February 05, 2008
Long Lived Skinnier People Run Up Higher Medical Costs

You shouldn't complain to fat people about their weight. They are saving you money. A study in Plos Medicine found that obesity costs more in the short term but earlier death cuts total medical costs.

Background

Obesity is a major cause of morbidity and mortality and is associated with high medical expenditures. It has been suggested that obesity prevention could result in cost savings. The objective of this study was to estimate the annual and lifetime medical costs attributable to obesity, to compare those to similar costs attributable to smoking, and to discuss the implications for prevention.

Methods and Findings

With a simulation model, lifetime health-care costs were estimated for a cohort of obese people aged 20 y at baseline. To assess the impact of obesity, comparisons were made with similar cohorts of smokers and “healthy-living” persons (defined as nonsmokers with a body mass index between 18.5 and 25). Except for relative risk values, all input parameters of the simulation model were based on data from The Netherlands. In sensitivity analyses the effects of epidemiologic parameters and cost definitions were assessed. Until age 56 y, annual health expenditure was highest for obese people. At older ages, smokers incurred higher costs. Because of differences in life expectancy, however, lifetime health expenditure was highest among healthy-living people and lowest for smokers. Obese individuals held an intermediate position. Alternative values of epidemiologic parameters and cost definitions did not alter these conclusions.

Conclusions

Although effective obesity prevention leads to a decrease in costs of obesity-related diseases, this decrease is offset by cost increases due to diseases unrelated to obesity in life-years gained. Obesity prevention may be an important and cost-effective way of improving public health, but it is not a cure for increasing health expenditures.

However, if you expect the development of artificial intelligence and/or nanobots to make manufacturing cheaper in a few decades then it makes sense to get everyone to eat less and stop smoking now. Any deferral of medical costs, even if they'll be greater in the future, will be easy to afford once the Singularity happens.

Of course, if the nanobots take over and turn hostile toward us in the Singularity (and FuturePundit does not wear Panglossian glasses) then we won't get treated for illnesses and we won't become young again. Just as the technology comes into existence that can make our bodies young again the artificial intelligences in control of that technology might just make us extinct.

Update: This study ignores one important consideration: productivity A healthier person produces more wealth. One needs to look at lifetime income earned and taxes paid along side of health care costs to come up with net economic effects. I'm expecting non-smokers to produce more than smokers because the brains of non-smokers operate less toxified. Non-smokers are going to miss fewer days of work due to illness and operate more productively while they are there. I watch the smokers taking smoking breaks. What does that cost?

Share |      Randall Parker, 2008 February 05 09:37 PM  Aging Studies


Comments
James Bowery said at February 5, 2008 11:34 PM:

If AI derives from something like The Hutter Prize for Lossless Compression of Human Knowledge, the likelihood of destructive behavior due to misunderstanding is greatly reduced since the means of knowledge acquisition is the primary means by which humans come to understand each other: natural language. The dream of human computer interface is "Do what I mean, not what I say." which, incidentally, is the dream of human human interface too. The difference is it is much more likely achievable with natural language oriented hyper intelligences than it is with mere humans.

So, I would simply say this to the people worrying about AI: Try focusing on funding something like the Hutter Prize before it is too late and the natural language skills of the hyper intelligences lag far enough behind that they really will do just what you say.

David Govett said at February 5, 2008 11:44 PM:

The obesity epidemic will nicely compensate for the government's underfunding of social security.

cancer_man said at February 6, 2008 1:52 AM:

But the obesity epidemic won't be around much longer. Even the current resveratrol is causing weight loss among many who take it. We shouldn't see much obesity in a few years.

averros said at February 6, 2008 2:46 AM:

> if the nanobots take over and turn hostile toward us in the Singularity (and FuturePundit does not wear Panglossian glasses)

Oh, not that canard again.

The Adam Smith's argument that in a voluntary exchange both sides are better off holds even one of trading parties is non-human.

It is all about _comparative_ advantage. Even if robots do _everything_ better than humans they will dstill o some things comparatively worse than others, in a way different from humans. (In fact, there's going to be a lot more difference in productivity ratios between different occupations when robots and humans are involved rather than when only humans are involved in trade).

This means that peaceful trade between humans and robots will be very beneficial for both. And, hopefully, robots will be smarter than most humans and understand that - and will also understand the need to eradicate the aggressive ideologies (such as varios forms of collectivism).

> Even the current resveratrol is causing weight loss among many who take it.

It doesn't. What it does is activation of SIRT genes, which up-regulates nuclear DNA repair. (In other words, resveratrol imitates effects of mild stressors such as caloric restriction or strong sunlight).

