Paul Alper points us to this news article which begins:
For every 2,000 steps you take each day, your risk for premature death may fall by 8 to 11 percent, according to research published in the journal JAMA Internal Medicine.
Along with the results from a related study, published in JAMA Neurology, the researchers also found that walking more, accumulating up to roughly 10,000 steps a day, was linked to a reduction in the occurrence of cardiovascular disease (including heart disease, stroke and heart failure), 13 types of cancer and dementia. . . .
Taking 10,000 steps a day (roughly four to five miles, depending on a person’s stride) has become a common health and fitness goal. . . . The new studies, however, found that health benefits also can be achieved by taking fewer steps. For instance, walking about 9,800 steps a day was found to lower risk of dementia by about 50 percent, but dementia risk was cut by 25 percent for those who walked as few as 3,800 steps daily.
What struck me was the way the article swung back and forth between cautious disclaimers and unrealistic precision. On one hand, your risk “may fall”; on the other, the drop in the risk is from “8 to 11 percent.” From a logical standpoint, this makes little to no sense: if you can’t even be confident about giving a causal interpretation to the result (regression coefficients from a large observational study), then how can it make sense to be so precise? The use of the phrase “may” in that sentence implies the effect could be zero (or, for that matter, negative, although I guess a negative result would go against our general theoretical understanding of the benefits of exercise). If the effect could be zero or it could be 10%, then I think it could be 5% too. (See Section 3 of this article for a discussion of this general point in the context of a different example.)
Reading further into the article, there is more of this oscillation between disclaimers and over-precision. For example, the second paragraph quoted above carefully uses the non-causal phrase “was linked to,” but then the third paragraph switches to the causal phrases “can be achieved” and “was cut,” and also there are the ridiculously precise numbers of “9,800 steps a day” and “as few as 3,800 steps daily.” I’m bummed cos I only took 3,700 steps yesterday. I better make sure to take 3,900 tomorrow to catch up.
Later on, we get the unqualified causal statement, “walking at a faster pace, or upping the intensity by power walking, for example, was found to have health benefits, too, with intensity amplifying the results,” immediately followed by the association language of “was linked to.”
My point here is not to slam this particular news article but rather to point out that writing about observational studies is hard! Writing about science is hard! On one hand, you want to keep putting in the disclaimers—association not causation—; on the other hand, if you’re reporting the study at all, the numbers should be relevant. I guess the right way to put it would be to say something like, “Walking 2,000 steps a day was associated with a 10% decline in risk of premature death,” but it’s hard not to slip into causal language.
To their credit, the authors of the journal article were more careful, using the word “Associations” in the title and abstract, and summarizing result as “more steps per day (up to about 10 000 steps) was associated with declines in mortality risks and decreased cancer and CVD incidence.”
This all reminds me of a class in college where we learned about the religious doctrine of predestination. Somehow it sounded weird to us, the idea that your success in life would reveal your predetermined destination. Getting saved by good works seemed more intuitive. But there’s be no empirical way to distinguish. In that case the difficulty is not from having observational data but the more fundamental problem of the outcome being unobservable.
Obviously, cause and affect is suspicious–healthy people tend to do the walking in the first place, rather than walking in of itself promotes health (even though I personally believe it does).
There are other aspects of this news article which are weird. The data (set) comes from the mysterious UK Biobank, https://www.ukbiobank.ac.uk/
“UK Biobank is a large-scale biomedical database and research resource, containing in-depth genetic and health information from half a million UK participants.”
The UK Biobank has appeared several times in Andrew’s blog regarding coffee drinking and mortality as well as exercise related to diet and mortality. The weirdness I referred to is that despite being a UK data bank, the authors (of whom there seems to be many) never appear to be from the UK. Besides, the data (sets) appear to be entirely self reported.
“The weirdness I referred to is that despite being a UK data bank, the authors (of whom there seems to be many) never appear to be from the UK.”
UK Biobank data is available to anyone, subject to a data use agreement, and a (hefty) fee.
“the data (sets) appear to be entirely self reported”
Yes. The genetic data is obtained from laboratories, as are some lab studies, but data such as exercise habits, diet, etc. are all self-reported (or extracted from medical records, which is just a second-hand self-report). That is how large observational data sets are typically built. Actually monitoring physical activity data and diet are not feasible at scale.
