Although it is certainly written by a human being, this article gives me a similar feeling to ChatGPT essays. The tone is compelling and plausible. But when I thought about it, I realized it didn't make much sense.
> Rarely do we even consider that the cost of doing something might outweigh the benefits.
Cost vs benefit is the main thing that people consider when making decisions. It's a core decision making framework taught as early as elementary school social studies. Institutional decision making is rife with processes for assessing risks and performing cost-benefit analysis.
> Intervention—by people or governments—should only be used when the benefits visibly outweigh the negatives.
This advice is the decision making equivalent of "just don't write code with bugs." Easily achievable but pretty useless and completely impractical. Real world decision making involves operating under uncertainty at almost all times. If it didn't, the topic wouldn't be worth writing articles about.
> Cost vs benefit is the main thing that people consider when making decisions.
I don't know what people you are talking about, but my experience is so different from this that I suspect your conclusion is purely platonic. You expect people to decide based on a cost/benefit analysis, or you hear people telling you they do.
People almost never do a cost/benefit analysis. On the best cases, they analyze one of that pair.
Iatrogenics happens when the benefits are clear and the harms are hidden. So “rational” people tend to intervene because it would seem the benefits outweigh the costs.
Correct. To parse that further, the costs may be hidden, involve unknowns and/or unknown unknowns.
This is the explanation many of the unvaxed smart people I know used to justify relying on their immune systems which were developed and tested for a million+ years over recently developed covid vaccines.
> Iatrogenics happens when the benefits are clear and the harms are hidden. So “rational” people tend to intervene because it would seem the benefits outweigh the costs.
Also as always you must remember that not every single person will be harmed equally. And in many cases solutions are chosen not because they are best(cheaper/least harmful), but because they harm specific portion of population. Think why swimming pools were filled with concrete and many more.
> > Rarely do we even consider that the cost of doing something might outweigh the benefits.
> Cost vs benefit is the main thing that people consider when making decisions.
This article also makes the mistake of equation "we", the average reader of the blog, with the doctors who are making prescribing decisions.
Understanding the risks and side effects of prescribing a medication or procedure is core to the practice of medicine. It's integral to everything doctors do.
The field of medicine is often criticized for not being aggressive enough with treatments, though these criticisms often come from people who only see the potential upside but not the risks. We saw this most recently with complaints about how long the FDA took to get vaccines out (which was actually lightning fast) or in how the field of medicine didn't rush to recommend theoretical COVID treatments based on small-scale studies and petri dish studies (which turned out to be the right choice, given that none of them replicated at scale in controlled studies).
It's also the reason why the most effective medication may not be the first-line treatment, because the first-line treatment is chosen as a tradeoff of risks versus benefits.
Doctors know all of this. Unfortunately, doctors who decline to give patients the medications they think they need or recommend against surgical procedures patients think they want are subject to a lot of negative pressure in today's medical system. If someone goes into a doctor and demands antibiotics for their cold, the doctor must find a way to talk them down without risking another negative review which can impact their career and compensation in many systems. Some doctors just don't care enough, and will hand out prescriptions as asked. It's becoming a real problem now that medical systems have so many layers of middle management trying to put NPS scores and other feedback loops into the medical system, combined with social media pushing so many people to think they know they need antibiotics or thyroid medicine or other treatments that come with more downsides than benefits.
I think they are saying that there are known costs and benefits but unknown costs are hard to quantify and unknown benefits are probably rare. Therefore the benefits need to be very material (several orders of magnitude as they state). Don't trade noise for noise.
Or their cost-benefit analysis said that the cost of removing kids from schools was worth it.
Do you have any evidence of ignored cost-benefit studies that were available at lockdown decision time (there wasn’t much time for studies when these decisions needed to be made)?
There was plenty of evidence that COVID was not as dangerous to children as it was to elderly and yet politicians kept schools locked down. This is why many parents took their kids out of public school and into private schools which re-opened, or at least families that could afford it did. There has been a massive impact to school-age children and the development loss is massive. See this article from 2021: https://www.medpagetoday.com/opinion/vinay-prasad/90658
Many of the studies referenced in that article are from 2020, yet schools had not re-opened.