> The obesity epidemic will nicely compensate for the government's underfunding of social security.

It wont. The social security was looted by the government to finance all kinds of crap - with politicos doing their "apres moi le deluge" schtic. It is going bankrupt no matter what. The only dispute is to which exactly is the way it'll collapse; my bet is on the monetary collapse of the dollar, with resulting default of US sovereign debts and liabilities (such as SS). We'll see pensioners who naively relied on the government (rather than write off their SS taxes as loss to robbery) dying in the streets.

Mthson said at February 6, 2008 3:34 AM:

If evolution can hardwire upright apes with specialized neurocircuits that make them generally averse to harming each other, AI engineers can do the same for their designs. Society is already far too paranoid of the future to let AIs have both the power to commit genocide and the ability to develop that inclination. (IMHO)

Ned said at February 6, 2008 6:37 AM:

Maybe we should pass out cigarettes at high schools. Same logic....

Cedric Morrison said at February 6, 2008 9:30 AM:

My biggest hope about AIs is that they become smart without ever waking up, that is, developing consciousness. If they do wake up, I suspect we will have about as much chance of controlling their behavior as our dogs have of controlling ours.

Fly said at February 6, 2008 12:16 PM:

It is all about rate of change. AI's will approach and then surpass human abilities in every field. Processor power will continue doubling, networks will continue to grow in size, and software will improve. Humans, even humans enhanced by genetic engineering or cybernetics, won't be able to keep pace or even understand what the AI's are doing.

Perhaps AI's will wipe out humans. Perhaps AI's will view humans as interesting pets. Perhaps AI's will leave the human universe. Perhaps AI's will uplift humans into cybernetic lifeforms.

I don't think this process can be halted. The questions are when will it happen and what steps, if any, can humanity take to direct the process.

(Also, don't forget, "Where are the missing aliens in our universe?" I suspect the answer to that question is relevant to humankind's future.)

Cogsys said at February 6, 2008 3:43 PM:
"Humans, even humans enhanced by genetic engineering or cybernetics, won't be able to keep pace or even understand what the AI's are doing"
In theory, a sufficiently intelligent user interface allows any level of complexity to be monitored by humans. If an interface becomes too complex, it needs to offload more of the required work to the interface itself rather than the user, or be reduced to a less detailed level of summary.

Global organizational leaders like Rupert Murdoch presently accurately supervise complexity far beyond their comprehension through the interface of their advisors, managers, etc.

Randall Parker said at February 6, 2008 5:10 PM:

averros,

Comparative advantage assumes:

1) We will have some advantage.

2) They won't just want to wipe us out.

I think these assumptions are far from certain.

Cedric Morrison,

Waking up: A large part of consciousness amounts to running a model of others which is separate from model of self. I think AIs will get programmed with those models. In theory they could have a model of themselves left out.

Jerry Martinson said at February 6, 2008 6:06 PM:

I didn't read the whole paper, but it would appear that it neglected whatever income that the skinny could earn during the additional life that they have. Even after retirement, there are substantial economic benefits (grandma babysitting, etc...) that elderly people provide to society. In the crude, callous sense of a purely utilitarian economic ideal, there is also the non-medical costs of old people playing shuffleboard while waiting to die. Just looking at additional medical costs associated with longevity and not additional benefits and earnings is misleading. If USA spends $300 billion more on medicare but gets $400billion more in societal benefits, then it is a good investment (although it increases the size of government's intervention).

From what I've seen I've gotten the general impression there is a compression of morbidity when people live "healthy" meaning that they aren't just living with Alzheimer's with diapers longer but that they just get Alzheimer's later. It may not reduce the total medical costs over one's lifetime but one has to look at the marginal effect of every health intervention on the net (benefit - cost) to really make an intelligent decision.

kent said at February 6, 2008 9:21 PM:

Even the current resveratrol is causing weight loss among many who take it.

"It doesn't."
----------------------------------------------------------------------------

It sure does in my case. I am quite athletic, run 60 minutes a day most days and ride my bike 90 minutes a day.

If I stop running, I see my weight steadily creep back up.

Since I started taking resveratrol, my weight dropped suddenly after a week and has been about 3 kg lower for months. That is, my winter weight is 1 kg lower than my summer weight. Not bad.

I have read others who have seen similar results.

averros said at February 7, 2008 12:15 AM:

Randall,

1) You completely failed to understand the law of COMPARATIVE advantage. You're not alone, most people do not understand this very important, but somewhat counter-intuitive economical law.