If you follow research in genomics, UK Biobank is the least mysterious dataset in the world. There are like a million papers that use it. The main reasons for its popularity are that it has a very large sample of genotyped individuals, it is openly accessible to all legitimate researchers, and it has a huge variety of phenotypic data. Most people using it aren’t from the UK because most researchers aren’t British. Its data are mostly self-reported, reflecting a tradeoff between statistical power and the quality of phenotyping.
“writing about observational studies is hard!”
I disagree. It’s not hard. Research summary: “Study finds people who exercise live longer.” See? When you tell the truth it’s not even news. Even the readers knew it was a waste of breath to generate this nonsense. One commenter pans the article like this:
“I want a trillion dollars to study whether or not eating right and fitness work. Not that I found that out in gym class.”
Not hard.
Chipmunk –
> I disagree. It’s not hard. Research summary: “Study finds people who exercise live longer.” See?
You’re avoiding the issue. Andrew was saying that writing about the findings of an individual observational study (even though he used the plural) is hard. Not that it’s (necessarily) hard to make. general statement about them.
> When you tell the truth it’s not even news. Even the readers knew it was a waste of breath to generate this nonsense.
Sigh. I guess no one should bother to study (and then write about) the effect of pretty much any type of intervention: Just ask Chipmunk to generalize about the related phenomena and move one – we’re all set.
“You’re avoiding the issue. Andrew was saying that writing about the findings of an individual observational study (even though he used the plural) is hard.”
OK, here you go: He’s wrong.
It’s not hard. Its cake. The writer probably even has a degree! Journalists are important people saving democracy, so they should have basic competence. Most of them can read, count, and understand the words like “could” and “may”. She’s a highly qualified journalist, employed by one of the country’s top newspapers. For people of this caliber, this should be simple stuff. Not hard.
” I guess no one should bother to study (and then write about) the effect of pretty much any type of intervention”
You’re avoiding the issue, which is that this is a useless study, which you know well enough.
No, I don’t think it’s useless – unless you try to get more out of it than it can legitimately give.
The study is information. It fits along with other information, like some of which is likely conflicting. In the end, over time, you put discrete pieces together into some kind of a matrix or taxonomy and then look for patterns and assess probabilities.
Saying it’s useless is basically a symptom of the same underlying problem as saying it’s dispositive of something: you’re looking at it in isolation and reading too much into it, trying too get more out of it than what it can give.
Ok chipmunk, is it better to walk 10,000 steps, power walk 8000, jog 5000, or sprint 3000 in high intensity intervals?
You neglected to mention the especially galling end note: “This article is part of The Post’s “Big Number” series, which takes a brief look at the statistical aspect of health issues.” Apparently this is what an article that *tries* to be numerate looks like!
I wonder how much of this is science journalists thinking in terms of style and heuristics (what words are more appropriate, what to avoid) rather than having been educated about what causes scientific uncertainty. Taking uncertainty seriously every step of the way while still generating excitement about a finding or seeming to have a point can be hard=. But maybe if this were made a priority in science journalism education it could be taught.
It is not possible to do a controlled trial of longevity. One would need thousands of subjects and decades of studies. Observational reports are the best we can get. Over the years we have been told that coffee is bad, no it’s good, no it’s bad, no it’s good. The observation that walkers live longer is probably valid, but the causality may flow in the other direction. Additionally, how many steps per day did today’s 87 year old take at age 57? And even if valid, how do you get people to get up and walk? Humility in health claims don’t get much publicity; a health adviser who measured his words carefully will not be a major party candidate for senator.
+1
The 10,000 steps a day bit is old news, dating back to at least 2003. Roger Ebert touted it on his blog. I forget where it came from originally, if I ever knew. In my anecdotal experience it is good advice. I quit a job at age 57 due to poor and worsening health, and walked my way back to health in a year. You could say I literally walked my butt off.
I got another job then and continued walking, to and from work, a half-hour of the noon hour, to shop, and a 15-minute walk break in the morning and afternoon at work.
I got some of my best ideas during those walk breaks. Somehow being away from a computer or even pencil and paper made me remember or develop methods that didn’t occur to me at my desk.
