Social Justice, Liberty

Design, Complexity, and Freedom

Design, Autonomy, and Complexity

I just got back from a grand rounds presentation with Neel Shah, a medical researcher and OB/GYN whose current research concerns overtreatment. One of the most important things to come out of his talk was the fact that the single biggest risk factor for cesarean section for a pregnant woman is the facility she walks into. It’s not her health status, it’s not her preferences, it’s not her birth plan, it’s not even her doctor. It’s the hospital. And it turns out there are a variety of really important hospital design elements that trigger overtreatment.  These include distance between patient rooms, distance between workspaces and patient rooms, limited ability to information share, and so on (full report is linked to here). The problem, of course, is that most women don’t know this. Medicine is supposed to take patient choice and patient autonomy seriously and most women don’t want c-sections.  Yet the c-section rate varies across hospitals tenfold in the United States, ranging from 7% to 70% and the variation is higher for low-risk women than it is for high risk women. So why do so many (32% of all American women giving birth) end up with a surgery they don’t want? Again, a big part of the answer (but not the only part, because incentives and risk also matter) is how the environment around women pushes their providers toward high-intensity treatment. What we have then is a kind of non-coercive violation of autonomy. Women aren’t (most of the time, at least) being actually forced into c-sections, they’re being led into c-sections by an environment that makes it very difficult to give birth vaginally.

This got me thinking about other ways in which we can get non-coercive violations of autonomy (someone in the comments can help me unpack that phrase, but I’m going with it for now). What are the other ways our built environments or our political, legal, cultural, and policy environments lead people into violations of their autonomous choices? Or put another way, how do our environments lead us into outcomes that are the exact opposite of what we would choose if unconstrained?

Other design concerns

Radley Balko’s work on the situation in Ferguson does a great job of demonstrating this problem in the political sphere. Ferguson’s political environment of hundreds of small competing jurisdictions meant that there were a million different ways to inadvertently break the law. Once involved with the legal system, there were a million different ways in which the environment prevented one from quickly or cheaply resolving the situation, involving long travel times, limited hours at clerk’s offices and courthouses, and obvious conflicts of interest. Some, but not all, of these design features were intentional. Many were just functions of understaffing and low levels of resources common to small towns.

Another important thinker in this area is Jane Jacobs, whose work I just read for a conference. While she’s often discussed in the context of urban design, it’s probably more accurate to call her an “anti-designer”, at least insofar as she believes large-scale design of urban areas is generally a bad idea. Her work emphasizes the way in which design can unintentionally change people’s behaviors and discourage neighborly cooperation and interaction (among other negative effects). Something as simple as a sidewalk encourages spontaneous interactions between neighbors, increases community vigilance, and provides an easy way to move about the neighborhood. Well-intentioned zoning laws, on the other hand, can strangle diversity, cut neighbors off from one another, and create dangerous dead zones at night.  Neighborhood design constrains our activities in potentially harmful ways.

What all these writings have in common is a general theme of how environments restrict people’s freedom. Of course, restricting freedom is not always a bad thing. The entire history of rule of law, for example, deals with the best way to restrict autonomy in some spheres so that people can better pursue their ends in others. But the violations of autonomy we’re talking about above are different from general restraints on bad behavior. These violations are leading people into worse outcomes than they would choose for themselves. They’re also paired with poor social outcomes including high health care costs, high incarceration rates, zombie neighborhoods. Thus, these kinds of violations are worth paying close attention to. It’s also worth noting that many of these environmental design issues are something like what is being referred to when people talk about “institutional racism” or the various “industrial complexes.”  Though in those analyses there’s even more intentionality than I think often exists. Many of the design problems we face are not those of poor motives (though again, those certainly exist), but of poor design, design that pushes people toward outcomes they would not choose themselves and that are bad for everybody.

Autonomy and complexity

Some of the problem of design has to do with complexity and how people react to complex environments. The problem with hospital birth is that too often the environment is too complex for people to make good decisions. When that happens, as Shah points out, people will revert from high-resistance modes of operation (waiting for a woman to give birth vaginally) to a low-resistance mode of operation (surgery). Rarely are the providers aware that they are doing this. They are reacting to complexity and to the situation in which they find themselves and they’re trying to find a simple way to deal with that complexity. Same thing for the Ferguson residents who don’t show up in court to deal with traffic tickets. Faced with a bewildering bureaucracy, logistical difficulties, and the high likelihood of further entanglement with the law, they choose the path of least resistance and skip their court date, inadvertently triggering an arrest warrant. Jacobs finds a similar pattern in urban design. City planners clearing slums inadvertently prevent people from solving their own housing problems and instead force them into public housing or other kinds of living arrangements they would not choose for themselves.