I don't think it does. If you have a ? in your cost-benefit spreadsheet for "cost in DALYs of locking down schools", the obvious solution (that one assumes governments actually used) is to ask your education expert for a guess and a confidence interval.
Getting that estimate wrong is not the same as failing to consider it.
Or their cost-benefit analysis said that the cost of removing kids from schools was worth it.
Cost is incurred by the children who have no voice, so zero cost to decision makers. Benefit is to one of the largest donor groups to politicians, so also large benefit to decision makers. Cost/benefit to society, population at large, the country? Seems to be willfully ignored.
It takes homeostasis, a biological mechanism, and tries to suggest that it applies to politics and economy, too. What a strange ideea...
Yes, the decisions to intervene should be taken more carefully, but I don't think that interventions "often leads to worse outcomes". In fact, where I live, indecision and lack of interventions are a very big problem.
There is a fairly famous story in management of a guy in the 1960s who fought against building housing projects for the poor by tearing down neighborhoods and building big high rises.
The idea was to give more people a better place to live.
What the opponent said was you're going to concentrate a bunch of poor people in a building that neither we nor they will maintain, it will be far away from any jobs, and it will lead to more crime and more poverty.
I don't think it takes much to say that people are short-sighted and some are only interested in their own personal short-term gain, certainly no metaphor or analogy is needed to state the obvious.
This is a very interesting story, but I feel like I'm missing something -- any chance you could dig up a reference to this for us?
Housing projects and private high rises seem like they would have similar problems...I'm not finding this easily, but confess I have no academic background in management!
> any chance you could dig up a reference to this for us?
you could search for Robert Taylor Homes in Chicago (torn down) or Pruitt-Igoe towers in St. Louis (torn down), for starters. Or the Cabrini–Green Homes in Chicago.
Ah, thanks, yes I'm familiar with the legacies of torn-down public housing projects.
It was the private ones and the implied comparison that I was after -- but now that I'm re-reading the parent post maybe there was no comparison, just a long sentence describing public housing projects as torn-down neighborhoods with high-rises...
> if a complex system has found some kind of steady state there must be some "restoring forces"
I think this is one of those things that is True with a capital T.
If you start with the premise that life is valuable, it is eventually derivable. Life tries to exist, and since existence is path-dependent making a change is likelier to be wrong than right.
I think you and the parent commenter are talking about the anthropic principle. It might be "derivable" but only in the particular situation in which we have found ourselves—and then it becomes less remarkable.
Complex systems can have multiple different 'steady' states (many valleys in a complex mountain range of energy minima and maxima, in that sense), so while there may be local restoring forces, climbing out of one steady state and into another is extremely common.
> making a change is likelier to be wrong than right.
This agrees with the slightly-neutral theory of evolution. A given mutation is more likely to break something a little bit than improve that something a little bit.
Imagine a future where espionage between nation-states becomes impossible not because of omnipresent surveillance but because the multinational advertisers realize that spies have nearly unlimited disposable incomes backed by other nation states so they have the best click through ROI and thus get targeted by the most bizarre ads that civilians aren’t.
The word “Intervention” is just fraught with unconscious bias and makes it seem like the intention that motivates an action is relevant when it is not. The characteristics noted in the essay apply to all actions.
No one would characterize the Industrial Revolution or adding lead to gasoline as “intervention” yet clearly the second-order thinking this article prescribes for “interventions” needed to be applied in those cases.
“The key lesson here is that if we are to act, we need a solid idea of not only the benefits of our actions but also the harm we may cause—the second and subsequent order consequences.” is more correct but doesn’t serve this author’s agenda.
Intervention wouldn't be so bad if it wasn't so hard to roll back on interventions set in motion that turned out to be a failure. When dealing with complex systems specially, our attempts to make things "better" often don't go so well.