The 10,000 steps things is _Asian_. Chinese and Japanese big numbers are written as groupsv of powers of 10^4, as oposed to the powers of 10^3 we use in the West*.
So there was a pedomiter in Japan years ago called a MAN-PO-KEI, where man is 10,000, po is step, and kei is measurement device. It was a big hit, and manpokei became the Japanese word for kleenex, oops, pedometer. (Manpokei was registered as a trade name in 1984 in Japan by the Yamasa Watch and Instrument Company, Inc. and they’re still at it: https://www.yamasa-tokei.co.jp/)
To beat a dead horse to death; the 10,000 number was selected for linguistic convenience and has no scientific basis whatsoever.
*: Thus one of the hardest things in Japanese to English translation (in technical and economics things) is getting the numbers correct. It’s a major pain.
agree on everything here, but also still think that 10k steps per day is a pretty reasonable approximately optimal quantity. I think the optimum is probably pretty broad, if you’re getting between 5000 and 20000 you’re doing well. Less is probably not as good, more might cause overuse injuries to knees or such if not careful… When I got a dog I automatically wound up in that band… 5000 is a light day, 20000 is a very heavy day, 6000 to 12000 is common. It’s definitely changed my life and improved my health.
Yes. Environmental circumstances are big determinants of exercise. When I lived in NYC and had a fairly long walk to the subway both from home and from work, it was a rare day that I got less than 12,000 steps, even though I never specifically set aside any time for walking. Often more: every trip to buy something included a walk of at least a few blocks each way. But since I moved away and began working from home, a home where most stores are driving distance, I have to make a concerted effort to get 8,000 a day, and for that I have to carve an hour out of my day just for the purpose. Which means that when I’m really busy, I don’t even get that.
Medical research has only recently begun to look at this kind of issue.
You can also conclude that 10K steps (~5mi) is a rough approximation of the necessary amount just from the caloric burn. Since you have to burn about 3500cal to lose a pound, if you reduce intake calories to basic necessity (e.g., what your body burns without additional exercise) and your 40lbs overweight, it would take about four years to reach normal weight by walking a mile a day, but less than a year to get there walking 5mi a day. Given the difficulty of reducing intake calories to the bare min – and that even just a cup of something like instant apple cider is nearly 100 cals – you might never get there, and might never lose weight walking a mile or two a day, but at five miles a day, even with a few failures on the calorie intake, you’d probably loose weight steadily.
“But since I moved away…I have to carve an hour out of my day just for the purpose.”
But you got back the time that you used to spend walking to the subway. And now you can get even more time back by doing an upper body workout while you walk or using a rowing machine or cross trainer to increase the intensity of almost every aspect of your exercise, reducing the amount of time you need to spend doing it and becoming even more fit in the process. 12K steps is about six miles or two hours and 600 cals. I can do about 900 cals on a stationary bike in an hour even without upper body movement. I’ve never done a full hour on a rowing machine….
Chipmunk –
Surely you know that the energy balance model has limited utility, body weight is not just math. Much depends on variations at the individual level like age, body type, characteristics of the digestive and metabolic systems, etc.
I’m not keen on what Daniel says. I have been trying to walk more (although I don’t count the steps), and the main thing I notice is that I’m losing weight — which can be good for someone who is overweight, but is not a good thing for a dinky person like me.
“also still think that 10k steps per day is a pretty reasonable approximately optimal quantity.”
Yes, it was a good guess. And there have been studies! One reported that 7,500 was a point above which diminishing returns hit, another found that 3,500 or so was a reasonable target for significant health benefits. (Numbers from memory.)
I’d guess that individual differences are also large: I read about folks who run and and/or gym in the morning and then work a well-over-40-hour-a-week job. And don’t believe it.
The experience here was that getting to 10,000 was really hard and left me tired for the rest of the day. Ditto for a stationary bike. I pushed my pulse rate up to recommended levels, gradually increased the time at that rate. And was tired for the rest of the day.
Now I do squats and push-ups during breaks. And juggling (working on 4 balls and club back crosses). Moving the arms has to help, I claim. Also, not having a car means I do some amount of walking in normal day-to-day things.