Noncoercive despotism?

All these examples remind me of Tocqueville’s description of democratic despotism at the end of Democracy in America. While he also ascribes too much intentionality to the “sovereign”, he does seem to understand the problem of systems that violate people’s wills not through coercion but through the gradual destruction of the ability to make good decisions. As he describes it, “it does not break wills, but it softens them, bends them, and directs them; it rarely forces one to act, but it constantly opposes itself to one’s acting; it does not destroy, it prevents things from being born…” Design – not only of our institutions, but of our built environments, of our health care facilities, of our communities – matters for freedom.  The next question is: can we design for freedom?  Can we undo poorly designed environments without falling into the overdesign/planning trap?

Are there other examples of design limiting freedom in harmful ways? Other authors who might be worth checking out?

Published on:
Author: Lauren Hall
  • gisborne

    The design of computer operating systems, networks, other software, and protocols all constrain freedom. The hardcore free software movement (in the most obvious guise of the GNU organization) are essentially digital libertarians.

    • Puppet’s Puppet

      They may constrain “freedom” in some appealing and important sense, and therefore the FSM may in some sense be referred to as “digital libertarians” for their advocacy of such freedom; but such a “digital libertarian” is a libertarian exactly as much as a “free-will libertarian” is a libertarian. They are merely various and distinct technical uses of the same word, and are wholly unrelated to political libertarianism. (Indeed, the FSM is almost certainly explicitly opposed to it, since they recruit state police power in the enforcement of their principles of design “freedom.”)

  • Sean II

    “Medicine is supposed to take patient choice and patient autonomy seriously.”

    You’re funny. Let me translate:

    “Medicine is supposed to take seriously the idea that people with average intelligence and no scientific literacy can, in a few 20 minute conversations, understand things which are considered tough going for the top 3% of people.”

    • Puppet’s Puppet

      It is appropriate for patients to be provided with multiple treatment options, with their respective advantages and disadvantages, in certain cases. In general a consumer shopping in any sector is often heavily dependent on the expert descriptions of the seller to make an informed choice. Medicine is indeed one of the more extreme examples, but there is no reason a patient should need a medical degree to effectively weigh factors like cost, pain, risks, etc. as they are presented to her, instead of having the doctor weigh those desiderata for her according to his chosen presumptions. (Also, I would take issue with the intelligence factor. Medicine requires extremely extensive expertise–as well as the patience and diligence to acquire that expertise–to understand. It does not require more than slightly above average intelligence to understand; how else would doctors themselves be able to understand it?)

      Overall, though, you are dead on. The piece uses a design-jargon concept of “autonomy” that has little to do with any political concept of autonomy that liberals would be concerned with. Indeed, the tone is if anything evocative of a rather creepy and very familiar “soft” paternalism.

    • DBritt

      If that’s true with respect to c-sections then why is the institution the primary risk factor? It should be nearly irrelevant. Clearly your view, while applicable in certain places, does not apply here.

      There are plenty of situations in medicine (obstetrics very much included) where someone has to make decisions on the basis of what risks are involved. It is possible to become informed about the risks of several scenarios in a short conversation, even if you don’t understand the underlying reasons for those risks.

      • Sean II

        There are perfectly good reasons why c-section thresholds should vary from hospital to hospital.

        If you don’t know those reasons and can’t imagine what they might be, stop having opinions about medicine.

        • DBritt

          Feel free to attack other arguments that also aren’t the argument I was making as well.

          • Sean II

            You said place of treatment should be irrelevant.

            You said it here: “If that’s true with respect to c-sections then why is the institution the primary risk factor? It should be nearly irrelevant.”

            You were wrong.

            Your other point – “It is possible to become informed about the risks of several scenarios in a short conversation” – that’s wrong too, by the way. Understanding risk means understanding statistics. Most people can’t.

          • DBritt

            Yeah but that was not remotely my main point. Not only that, your response was a misleading criticism of what I said.

            My main point was that your claim that people making decisions about their own care is somehow technically impossible because they coughs never understand the complexities of medicine is not only wrong general, but especially wrong in this context. People should absolutely be making choices about their own childbirth processes after having been duly informed of the risks.

            The related but less important point about is maybe not worth defending, but I’ll just say that there is a big difference between the observation that institutions are correlated to rates of c-section and the statement that those institutions are *themselves* the primary source of risk.