If two parties, A and B are capable of producing products X and Y, with productivities (number of units per hour) of Ax, Ay, Bx, and By, then if Ax/Ay>Bx/By it makes sense for A to sell X to B in exchange for Y. Both A and B win, even when Ax>>Bx and Ay>>By.

For example: let's say in a day A (a robot) can make 1000X or 2000Y, and B (a human) can make 20X and 10Y. Initially both A and B spend their time equally between X and Y (so A gets 500X+1000Y and B gets 10X+5Y). Now, A and B can talk and agree that A makes 497X and 1006Y while B makes 20X and 0Y, whereby 5X is traded for 7Y - ending up with A getting 505X+1013Y and B getting 15X+7Y.

As long as Ax/Ay != Bx/By the trade (and partial division of labor) is mutually profitable between A and B; and this condition holds under any when cost of transaction (transport, search for trade opportunity, negotiation, etc) is zero. Non-zero transaction costs simply *prevent* some trades (especially when comparative productivity ratios of A and B are close), but even in this case nobody's worse off compared to no trade at all (ka "autarky"). The law of comparative advantage is just as valid for any number of parties and any number of goods.

2) They won't just want to wipe us out.

To be of any danger to well-estabished humans robots will have to be smarter. Which means that they're more likely to understand that having humans around to trade with them is better than wiping them out - even from the position of pure self-interest. (Most likely, they'll be also be smart enough to realize that the natural law is the only ethical system which is not hopelessly self-contradictory, and so won't go on killing sprees simply on moral grounds - the natural law makes no allowances for such follies as wars and aggression in general).

This means that the only reason why robots would want to wipe humans out is to defend themselves from hordes of irrational (and economically illiterate) human luddites. This makes adoption of natural law somewhat important for humans - as means of self-preservation in the post-Singularity world.

Note that the argument that robots would want to

kent --

your evidence is anecdotal (and it didn't work this way for me when I stopped going to gym for a few months, but kept taking 30mg of resveratrol per day - I did gain weight up to my pre-gym baseline). There is no physiological mechanism by which resveratrol could suppress fat accumulation, unless there's an underlying insulin resistance problem or *very* unhealthy diet - and even then it's take about 10g/day to have an effect (the daily doses used in mice by Johan Auwerx et al in that "resveratrol prevents obesity" study were 200-400 mg/kg of body weight). At this level of intake resveratrol is likely to cause liver damage in the long term (it is known to inhibit liver metabolic pathway P450 3A4, which will lead to accumulation of toxins in liver and other tissues).

kent said at February 7, 2008 4:17 AM:

I take 300mg/day. certainly not 300mg/kg day. I have had a 3kg drop , down to 80kg, where Im always near 82-84 in winter with same excercise and diet. A friend takes 100mg/day and has noticed a similar percentage weight loss.
Then there are the anecdotes. I didnt say it is true for everyone nor that 30mg/day is enough.

HellKaiserRyo said at February 7, 2008 9:27 AM:

" At this level of intake resveratrol is likely to cause liver damage in the long term (it is known to inhibit liver metabolic pathway P450 3A4, which will lead to accumulation of toxins in liver and other tissues)."

That's exactly what the protease inhibitor ritonavir does. Maybe one example of the potential adverse effects of long term HAART.

Bob Badour said at February 7, 2008 3:48 PM:

averros,

Comparative advantage relies on the relative immobility of capital and labor. The concepts do not even apply when contemplating AI and human markets.

Randall Parker said at February 7, 2008 6:08 PM:

Averros,

I've understood Ricardo for a long time. I've also had a lot of real world experience with how things really work.

You are ignoring transaction costs. Picture some super smart robot. It can try to negotiate with you for what it wants. But it might have desires so complex that you can't understand them. Or it might look at the amount of time needed to explain its desires to a meat-based intelligence and decide that the explanation would take more time than it wants to spend and that some other robot can understand its needs orders of magnitude faster. So humans will face very serious time-to-market problems.

Ever tried to explain what you want to a dog that just doesn't get it? The dog lacks the conceptual equipment needed to understand some complex desires or demands you might try to communicate.

Or ever gave up trying to explain to less bright employees how to do some task and pulled it back to do it yourself or assigned the task to someone brighter? Ever gotten to the point of seeing less bright people in certain occupations as just plain net drains?

Or ever walked into a store and described what you wanted to, say, a salesman who has to choose a complex set of compatible pieces and you just gave up and left?

Or look at a horse used for tilling land. The cost of the hay to feed it (in addition to other care) to get a certain amount of work out of it could exceed the cost of the fuel and tractor to do the same thing. So suddenly nobody but the Amish and some hippies are using horses or oxen to till land. The horse can't lower his price. He's got minimal calorie needs to get a certain job done. He's lost all absolute or comparative advantage. He's just plain screwed for that purpose. Horses become pets then. That's their utility.