Martha: perhaps a consult with your doctor and a personal trainer could help. Maybe give you some strength building exercises, and a diet modification for muscle weight gain? what works for 30-40 year olds is definitely not going to be the same for people 70+ so probably take everything with a grain of … well maybe not salt … I don’t know, maybe curry powder :-)
Joshua said: “body weight is not just math”
Huh? Is there an organ in humans that creates or destroys matter?
It’s math alright. What goes in must stay in, be burned as fuel, or come out as waste or through evaporative loss. Maybe there is another option I haven’t covered?
The percent of calories you purge as waste can vary quite a lot between people. In the extreme case, those with a hyperactive thyroid (graves’ disease) will eat and eat and eat and eat and the calories will go in one end and out the other as they waste away. There’s also a lot of difficulty in measuring calorie burn. A significant majority of calories are burned just maintaining homeostasis, and this “ambient” calorie burn varies drastically with temperature. The body is not magic; it has to maintain body temperature, and the heat energy must come from somewhere.
As such, it’s very difficult to measure the number of calories that are actually absorbed and expended by the body. The gold standard is to put a human in a sealed environment and measure the temperature very precisely for some length of time
https://www.sciencedirect.com/topics/medicine-and-dentistry/direct-calorimetry
Most of the time, they use a less expensive method like bomb calorimetry which basically lights food on fire in a fixed volume container and measures the heat output. If it sounds like that would only loosely correlate with the energy a human body would absorb that’s because it does.
So yeah, nutrition is just arithmetic, but arithmetic where you don’t know any of the numbers. Counting calories and exercise can help pin you down to a stable point, but you won’t know *where* it is in terms of net calories. You might calculate based on food labels that you’re at x calories in and based on bodyweight and exercise that you’re at y calories out, but you must observe the results yourself over a time period and adjust accordingly, you can’t just trust x – y.
Somebody:
“Nutrition is just arithmetic, but arithmetic where you don’t know any of the numbers” . . . well put!
Chipmunk –
I didn’t say it doesn’t involve math. I said it’s not JUST math.
somebody elaborated upstairs, but the energy balance model is simplistic to the point of being generally useful but not very instructive.
For example, there have been times when I increased my activity and consumed fewer calories and lost weight. And here have been times where I increased my activity and consumed fewer calories and didn’t lose weight. That’s (presumably) because my metabolism adjusted due to other factors, or there were additional variables involved.
Correction:
It looks like more precisely, the energy content of food is based on the Atwater system, which uses standard energy densities per gram of protein, fat, carbohydrate. These energy densities are calculated by combining heat of combustion measurements as described with “digestibility coefficients”, the origin of which I’m not sure of. In any case, these can still pretty inaccurate; this study finds a 32% discrepancy with almonds
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3396444/#abstract-1title
So it seems those numbers really cannot be taken literally
Chipmunk –
Let me put it this way. Select any two people and give them the exact same amounts of the same food and have them engage in the exact same amounts of the same activities.
Do you think their subsequent weight gain or loss will be exactly the same?
That’s what I meant by its not “just” math.
Looks like they set the food on fire and then also set the poo poo on fire, based on equation 1 in
https://academic.oup.com/ajcn/article/86/6/1649/5064893#E1
Somebody said:
“So yeah, nutrition is just arithmetic, but arithmetic where you don’t know any of the numbers.”
It doesn’t matter that you don’t know the numbers. What matters is that you know there is an equation that must be resolved exactly, that there are terms you can control, and that there are limits to the terms you can’t control.
That allows **very useful** approximations – as I indicated in my first comment to Daniel. The fact is that you can control your calorie intake and burn well enough, and both can be done with little enough effort, that *most* people should be able to succeed without very precise numbers. Do 400-700 cals or more per day exercise and keep the carbs – esp wheat, rice, sugar, corn products – in check and your sure to lose weight unless you have some disease that prevents it.
—-
Joshua said:
“Do you think their subsequent weight gain or loss will be exactly the same?”
If they are clones of the exact same weight and fitness, yes. If they are not clones, not of the same gender, weight, height, or fitness, no. However, if both are of the same gender, height and weight, and both reduced consumption of useless calories from things like corn, bread, rice, and sugar by the same amount; and they both initiate a daily exercise routine of over say 500 cals, I’m confident both would lose weight.