          • Sean II

            Your main point is even more wrong than your minor one.

            Your wrote this: “People should absolutely be making choices about their own childbirth processes after having been duly informed of the risks.”

            And you’re wrong.

            The idea of “duly informed patients” is a lie that no one really believes in past intern year. It’s a punch line.

            The *minimum* level of statistics needed for someone to be informed about health decisions is NNT vs NNH. Maybe 10% of the population can understand that, with a painstaking explanation, (and no doctor worth having would waste his time thus).

            And when you try to teach a patient NNH, all that happens is: they hear the word harm, and freak the fuck out. No matter how big the number is.

            So you know what smart doctors do instead? They say: “Okay, Myrtle. Your symptoms suggest X. So we’re gonna do Y and Z. Sound good? Okay great let’s get that started…”

            That’s what real-life informed consent looks like. Because that’s about as much of being informed as most patients can handle.

          • DBritt

            Based on my experience having been in the room for two births and with my own family members who are doctors I think you’re quite overconfident in your conclusion that no doctor would every try to inform patients so that they can make their own decision. I can think of at least one conversation I’ve had with my sister (ortho resident) about exactly that. And anyway, you’ve still failed to address the original problem with your statement, which is that in this particular context the *hospitals* are the source of c-section risk. So this is not a case of “doctor knows everything, patient knows nothing.” If they really did then it would be *patients’ condition* that predicted c-section, not the hospital.

            Most cases in medicine you tell Myrtle what the treatment is because there isn’t ambiguity. But medicine itself has demonstrated that c-section is not such a case.

          • Sean II

            1) “Based on my experience having been in the room for two births and with my own family members…”

            Compelling sample you have there.

            2) “Most cases in medicine…there isn’t ambiguity.”

            Exactly wrong. The thing that’s rare in medicine is a clear path from presentation => diagnosis => treatment. Ambiguity is common.

          • DBritt

            1) Lol, okay then how large is your obstetrics sample? Or if you’re an OB how large is your ortho sample? Just being an MD doesn’t magically give you insight into how all other docs operate. And I’m highly skeptical of your claim that advocacy of patient autonomy is a massive conspiratorial deception.

            1b) Besides which, you’re consistently dodging the fact that, given the evidence around c-sections, the approach you advocate is a proven failure.

            2) You are consistently expanding positions I take into straw men that are easier to attack. To the extent that there is ambiguity in *treatment options* after presentation, after diagnosis, after treatment options have already been narrowed down, after by far most of the ambiguity has been eliminated, at that point *the patient should absolutely be involved.*

          • Sean II

            1) “And I’m highly skeptical of your claim that advocacy of patient autonomy is a massive conspiratorial deception”

            Yeah, it’s about as crazy as the idea that education is really just signaling.

            2) “…the evidence around c-sections…”

            Of which there is none. Like I said, that supposedly shocking variance is easily explained.

            3) “You are consistently expanding positions I take into straw men…”

            Sure. This must be why I keep quoting your exact words before each point.

          • DBritt

            I’m probably done here because you are clearly not arguing in good faith. But I’ll give it one last go.

            1) No comment, except to say that if you really think patients should have no say in their care it’s curious that I find you on a libertarian blog.

            2) The point of a paper like the one linked above is that it tries to deconvolve the various factors that contribute to an outcome. Maybe they did a shitty job, but if they didn’t then it *really is the institutions* that are responsible, not the demographics they serve, etc. I would hope you can understand that. Correlation isn’t causation, but here we are talking about institutions as the actual *cause.*

            3) So it’s not possible to quote someone and then interpret their words misleadingly? The first time it happened I described the institution as the “primary risk factor.” That implies that the institution itself is responsible for the risk (see #2). Not the demo it serves, etc. You misinterpreted that at two separate points, so maybe you don’t understand what I’m saying. The second time I referred to ambiguity in possible treatments and you expanded that to include ambiguity at all phases of interaction including diagnosis, which I clearly didn’t intend.

          • Sean II

            “…if you really think patients should have no say in their care it’s curious that I find you on a libertarian blog.”

            That’s a revealing comment. Sounds a bit like: “Why aren’t you letting ideology drive the bus?”

            But of course that’s the point. What patient’s should have has little to do with what they can or do have.