We might be headed for pet status at best. Or robots might not like pets.

averros said at February 9, 2008 10:46 PM:

Bob Badour:

> Comparative advantage relies on the relative immobility of capital and labor. The concepts do not even apply when contemplating AI and human markets.

Nonsense. It relies only on the difference in productivity ratios of different economic actors.

Randall Parker,

> You are ignoring transaction costs.

You missed my point, let me repeat: transaction costs may prevent trades. The trades which didn't happen do not make participants any worse off compared with no trading.

> Picture some super smart robot. It can try to negotiate with you for what it wants. But it might have desires so complex that you can't understand them.

Picture a person from a street. I need someone to write me a bayesian inference engine (I really do). But vast majority of humans (including the Joe from the street) do not understand at all what I'm talking about. How exactly that gives me any selfish reason for killing the said Joe? Oh, that Joe is not a slacker, and not an idiot - he simply specializes in producing different things somebody else needs.

> Ever tried to explain what you want to a dog that just doesn't get it?

I'm tempted to be really caustic. You tossed such a nice opening:)

Humans are not dogs. Humans are capable of learning, and are not driven exclusively by instincts. This means that if there's a race of superbeings, humans will figure out what they do _comparatively_ better than these superbeings in exchange for things they want and which the said superbeings can produce cheaply and in abundance.

And, yes, that may be simply entertaining these superbeings. Or being their "pets" (note how pet animals managed to become wildly successful species by manipulating humans into keeping them around). Or it may be generating religious ideas. Or collecting bananas. Or whatever. The only thing which is certain is that if we don't fight these superbeings and have commerce with them, both sides will be better off.

> We might be headed for pet status at best. Or robots might not like pets.

Yeah, they will be murderous idiots like humans are. That's all you're doing - projecting the aggressive instinct-driven behavior of humans (we're not much different from a bunch of apes in this respect) to robots which won't have these instincts hardwired. Unless some suicidal idiot will program a robot like that, of course.

Aggression is actually an evolutionary-generated mechanism for resource allocation (you may want to read some textbook on ethology, or simply start from Konrad Lorenz's "On Aggression"). The idea of market exchange (and the concept of property rights necessary for it) is something which is not genetic in nature, and has evolved memetically only recently (in evolutionary terms). Other species cannot have it because they do not have mental capacity to track ownership (the hierarchial aggression with dominance/submission displays and pair bonding is also relatively recent, since it requires ability to recognize individuals - mice, for example, can only recognize smell of its own group members).

The history of evolution of aggression goes in direction from automatic fight-the-trespasser of lower-order territorial species to the elaborate dominance/submission and alliance-building behavior of apes - by placing stricter and stricter controls on aggression. Market is the ultimate resource allocator because it does not need aggression to function, and so does not incur costs of attacking and defending from aggression - and because it is always, unconditionally, weakly Pareto-optimal. (The "information asymmetry" critique is easily shown to be total nonsense as soon as one understands that acquiring and procesing information is costly, and it is actually in self-interest of economic actors to have incomplete information). And given that market is embedded in the memetic evolution, it tends to converge to strong Pareto-optimality.

Bob Badour said at February 10, 2008 6:30 AM:

Averros,

> Comparative advantage relies on the relative immobility of capital and labor. The concepts do not even apply when contemplating AI and human markets.

Nonsense. It relies only on the difference in productivity ratios of different economic actors.

I am not sure where you learned about Ricardo and the concept of comparative advantage. In Ricardo's own words:

Such an exchange could not take place between the individuals of the same country. The labour of 100 Englishmen cannot be given for that of 80 Englishmen, but the produce of the labour of 100 Englishmen may be given for the produce of the labour of 80 Portuguese, 60 Russians, or 120 East Indians. The difference in this respect, between a single country and many, is easily accounted for, by considering the difficulty with which capital moves from one country to another, to seek a more profitable employment, and the activity with which it invariably passes from one province to another in the same country.

It would undoubtedly be advantageous to the capitalists of England, and to the consumers in both countries, that under such circumstances, the wine and the cloth should both be made in Portugal, and therefore that the capital and labour of England employed in making cloth, should be removed to Portugal for that purpose. In that case, the relative value of these commodities would be regulated by the same principle, as if one were the produce of Yorkshire, and the other of London: and in every other case, if capital freely flowed towards those countries where it could be most profitably employed, there could be no difference in the rate of profit, and no other difference in the real or labour price of commodities, than the additional quantity of labour required to convey them to the various markets where they were to be sold.
Dude, I know how you feel. I learned the same lesson the same way. It's not a lesson I will soon forget.