The sunk cost fallacy applies to both interventionism and noninterventionism. "An ounce of prevention is worth a pound of cure" is an idiom for a reason.
What made interventions such as the Berlin Air Lift work is that they were stuck with until they were no longer needed. An unanticipatable effect of the Berlin Air Lift was the "Tear down this wall" speech and the end of the USSR.
How do you know that was an unanticipated effect? Also, what about all other cases where interventionism didn't work, that is, made things worse? How can you tell a priori wether intervening will have good results? Moreover, if you're a decision-maker, will you bear any cost if your decision turns out wrong? If not, how is that supposed to be ethical?
I'm just stating that lack of action is an action. I'm not arguing the points you're asking me questions about.
> How do you know that was an unanticipated effect?
Wanting to support Berlin so that the communist regime would eventually collapse might be a desired outcome, but the way it happened couldn't have been anticipated, it could at most be wished for. Wishfulness is the very thing the parent article is arguing against.
> How can you tell a priori wether intervening will have good results?
You try to learn history and figure out parameters for successful predictions?
> Moreover, if you're a decision-maker, will you bear any cost if your decision turns out wrong? If not, how is that supposed to be ethical?
I don't disagree? Ethics effects exist to encourage people to take precautions when it comes to the effects of their actions or inactions. There are obviously entrenched interests that push back against ethical consequences.
I read something recently here, comparing this phenomenon to the phenomenon of overfitting.
The gist is that we want to leverage our ability to measure things that are easy to measure, which are correlated with things that are hard to measure, and then optimize for those things. The downside being that you will always implicitly be sacrificing the things you don't/can't measure, and these sacrifices tend to occur in greater marginal amounts as the low hanging fruit gets eaten.
There are obviously interventions which are so ill convceived that they have no place being implemented in the first place, but this mechanism applies to every intervention eventually.
Related to N. N. Taleb, his colleague and CIO of Black Swan fund Universa Investments Mark Spitznagel [1] discusses iatrogenics in the context of hedge funds in a recent interview [0], at 18:20.
I learned this big time as a parent: If the only solution you have is guaranteed to make it worse, it's better to stand there and watch it burn than to put out the fire with gasoline. The piece I wrote about it (titled The Hand Licking Incident) did well on HN if you want to read a variation on this theme.
It's a good read, because these things happen every day around the world but the world usually finds people too dumbfounded in this to learn from it, much less write a cautionary tale!
Doctors are always thinking about the potential for harm in their therapies but you don't always have a lot of options. For example, amoxicillin is the standard of care for strep but some people are allergic. Azithromycin is also standard but overuse of macrolides is driving resistance in organisms that are already resistant to other antibiotics.
I understand the article is really trying to reach into medicine to find the reason to push back on other actions. If doctors recognize they can cause harm when they try to heal, so must other practitioners, presumably politicians, bureaucrats, engineers, marketers, activists, parents, etc.
It's not enough to identify the problem. We can bias for or against intervention. But neither bias is scientific or necessarily likely to produce a good outcome without externalities or adverse effects. Ideally we would try more things with small samples and also try to understand externalities, comparing that to other possible interventions or nonintervention. However, the status quo bias also has an implicit blind spot, which is assuming that things will remain the same. Unfortunately, conditions do change.
We also have a problem with the moral aspect of the question. Who has the right to intervene? Who has an obligation? Is it moral to decline to intervene? If you intervene for good reasons, but cause bad outcomes, is that a moral failing? Is one person's assessment of the outcome more morally correct than another's?
> The key lesson here is that if we are to intervene, we need a solid idea of not only the benefits of our interventions but also the harm we may cause—the second and subsequent order consequences. Otherwise, how will we know when, despite our best intentions, we cause more harm than we do good?
I've watched enough time travel shows to know that we don't fully know the "benefits" of our intervention, in addition to the "harm".
