adult obesity trends based on individual BMI trajectories (2024) - DOI: 10.4054/DemRes.2024.51.13, n=26k
Wing, Phelan Long Term Weight Loss Maintenance (2005), DOI: 10.1093/ajcn/82.1.222S
AE Field et al, Relationship of a large weight loss to long-term weight change among young and middle-aged US women (2001) doi: 10.1038/sj.ijo.0801643., n=47k
Hajjaj et all, A systematic review and meta-analysis of weight loss in control group participants of lifestyle randomized trials (2022),
METS 1: Donelly et al (2003) DOI: 10.1001/archinte.163.11.1343
METS 2: Donnely et al (2012) DOI: 10.1016/j.cct.2012.03.016
Washburn et al, A Randomized Trial Evaluating Exercise for the Prevention of Weight Regain (2021) DOI: 10.1002/oby.23022
On diet, technically, they provided information and “recommended continued consumption of a minimum of two portion-controlled entrées, three low-calorie shakes, and five servings of fruits and vegetables per day” these things “were encouraged, but not required, and portion-controlled entrées and shakes were no longer provided by the trial.” They also provided a list of suggested things at local stores: “Participants were asked to purchase portion-controlled entrées and shakes with acceptable energy and macronutrient content available at local supermarkets from a list provided by the trial.”
Initially they had provided meals in the 3 month “weight loss” period before the weight loss maintenance + exercise portion of their study.
So this wasn’t even “ad libitum” food intake, it was “heavily advised to be lower calorie and healthier,” as well as providing lists, and the participants still gained several kilos each.
Semaglutide, Liraglutide, Tirzepatide, Exenatide. A compounding pharmacy is the way to go, because it’s about 10x cheaper per month, and they don’t have supply shortages.
Courcoulas et al, (2018), DOI: 10.1001/jamasurg.2017.5025
Chandrakumar et al, The Effects of Bariatric Surgery on Cardiovascular Outcomes and Cardiovascular Mortality: A Systematic Review and Meta-Analysis (2023), DOI: 10.7759/cureus.34723
Aminian et al, Association of Bariatric Surgery With Cancer Risk and Mortality in Adults With Obesity (2022), DOI: 10.1001/jama.2022.9009
Matching of controls was done by KNN-ing to the closest 5 patients in the control who matched on a propensity score calculated from age, sex, race, BMI band, smoking, diabetes, Elixhauser comorbidity, Charlson comorbidity, and state.
Suissa et al, Bariatric surgery and all-cause mortality: A methodological review of studies using a non-surgical comparator (2024), DOI: 10.1111/dom.15771
n=2717, https://doi.org/10.1038/s41366-023-01312-6
n=554, DOI: 10.3389/fendo.2022.915394
doi.org/10.1007/s13679-018-0320-0
The American College of Sports Medicine recommends at least 60 minutes of activity for children every day - link.
Some recommended activities:
Tag or follow the leader
Playgrounds
Bicycles or tricycles
Walking, running, skipping, dancing, jumping
Swimming
Sports
Gymnastics or tumbling
This is good for their brains, too:
“This systematic review synthesizes the high-quality experimental evidence available regarding the effectiveness of physical activity on motor skills and cognitive development in 4–6-year old, typically developing children. Findings favor causal evidence of relations between physical activity with both motor skills and cognitive development in preschool children, with increased physical activity having significant beneficial effects on motor skills and cognitive functioning.” - doi: 10.1155/2017/2760716
“Dietary factors contributing to obesity risk in children and adolescents include excessive consumption of energy-dense, micronutrient-poor foods; a high intake of sugar-sweetened beverages; and the ubiquitous marketing of these and fast foods.39, 40 The relative effect of other factors such as specific eating patterns (eg, frequent snacking, skipping breakfast, not eating together as a family, the window of time from first to last daily meal), portion sizes, the speed of eating, macronutrient intake, and glycaemic load on obesity development remain unclear, although all might be important.41, 42”
“The link between screen time and obesity in childhood and adolescence was initially documented through cross-sectional and longitudinal studies of television viewing.43, 44 The past two decades have seen the increase of mobile and gaming devices. Screen exposure influences risk of obesity in children and adolescents via increased exposure to food marketing, increased mindless eating while watching screens, displacement of time spent in more physical activities, reinforcement of sedentary behaviours, and reduced sleep time.44, 45”
Great article! What desk treadmill do you recommend?