            It would be nice to live in a less cognitively stratified world. It would be nice if everyone understood statistics. It would be nice if patients could be rational…instead of doing that self-destructive thing humans do, getting emotional at precisely the moments when reason is most needed. It would be nice if doctors weren’t busy, and had unlimited time to spend teaching medicine to patients.

            Come to that, it would be nice if there was no such thing as scarcity – of money, time, information, etc.

            Of course, if we’re just wishing anyway, why not simply eliminate disease itself?

            It wouldn’t be that much harder than making all these other dreams come true.

    • Lauren Hall

      You’re intentionally misunderstanding. The vast majority of women do not want a medically unnecessary c-section. Yet 45% of c-sections are, in fact, medically unnecessary. Therefore, doctors are performing major surgery on women that those women do not want for reasons of their own, whether those reasons are convenience, fear of liability, or financial incentives. I’m not advocating that people perform their own neurosurgery. But it seems very clear that patients should have a say in what happens to their bodies, even if their decisions end up being poor decisions, which is not the case here. Patient choice and autonomy matter, particularly in a situation like childbirth where around 85% of births do not (or should not) require medical intervention at all.

      • Sean II

        Oh I understand perfectly well.

        You and some other people think you’ve got a better idea when c-sections are “medically necessary” than the doctors who perform them.

        Like a crazy man, I’m betting the other way.

        In part because I can see at least three crucial factors you and those other people carelessly overlooked.

        • Lauren Hall

          Those statistics (that 45% of c-sections are unnecessary and that 85% of births do not require medical intervention) come from obstetric researchers like Neel Shah and Chris Glantz. These are people with medical degrees whose life work is doing research on pregnant women. If your credentials can match theirs, I would be happy to hear you out. But right now, it seems like the accusations of “staggering ignorance” you so cheerfully throw around fit you better than me. Actually read the obstetrics literature and then we can continue this conversation.

          • Sean II

            Funny how the sources you cite to turn out to be acting as consultants, not scholars.

            I can think of many reasons why a young genius with a massive scoop would forego NEJM to publish under the illustrious banner of…Truven Health or Ariadne Labs.

            Meanwhile ranged against you is the whole of obstetric practice in the world’s best health care system.

            You are making an extraordinary claim – that nearly half of all surgeries performed in a given context are *unnecessary*.

            To back up that claim you would need some extraordinary evidence. You’d need a really compelling story about why all these expert practitioners just keep getting it wrong. So far you’re not even close.

            And I don’t mind admitting that you piss me off. Your pattern of thinking is a close cousin of what one finds in anti-vaxxers. Except the stakes are higher, because pregnancy and child birth are MUCH more dangerous than measles, polio, etc.

            Indeed “natural childbirth” might well be history’s most prolific killer. Hard to beat something that once culled 25% of everyone born, plus also 5% of mothers.

            Medicine ended that horror. The fact that 85% of births now go pretty well is a miracle of applied science, not – as you’ve confused it – a blessing of nature.

          • Adam Bowers

            SeanII, you often make good points but frequently come across as a first-rate asshole. It’s a distraction from otherwise useful debate.

          • Sean II

            I’m sorry for that, but what’s worse:

            1) A couple people who have no idea what they’re talking about accuse half the doctors in a given speciality of grevious malpractice, without proper evidence, without pausing to think of benign explanations for the behavior that puzzles them, without worrying about how implausible it all sounds…


            2) Guy who points this out gets snarky in the process.

            I just don’t see how 2) can be worse than 1).

            Especially not when the ignorance of 1) is what provoked 2).

          • Adam Bowers

            In my opinion, #2 is worse. In most cases, ignorance is the default state of things and can be changed. Doing #2 makes it more difficult than is necessary to change #1. I say this knowing that I am guilty of mountains of #1 and #2.

          • Sean II

            I dispute the idea that tact => persuasion. Historically that claim is not well supported.

            One of my favorites examples is the very rapidly successful gay rights movement.

            Aggressive behavior, slogans like “Get used to it!”, relentless mockery of traditional mores, that’s what finally worked.

            The “tone it down, careful
            they might hear you” crowd, that’s what didn’t.

          • Lauren Hall

            This will be my last comment in response to Sean II, probably for a long time.

            I’ve been doing research on childbirth practices in the U.S. now for 5 years and am in the middle of a book project on the subject. I have been combing the obstetrics literature and doing interviews with maternal-fetal medicine specialists, OBs, GPs who deliver babies, midwives, and lots of people in between. I’ve also read ACOG’s recommendations and the sources that back those recommendations up. I’ve read the obstetrics journals, as well as JAMA, NEJM, the British Medical Journal, Lancet, and many others. My research cites all of these peer-reviewed sources and I’ve had academic obstetricians fact check my chapters to ensure the research is balanced and accurate.