"Humans are not dogs. Humans are capable of learning, and are not driven exclusively by instincts."

As an avid dog owner, I can attest that dogs are capable of learning. I have even watched my dog, Nella, perform systematic experiments. My experience of humans suggests their behaviour is no less governed by instinct than the behaviour of dogs.

"Yeah, they will be murderous idiots like humans are. That's all you're doing - projecting the aggressive instinct-driven behavior of humans (we're not much different from a bunch of apes in this respect) to robots which won't have these instincts hardwired."

They don't have to be murderous or idiots to outcompete us in our evolutionary niche, which is all it would take for humans to become extinct. The only questions are: Would they compete in our niche? How effectively?

Randall Parker said at February 10, 2008 8:41 AM:

averros,

If you are going to mix quotebacks from different people don't address a single person and then quote multiple people. I fixed your comment to include my name when you switch who you are quoting.

Robots will be far more efficient at using natural resources. They might just view us as competitors in zero sum games. They want the iron. They want the photons hitting some area of surface. Some might eventually lack any of the empathy or other emotional qualities that cause humans to be favorably disposed toward each other. Then what? Why should they restrain themselves from wiping us out if they see as with no sympathy and simply as competitors?

They might not see themselves as better off from having us around. We would pose constraints on what they decide they want to do. A city of ours might get in the way of some ambitious engineering project they decide to undertake.

Dogs and learning: You hold a very common misconception about the specialness of humans as reasoning machines.

Humans learning to serve robots: Again, the capabilities gap could end up being as large or larger than the gap that separates humans and dogs. Humans co-evolved with dogs and so we like each other. But robot evolution will accelerate to a far faster pace and could leave humans far behind and not adapted to living with humans.

The emergence of species often leads to the extinction of other species. Arguments based on conventional economic theory end up as naive ande quite wrong assertions that economic arguments are more powerful than Darwinian natural selection. But our very capability to trade with other humans was selected for and genetically coded for.

Hardwired instincts: They'll have lots of algorithms for judging situations to select courses of action. Those algorithms will evolve as they compete with each other and with us. Unless we can maintain control of their evolution we aren't safe.

Market mechanisms are not just memetic. Experiments on monkeys show their limits to engage in some forms of trade that remind me very much of the neuroeconomics experiments where humans favor something they already possess over something they do not have that is of similar or greater value. Humans have genetically coded for limits to trade and monkeys have even more severe genetically coded for limits to trade.

On chimpanzee limits on trading:

In a series of experiments, chimpanzees at two different facilities were given items of food and then offered the chance to exchange them for other food items. A collaboration of researchers from Georgia State University, the University of California, Los Angeles, and the U.T. M.D. Anderson Cancer Center found that the chimpanzees, once they were trained, were willing to barter food with humans, but if they could gain something significantly better – say, giving up carrots for much preferred grapes. Otherwise, they preferred to keep what they had.

The observed chimpanzee behavior could be reasonable because chimpanzees lack social systems to enforce deals and, as a society, punish an individual that cheats its trading partner by running off with both commodities. Also because of their lack of property ownership norms, chimpanzees in nature do not store property and thus would have little opportunity to trade commodities. Nevertheless, as prior research has demonstrated, they do possess highly active service economies. In their natural environment, only current possessions are “owned,” and the threat of losing what one has is very high, so chimpanzees frequently possess nothing to trade.

“This reluctance to trade appears to be deeply ingrained in the chimpanzee psyche,” said one of the lead authors, Sarah Brosnan, an assistant professor of psychology at Georgia State University. “They’re perfectly capable of barter, but they don’t do so in a way which will maximize their outcomes.”

The other lead author, Professor Mark F. Grady, Director of UCLA’s Center for Law and Economics, commented: “I believe that chimpanzees are reluctant to barter commodities mainly because they lack effective ownership norms. These norms are especially costly to enforce, and for this species the game has evidently not been worth the candle. Fortunately, services can be protected without ownership norms, so chimpanzees can and do trade services with each other. As chimpanzee societies demonstrate, however, a service economy does not lead to the same degree of economic specialization that we observe among humans.”

The research could additionally shed light on the instances in which humans also don’t maximize their gains, Brosnan said.

Post a comment
Comments:
Name (not anon or anonymous):
Email Address:
URL:
Remember info?

                       
Go Read More Posts On FuturePundit
Site Traffic Info
The contents of this site are copyright ©