American politics are riddled with examples - look at the Patriot Act and the Affordable Care Act. It seems to happen most with sweeping legislation. Of course, large legislation will have additional consequences because the point is it intervenes more. The flipside is that administrations that don't pass laws are referred to in derogatory ways - they are a lame duck. We've basically forced action, any action, no matter the consequences.
That misses a lot of context. The Affordable Care Act is considered by many to be an attempt to meet medical care needs with a market solution while most competing ideas go a completely different route such as Single Payer. It isn't just a patch to a system, but an attempt to fit some difficult constraints. Portraying extremely complex systems as simple may be the most common form of the first failing listed here.
What context was missing? You can argue that the ACA has benefits or the US system is complicated but I did not simplify anything. From the article "Intervention—by people or governments—should only be used when the benefits visibly outweigh the negatives." My argument was 'damned if you do, damned if you don't' and politicians do not worry about the consequences of their actions.
You are asking a question yet seem committed to ideas and political views already. here is an attempt to answer:
Through the Industrial Revolution medicine increased in scope and complexity and cost. By the 1970s people expected access to hospital care which was increasingly expensive to provide. Around the world most developed nations came up with some kind of universal insurance or single payer system to handle this. The US went with employer based care which limits who is covered and puts a burden on business.
Ongoing calls for universal basic care threatened market based care in the US, so Mitt Romney came up with a plan to provide universal care through government regulated insurance markets. Obama tried to bring this plan to a Federal level but was stopped by a complex array of social and political forces. As a notable example of what happened, Frank Luntz who often has interesting ideas to contribute to this day insists that the ACA was not an insurance program even though providing health care through private insurance was the central concept.
So now there are people who see the ACA as a compromised attempt to save market based insurance provided medical care and others who see it as some kind of strange intervention amounting to socialized medicine which oddly enough is exactly what it was designed to avoid. Where ACA was adopted it seems to be working reasonably well, and in Massachusetts where the original Romney plan was adopted everyone is insured and gets care through private insurance markets coordinated by the government.
The really big irony here is that by compromising ACA and then not even adopting it in some states it is increasingly likely that some kind of Single Payer system will end up being adopted at a Federal level. In this sense those who attempted political intervention have generated the most harmful possible result. That is the increasingly likely possibility that the existing system of providing health care in the US through insurance markets will be completely swept away instead of being reformed as Mitt Romney and Barack Obama advocated.
It seems like you're focused on the implementation of the ACA here. I used the ACA as an example of secondary consequences. My question of context was because that's literally what you said. I was wondering how context related to my post which I could have been more clear about. You seem to be saying that context is everything no matter the secondary consequences and go so far as to give more context. Many people characterized the ACA as a Trojan horse at the time. Was that it's purpose? That's maybe a more appropriate question to all this - why doesn't congress add purpose to bills? We need measurable benefits to weigh. By sidestepping that part of the process, it's a guessing game - all bills are just a mash of secondary consequences. That's convenient for politicians and political parties but not the people.
No, lame duck administrations are when you’re in office with a selected successor.
Both of the laws referenced were very strategic laws that largely accomplished their purpose. The ACA in particular was designed to appeal to republicans first - it was in alignment with old guard GOP values. It patched some of the holes in healthcare (pre existing conditions, etc) and drove the economic activity that will drive future universal healthcare (the consolidation of providers into health networks) when political conditions allow.
Overall US governance is deliberately designed to make doing something impossible. Think of any issue of the day that would benefit from Federal action. In general, nothing at all happens unless it’s some spin on an existing law.
I agree that I didn't use lame duck in the textbook definition but it does show the derogatory nature of not doing anything. Likewise congress would be called ineffective if they didn't pass laws. We often hear about gridlock when there's a balance of power with the 2 party system. I think saying 'deliberately designed to make doing something impossible' is an overstatement. Compromise is required. I see more and more that neither party wants to compromise, though. From my perspective, it's hard to talk about any Federal action that would be of benefit. Granted I'm jaded from seeing and being part of second order consequences. What kind of Federal action did you have in mind?