Ok, I managed to find it. It looks like based on the FDA guidelines here
https://www.accessdata.fda.gov/scripts/cdrh/cfdocs/cfcfr/CFRSearch.cfm?fr=101.9&SearchTerm=nutrition%20label
you can use bomb calorimetry directly, or use precomputed factors per carbohydrate, fat, protein. Those factors come from table 13 here
https://www.ars.usda.gov/ARSUserFiles/80400525/Data/Classics/ah74.pdf
and are also based on bomb calorimetry multiplied by a “digestibility factors” like
(amount in food – amount in feces) / amount in food
In any case, these numbers can be extremely off.
@Chipmunk
Yes, I understand that it is possible to lose weight by increasing activity and decreasing calorie intake. I’m just responding to your contention that it’s “just math”. The fact that a mathematical expression exists is irrelevant; a mathematical expression exists for everything so long as you believe the universe is governed by consistent physical laws. In practice, attempting to actually do the math is pretty unhelpful; actually doing it is a lot more complicated than that.
A more useful model for calories is perhaps a compartment model… food enters the digestive system. Some variable fraction of it is absorbed into the bloodstream, some fraction is excreted as stool. Once in the bloodstream, some variable fraction will be metabolized to produce heat for homeostasis, some fraction will be used to power organs like muscles, the brain, etc, and some fraction will be stored as fat. In addition, some amount of sugar might be excreted in the urine esp. for diabetics, and of course any alcohol you consume will be metabolized away or excreted in the urine.
Each of the “some fractions” is both time-varying, dependent on things like hormones, external temperature, and of course activity level. As you become more fit you will burn fewer calories doing equivalent work, because you become more efficient. As you gain muscle mass you burn more calories doing maintenance of that mass. As you put on subcutaneous adipose tissue you will spend fewer calories maintaining body temperature in cool weather, but will suffer more heat related issues…
I agree with the assessment of “it’s just math but math where you don’t know any of the numbers”. If you follow things in detail for many weeks you might be able to fit some kind of Bayesian ODE or difference equation model successfully.
Somebody, Daniel:
“The fact that a mathematical expression exists is irrelevant; ”
I disagree absolutely. Many people believe in magical ideas about weight gain and loss. It’s not magic. Despite variation in internal processes, most people’s weight is surely controlled by their ratio of calorie intake to calories burned in exercise. In a way, you are propagating magical ideas with your claim that it’s all just too much to understand and there’s no way to quantify it, when in fact there is a way to quantify it – by making reasonable estimates about the parameters in the equations and operating beyond them. We don’t have to understand every term or know it’s value to succeed.
Your position here is just more can’t-see-the-forest-for-the-trees. Rather than elucidate problems with useful data, you obscure them with irrelevant data.
Chipmunk –
First the article linked below by Alex C. gives a good overview of what I was getting at.
Second, you say:
> most people’s weight is surely controlled by their ratio of calorie intake to calories burned in exercise
I’m the least math literate person here, and I realize I’m in a minority, but I think it’s useless to the point of inaccurate to say that a person’s weight is controlled by that ratio (or “just math”), if that “ratio” constantly changes as a function of myriad external factors. It’s some magical ratio that can’t actually be quantified or described. It changes to the point even where, it seems, as you become more fit or if other parameters change you approach the point where reducing your calories can lead to no weight loss (or even weight gain)
You appeal to the first law of thermodynamics, but again that seems to me to be so simplistic as to be not very instructive. I don’t know any more about physics than I do math, but the following seems to me to a valid part of the discussion:
The body is an open system, not a closed system. The first and second laws of thermodynamics confirm that a calorie is not a calorie: meaning that metabolic pathways determine the rate at which energy moves from its usable to unusable state. These two laws do not confirm that the body cannot increase in mass in the absence of excess energy coming in from the outside world.
https://edinstitute.org/paper/2015/6/13/gaining-weight-despite-calorie-restriction
I don’t know if that analysis stands up to scrutiny from people who understand physics, but it gets to the heart of what I was going for. The “ratio” of how calories TURN INTO WEIGHT isn’t fixed, it’s a dynamic, it a ratio as far as I can see.