            U.S. maternity care is abysmal and maternal morality rates are increasing as opposed to decreasing (even for low-risk women), in large part as a result of our skyrocketing interventions. This is all widely accepted in the obstetrics community. What is less widely agreed upon is what to do about it. The liability threat is a major part of the problem and most doctors express very little hope that we can turn the tide without massive systemic changes there.

            Your evidence meanwhile, pulled from nowhere: “ranged against you is the whole of obstetric practice in the world’s best health care system.” Every major medical association and research organization in the United States disagrees with your contention. It’s just flat out false. We have the WORST maternal and infant mortality rates in the developed world. Do your research. And do not tell me because you don’t like the links I provided on a BLOG POST that I have not done my research or that I “have no idea what I’m talking about.” If you had shown any interest in engaging in an actual exchange of ideas, I would have been happy to comply. Alas, that was clearly not your motive.

            If there are others in this thread who would like actual evidence-based peer-reviewed literature on the maternity care crisis in the U.S., I would be happy to provide it. Just start another comment thread so I don’t have to come back to this dumpster fire of a conversation.

          • Sean II

            Well, at least you finally mentioned liability. Most accurate thing you’ve said on the subject so far. You should have led with that.

            But then you lapsed right back into foolishness by forgetting to control for demography when comparing outcomes.

            Hey Lauren, guess what happens when you do both – i.e. control for legal environment and demography?

            These things you find so puzzling start to make sense.

          • Adam Bowers

            “I dispute the idea that tact => persuasion.”

            I disagree. The training I’ve taken, the literature I’ve read, and my personal experiences all support being more tactful, not less.

            “Historically that claim is not well supported.”


            “One of my favorite examples is the very rapidly successful gay rights movement. Aggressive behavior meant to shock, slogans like “Get used to it!”, relentless mockery of traditional mores, that’s what finally worked.”

            This is wildly inaccurate. Like most civil rights issues, this issue boiled for many decades. There were thousands of small advances that slowly bring the issue into the orbit of more and more people. Lawsuits wind their through the courts, legislation is passed, people slowly change their minds. You get the picture. Yes, there are protests, and they get the most press, but that’s not what changes hearts and minds.

            Here’s a rough timeline I found of the gay rights movement in the US.


          • Sean II

            Here’s what I see: a lot of nothing, then Stonewall, big turning point. After that, many victories.

            But I don’t think they call it a riot because of how tactful it was.

          • Peter Weller

            -Obvious ad homimen.

            -“Best healthcare system in the world” by what measure? Seems an extraordinary claim.

            Most of the excess c sections are justified on social reasons. On strict medical necessity they are unnecessary.

          • Sean II

            That’s why Lauren should be writing a book on liability reform instead.

            There is nothing to be gained by excoriating doctors, because they didn’t create this situation, and they can’t change it.

            Look, our society has chosen – speaking through its courts, and various other amplifiers – to insist on a trade-off where the small but certain harm of a c-section must be preferred over a list of large but improbable harms to the baby.

            Ironically, from the standpoint of Lauren’s wider argument, our society did this for the well-known reason that people suck at understanding risk. Which is precisely why I scoff at the idea that “let’s empower the innumerate morons” is any kind of solution here.

    • Peter Weller

      A doctor is not a technician: their role is also to advise the patient on the best course of action, given the patient’s values, communicating at a level that person can understand.

      • Sean II

        “…communicating at a level that person can understand.”

        Okay, you tell me:

        How does one communicate NNT vs NNH at a level that an 85 IQ patient can understand?

  • DBritt

    I think science funding is a great example of what you’re talking about. An individual scientist has all sorts of incentives to pursue low-risk, “low-hanging fruit” work that will ensure tenure and continued grant money. There is basically zero incentive to do high risk work. Even agencies that are supposed to fund high-risk, high-reward work tend to fund projects that are basically guaranteed to have *some* output in papers, etc. By design we have trained a couple generations of scientists to go after the easiest/most publishable problems.

    I even recall several years ago a lecturer in mathematics who solved some long-standing problem in a shocking way and the comments from tenured profs were super dismissive. One even said something to the effect that someone of this guy’s skillset isn’t cut out for tenure. You can *succeed* in high-risk work and not even be recognized! Of course that’s because of the low likelihood of a future high-risk success.