Edit
I didn't address one of your points. You said the laws largely accomplished their purpose. What did you see as their purpose?
The Senate is a brake on change. In the most egregious example, the House passed anti-lynching legislation for 50 years, which never made it out of Senate committee.
Patriot act was ultimately designed to avoid the types of failures to coordinate that stopped the authorities from preventing 9/11 attacks. It was very successful at that aim, but created a bunch of other problems in the process. My statement isn’t an endorsement of it.
ACA reduced the number of uninsured individuals by 50%, eliminated a few key policy holes (ie pre-existing conditions) and slowed cost growth.
How successful was the Patriot Act? Without measurable benefits we can't really say.
For the ACA, the numbers I'm seeing are much less than 50%. From the Kaiser Family Foundation "While uninsured rates decreased across all income groups from 2013 to 2016, they declined most sharply for poor and near-poor people, dropping by 9.7 percentage points and 11.4 percentage points, respectively." From that same source "In 2017, the uninsured rate reversed course and, for the first time since the passage of the ACA, rose significantly to 10.2%." You bring to the table an important point, though. We need measurable benefits. The fact that there's even guessing at the purpose of bills shows the flaw. We should require goals with every bill. Without that we just have bickering on specifics which, by design, have never been communicated. Politicians use this lack of information to their advantage.
He talks about bloodletting but ignores the elephant in the room: circumcision.
You go to the US and you have to be careful if you have a boy because they will circumcise your child without asking for permission.
If you ask Americans they will tell you that it is an hygienic, prophylactic measure, but the real reason is that it became fashion to copy the British royal family that did it, and after that they just continued the tradition.
People support it not because they "don't have skin in the game" but because they have. Because the majority of them were already circumcised and there is no way back,or experience the alternative, so they want to believe there is a real reason for that, instead of the bogus Royal thing.
I asked a lot of people there and they painted a terrible world if you dare not to do that. I come from a country in which 90% are not circumcised and the image they had about the alternate reality was surreal and made no sense.
Another thing that Americans do is boiling eggs requirement, for "hygienic" reasons, destroying natural antioxidants, or natural antibiotics and antifungi so eggs could be way more dangerous than natural ones after time.
In France we had irradiated food for a long time. Crazy if they don't let you choose.
Today there are lots of things that could lead to secondary effects:
You have forced Water fluoridation.
You have forced GMO food you could no opt out.
Forced electromagnetic signals everywhere you can not opt out.
Forced flame retardant in foams, sofas and bed mattresses.
Forced COVID vaccination.
But the most absurd of all is the Daylight Saving Time.
I believe the real culprit of that behavior is human hubris. People simplify the world, and under the simplification a new idea is fantastic.
So for example you take the statistics and see smokers usually burn their houses while they smoke in bed and something happen, and forcing everybody to use flame retardants will save the risk of those people, but you can not see the new dead people that you create after they get cancer that will never smoke in bed.
> You go to the US and you have to be careful if you have a boy because they will circumcise your child without asking for permission.
No, they won't, unless they really want to lose a slam-dunk lawsuit and possibly face criminal charges. Outside of a very limited set of circumstances, you can't just perform surgery on a person without their consent (or in this case, the proxy consent of the parents).
I had an interesting experience as a child. I grew up in Barbados, where circumcision is not a common practice. In Biology class on day, the teacher talked about the practice, and basically everyone in the class went “Eww”. I moved to the US the next year, and when the same topic came up in Biology, the teacher mentioned that in other countries, most men are uncircumcised. And the class went “Eww”.
The foreskins are used for expensive skin rejuvenation cosmetics and other high end products.
Modern circumcision is evil. The foreskin is analogous to the clitorus. It is filled with nerve endings and keeps the head skin from keratinizing (hardening and losing sensitivity).
Biblical circumcision was a notch in the foreskin. It was not the radical removal of it entirely.