Leaving aside the “fatttist” rhetoric from that link and other parts related to the psychosocial aspects (which are relevant to the practical reality of “it’s just math” but not greatly germane to the theoretical question at hand), perhaps this would be the more relevant quote:
That means restrict a bit and the metabolic response will not only match that deficit, but it will also overcompensate such that excess energy is now available within the system. Some individuals may release that excess into the greater environment (carbon dioxide and water) and others may partition the energy within the system (growth, repair and chemical bonds).
And while I think it’s over the top to talk about “fattists,” I do think that our difference here reflects a political disageement – one that you and I have run across before. I recognize that there’s a way that saying it’s nit “just math” can come across as a librul do-gooder trying to coddle people, and prevent them from dealing with reality and facing adversity head on.
But on the other side, simplifying the complex process and saying its “just math,” or boiling it down to a simplistic energy balance model, has proven to be a largely ineffective and more importantly misleading way to understand body weight. In pint if fact, even beyoind the psychosocial aspects, just saying it’s a matter of energy in verus energy out can be b broadly informative but in the end, has very limited utility for managing adiposity
Last comment on this (I promise)…
Just Googling on the physics angle a bit, I think the article below offers an interesting critique of my comments, but also suggests that indeed, it’s not a simple question.
My guess is that like with many things, how you evaluate applying the first law of thermodynamics to body weight gain or loss largely depends on your starting premises (should I say priors?)
https://physics.aps.org/articles/v12/47
As I said before:
Your problem is that you are fundamentally incapable of distinguishing between things people are actually saying, including you, and some broader point being made in your head.
“For every 2,000 steps you take each day, your risk for premature death may fall by 8 to 11 percent, according to research published in the journal JAMA Internal Medicine.”
Am I really the only one who read that sentence a bit weird? ‘Steps a day’ is a daily measure, but ‘premature death’ is not measured by day. So I read this as saying that every day I take 2000 steps, my overall risk of of a premature death is reduced by 8-11%, in other words, a few weeks of fairly short walks and I’m set for life!
I think this is a bigger issue.
I recently had a number of online arguments with scientifically sophisticated people (and some who weren’t) who absolutely insisted that observational studies are sufficient for drawing causality.
Maybe it’s an interesting question as to whether controlled studies are any less insufficient, but although many people can understand in a sense that “correlation doesn’t equal causation,” understanding the limitations of observational studies is a tough concept. I think that’s in some part because inferring causality from associations is basically how we’re wired as humans and to a large degree it works! The problem is that it doesn’t work as well when examining phenomena at a higher resolution, in the sense that takes place in scientific study.
For that reason, yes, I think that’s deceptively difficult to not slip into the default mentality, and to always maintain the discipline needed to not speak in causal terms when describing associations that we think might be a result of causality.
BTW, in addition to the fixed limitation, that identified associations should not be described as causal (without additional evidence of the sort outlined in Bradford HIll’s criteria), neither should cross-sectional data be described as casual (IMO, you need longitudinal data to describe causality – or at least direction of causality). And that’s a mistake that I think occurs even more often.
Joshua wrote:
“I recently had a number of online arguments with scientifically sophisticated people (and some who weren’t) who absolutely insisted that observational studies are sufficient for drawing causality.”
The implication of this statement is that an observational study cannot be used to determine causality. This implication fails the Okla Natural Nuclear Reactor test. A single observational study resulted in a robust determination of causality which is deeply rooted in basic physics and has not been subsequently challenged.
The general point, that observational studies are generally insufficient to determine causality, is fine. But it is also true that a full coverage approach (“fault tree”) gets you to causation as well, in fact better than any other way including controlled, forward-causation trials. By that I mean if you are able to test every plausible root cause with observational studies and get results in each case that are easily interpreted, you can indeed determine causation outside of the specific conditions of the Okla situation. This actually happens in some fields!
Matt –
> The implication of this statement is that an observational study cannot be used to determine causality. This implication fails the Okla Natural Nuclear Reactor test.
I can’t really respond based on that test – I’m totally unfamiliar with it in any way.
> A single observational study resulted in a robust determination of causality which is deeply rooted in basic physics and has not been subsequently challenged.
Maybe my statement was too broad – but it does seem to me that you’re brining in here confirmatory evidence related to physics. So yes, as there’s sufficient confirmatory evidence from outside of the observational study itself, the power of an observational study to “determine” causality grows. But in the context of this post, I would argue that the confirmatory evidence regarding mechanisms of causality is pretty damn mixed.