This misses the problem of scale. Often the only way to patch an existing system is to try small and localized changes and then propagate them slowly. Even correct changes may require unexpected related details to be in order for a successful result. Going slowly allows for analysis of related factors and anticipating range of variation.
This reminds me of Eisenhower's old quote to his advisors: “Let's make our mistakes slowly"
Which, ironically enough, if you make changes slowly, and have a good way to back them out when they go wrong, you can change much faster. Mostly because you don't have people screaming about a change being the end of the world so hard that you need to go to war in order to change the smallest thing.
Was hoping to find an intelligent and well researched article on medical interventions. Instead it's just some extremely hand wavy libertarian college essay.
this reminds me of the 'leading change' performance metric some mil leaders were (are?) evaluated on. the incentive is to be seen making changes, but there isn't anything about following up to ensure the change accomplished anything.
While conservatives love the stories of interventions failing with second order effects being worse than the initial situation, in practice these are only rare instances among plenty of successful intervention we don't even think about.
The main factor leading to a disaster is usually the desire to make grandiose interventions (instead of steadily increasing ones that have proven successful at small scale) and refusal to listen to early signs of something going wrong. That is, if you're Mao Zedong or any kind of dictators with a “brilliant plan”®, then your intervention have a high risk of such failures, but otherwise the odds are low.
In my experience, bad mental models in medicine lead to all kinds of horrors for entire classes of patients. If they decide to call it a genetic disorder, they throw their hands in the air and give up on actually making you better and trot out phrases designed to politely break you of demanding such, like "the normal progression...", a phrase intended to make you accept that you will get steadily worse, you hurtling towards your death is nothing to be concerned about and please stop annoying your doctor by expecting real help.
Medicine has become monopolized, politicized, and propagandized. There are many treatments for modern ailments that is legally forbidden for doctors to recommend or prescribe, or will cause the doctor lose their license.
Doctors operate under the Standard of Care rules. If the doctor does not follow SoC they open themselves to medical malpractice liability. Problem is that the SoC is imperfect, especially for less common afflictions. Doctors often know their prescriptions will not help their patients or even worsen conditions, but recommending non SoC treatments will potentially cost them the medical license that came to them at high cost and represents their livelihood.
It can be argued that the SoC maximizes EV and minimizes cost across the population, however, even if that were true on day 1 of the implementation it would quickly degrade as bureaucratic lockin blunted the learning and innovation that doctors have historically used to advance knowledge.
In the medical field, cases of interventions being harmful are not rare, and entire classes of interventions have been reconsidered in light of evidence.
There's several billion medical intervention of all kinds every year, harmful interventions do exist but they are only a very tiny fraction of the total.
Yes, but there are _kinds of_ interventions which turned out to be systematically harmful. For instance, in most developed countries it was a policy to thoroughly wash newborns, until they realized it was harmful.
> Given virtually all interventions have no control how do you know they’re successful?
Nobody ever ran a controlled trial on parachutes, how can we do know that they work?
Controlled trial aren't the alpha and omega of knowledge. They are powerful tools, but they also have a fairly narrow scope of applicability (and they are also a quite recent addition to the collection of scientific tools).
Homeopathy has been widely ridiculed as having no scientific basis, but in terms of outcomes, it might have been better than the heroic interventions of times past, such as bloodletting, mercuric chloride enemas, and whatnot.
Of course, there are no 'crystal signatures' in water and homeopathy has no more scientific basis than astrology does as an actual cure for anything - just a less bad approach compared to the above.
> Rarely do we even consider that the cost of doing something might outweigh the benefits.
Cost vs benefit is the main thing that people consider when making decisions. It's a core decision making framework taught as early as elementary school social studies. Institutional decision making is rife with processes for assessing risks and performing cost-benefit analysis.
> Intervention—by people or governments—should only be used when the benefits visibly outweigh the negatives.
This advice is the decision making equivalent of "just don't write code with bugs." Easily achievable but pretty useless and completely impractical. Real world decision making involves operating under uncertainty at almost all times. If it didn't, the topic wouldn't be worth writing articles about.