> The general point, that observational studies are generally insufficient to determine causality, is fine.
Ok, so I’ll go with that.
> But it is also true that a full coverage approach (“fault tree”) gets you to causation as well, in fact better than any other way including controlled, forward-causation trials. By that I mean if you are able to test every plausible root cause with observational studies and get results in each case that are easily interpreted, you can indeed determine causation outside of the specific conditions of the Okla situation. This actually happens in some fields!
Again, I can’t comment on the Okla test situation, but sure, a battery of coordinated observational studies, particularly if coordinated to inform various theories about mechanisms of causality, in particular if in aggregate they effectively represent longitudinal data, could make a compelling case!
Are we in disagreement?
Daniel said, “Martha: perhaps a consult with your doctor and a personal trainer could help. Maybe give you some strength building exercises, and a diet modification for muscle weight gain? what works for 30-40 year olds is definitely not going to be the same for people 70+ so probably take everything with a grain of … well maybe not salt … I don’t know, maybe curry powder :-)”
Ah, how naive you sound. Finding a high quality primary care doctor around here is a very difficult task –there was one who tended to try to do everything himself, rather than referring me to a specialist; that resulted in a couple of botched attempts at treatment. Then there was the one who seemed to think that vitamin D2 was something like a panacea. And then the guy who seemed to get his jollies by doing pelvic exams on women patients (fortunately, he moved out of town.)
Oh dear. It can be something of a challenge in my area (Los Angeles county) but the options sound vast compared to your area. Sorry to hear it. Perhaps a telemedicine consult with a gerontologist since we’re not really talking about diagnosis or treatment of a specific disease etc, in person might be less important than just talking with someone who knows something useful about your type of situation… good luck!
With regard to weight loss: I don’t think that exercise is an effective way to lose weight. See, for example: https://www.vox.com/2016/4/28/11518804/weight-loss-exercise-myth-burn-calories
Case in point: I started an exercise program (running and riding a stationary bike) about three years ago, at the start of the Covid lockdown. I’ve kept it up since then. I haven’t lost any weight at all. Maybe I’d lose weight if I altered my diet, but I haven’t done that. Dieting is hard! Much harder than sticking to an exercise program.
Exercise works fine as long as you don’t start eating more to compensate for the extra calories being burned. Which is of course exactly what people will do if they can.
So even though it really is about calories in < calories out, the first is controlled by food cravings. Keep it under ~100 grams of carbs per day (~400 calories; ~20% of calorie intake) and most people find they *want* to eat less. Twenty grams per day works even better. But it is like an addiction, if you go back to the original diet then the weight will come back.
Different things work for different people. For me, high-intensity exercise has a big effect on my weight, but lower-intensity exercise does not. Here’s an old blog post about this. https://statmodeling.stat.columbia.edu/2018/07/10/exercise-weight-loss-long-term-follow/
I mentioned in that thread, but again: 5-10 lbs is the size of a water weight fluctuation (0.5-1 kg of glycogen + 1.5-4 kg of water + 0.1-0.5 g of whatever ions like potassium). So that is not necessarily a “big effect”. It is something people can lose via crash dieting within a few days.
It would be better to see the timeseries and also know when you measured the weight (time of day, post-prandial, etc).
Anon,
My most extreme experience with hydration effects is from about ten years ago. I biked from Berkeley to the summit of Mount Diablo and back, via a hilly route on a very hot day. It’s about 100 miles. I started with two full 16-ounce water bottles and stopped just about everyplace I know on the way where I could fill them. At each water stop I would fill one of the bottles, chug it, then refill both bottles and continue my ride. I weighed myself before starting, and upon return. By the end of the day I had drunk 20 pounds of water and 12 ounces of Coke…and I had lost over 7 pounds.
But although I _can_ lose 5-10 pounds (or perhaps more) through fluctuations in water weight, in fact I do not except in extreme circumstances. I weigh myself infrequently but there have been times I have weighed myself on consecutive days and I’ve never had more than a 2 pound variation. A fraction of a pound is typical.
Perhaps it’s worth mentioning that when I weigh myself I do so in the morning, after going to the bathroom.
It is normal for water weight to work like this. It is steady then suddenly increases/decreases over a few days, or even one day. So it’d be informative (to you) to know how gradual the weight loss/gain is.
But either way, my point is 10 lbs isn’t really a lot, I’d say it is right at the lower bound of meaningful.
It sticks out to me because when I checked the nutrition literature there were tons if papers reporting (statistically) “significant” weight loss of a few lbs. All happily citing and misleading each other, of course. Losing less than a daily fluctuation after months on a diet is not a successful intervention.
No, no, ten pounds is not at the lower bound of being meaningful. I feel much better at 192 pounds than at 202: more energetic and more able to do things like walk or jog up a steep hill or bike quickly up a hill. Of course much of this is correlation rather than causation: when I’m ten pounds lighter because of exercising, I’m not just lighter I’m also fitter. Indeed this is why I only weigh myself infrequently: I don’t really care what I weigh per se.
But when I dropped from over 200 to the low 190s, the weight I lost was all fat. Actually I lost even more fat than that, I just replaced some of it with muscle. That fat wasn’t doing anything for me as far as providing power for biking or hiking. Dropping 5% of my weight without losing any muscle meant my power:weight ratio improved 5%. That is not “borderline of being meaningful”, at least not to me.
“But either way, my point is 10 lbs isn’t really a lot, I’d say it is right at the lower bound of meaningful.” – Anoneuoid
“No, no, ten pounds is not at the lower bound of being meaningful.” – Phil
I suspect both of these are correct depending on the population you’re talking about.
I used to be an elite amateur road cyclist and race elite-am and pro road bicycle races. As Phil says, weight is pretty important to this sport (as it is in many others). Using strict training, diet, and hydration practices, I could keep my weight within a pound or so (weighing at the same time every day). You could detect small weight increases and decreases this way (and, it wasn’t hard to confirm, as I was already so lean, that small additional fat loss was noticeable visually and via skinfold measurements). After all, if you’re chasing down wheels and pedals that weigh 50g and 25g less, then an extra pound or two of fat or upper body muscle is ridiculous. Anecdotally, I hear that in some big pro teams today (for example DSM last year), all the athletes are given only the amount of food calculated to be needed to maintain weight in big stage races like the Tour. This is accomplished by data on work done on the bike (via power meters that measure work in KJ from strain gauges), logging all food/drinks taken during the race, and likely lab studies on efficiency for the riders. The riders are then only given the food they need to prevent weight loss/gain.
On the other hand, several years ago, I briefly worked on a project where I tracked the weight and nutrition (via food logs) of a couple of dozen ‘normal’ (i.e. not elite athletes) people who were overweight and trying to lose weight. I was very surprised at the dramatic weight swings that I saw (which if I remember correctly were in excess of 10 pounds). At the time I thought this was due to water weight fluctuations as Anoneuoid mentions and weird eating practices where people would do great all week on their diet and then go to Chiles and blow it on the weekend. In this case, the 10 pound weight loss during the week wasn’t ‘real’, as Anoneuoid describes.
Yes, if Phil indeed lost 10 lbs of fat then that is substantial. But the weight shown on the scale cannot tell us if that was the case.
If it was gradual, then it was probably fat. If it was suddenly lost and gained back when he stopped exercising then it was probably water weight.
By suddenly I mean over the course of a few days vs gradual being 1-2 lbs per week.
Btw, another issue with studies on this topic is only showing the average timeseries. If some suddenly lose 10 lbs after 1 week, others after 2 weeks, etc then the average weight gradually decreases, which is completely misleading as to the mechanism.
Daniel wrote,
“Oh dear. It can be something of a challenge in my area (Los Angeles county) but the options sound vast compared to your area. Sorry to hear it. Perhaps a telemedicine consult with a gerontologist since we’re not really talking about diagnosis or treatment of a specific disease etc, in person might be less important than just talking with someone who knows something useful about your type of situation… good luck!”
In an earlier comment, you suggested a good trainer — that is more findable than a good primary care physician, although there is still the question of their availability.
And making things still more complicated, the weight loss started shortly after a colonoscopy to verify if I still had diverticulosis. And further complication: The gastroenterologist’s PA said I should have a regular gynecologist, because there is noticeable overlap between symptoms of diverticulosis and common “symptoms” of being postmenopausal.