Catheter Infection Law

Consider the vast legal apparatus we maintain to reduce the US’s ~30,000 annual car crash deaths. This includes a vast complex of traffic laws, such as re speeding and stop signs, auto safety rules and tests, and huge police forces and courts devoted to enforcing all this.

Now consider that catheter-related infections at hospitals killl a similar number of folks, and can be almost completely stopped via a simple five step procedure, and yet we have almost no related legal apparatus. No laws require hospitals to follow these procedures, or to pass tests showing they can reliably do so, no police checks to see that they actually do so, nor do courts adjudicate many disputes here. Instead we mainly hope for results from “federally funded programs” that pay hospitals who start programs where they say they will look into the problem. Todays’ Post:

An estimated 80,000 patients per year develop catheter-related bloodstream infections, or CRBSIs — which can occur when tubes that are inserted into a vein to monitor blood flow or deliver medication and nutrients are improperly prepared or left in longer than necessary. About 30,000 patients die as a result, according to the [CDC]. … Yet evidence suggests hospital workers could all but eliminate CRBSIs by following a five-step checklist that is stunningly basic: (1) Wash hands with soap; (2) clean patient’s skin with an effective antiseptic; (3) put sterile drapes over the entire patient; (4) wear a sterile mask, hat, gown and gloves; (5) put a sterile dressing over the catheter site. …

A federally funded program implementing these measures in intensive-care units in Michigan hospitals reduced the incidence of CRBSIs by two-thirds, saving more than 1,500 lives and $200 million in the first 18 months. … “The cost of implementing [such programs] is about $3,000 per infection, while an infection costs between $30,000 to $36,000.”

From the numbers above, the cost to society of each infection is ~3/8 of a life, which is at least a million dollars. So hosptials internalize ~3% of the social cost of such infections. In contrast, car drivers typically internalize at least 30% of the social cost of car accidents they influence.   By this measure, it would make more sense to completely eliminate all traffic laws and trust personal incentives to moderate car accidents than it does to trust hospital financial incentives to cut catheter-related infections. So why do we trust hospitals so much more than car drivers?

Added: Apparently attempts to try these checklists were banned for a while as unethical medical experiments, because careful records were being kept.

GD Star Rating
Tagged as: ,
Trackback URL:
  • Cyan

    Why do you introduce the post by comparing 30,000 traffic deaths under a huge regulatory regime to 30,000 catheter infection deaths under a regime of almost no legal regulations? These numbers aren’t comparable, despite being numerically equal.

    • Jim Stone

      Yes, I’ll second that question. (Love this blog, by the way).

  • A guess: people do think of themselves as driving cars, but they don’t think of themselves as lying in a hospital and receiving a catheter. Hospitalization happens to other people, and is best forgotten as soon as possible.

    This could also explain why people don’t want to invest in health insurance (“I am not going to get sick”) and also why other people do (“When I get sick I want it taken care of without worry”). Hence, because it can explain anything, this theory is useless 🙂

  • Yuck! Keep in mind, catheters are also what they stick up your urinary tract when they need to collect something from your bladder — they’re not just for the bloodstream. And considering the trivial steps needed to prevent these infections, it’s pretty bad that they won’t do this, and even worse that it’s consider unethical to experiment with lists. Good for them that they eventually figured out that it’s not unethical, but what process led to such a judgment in the first place?

    But yeah, Cyan’s right. 30,000 driving fatalities is what you get *with* the big enforcement regime; 30,000 catheter infection deaths is what you get *without*. So presumably, if both had equally stringent enforcement, the outcome would still be more deaths on the road. So while you’ve made the case that this is a low-hanging fruit, the comparison was misleading.

    • Jayson Virissimo

      Why should we assume that the central planning of safety regulations leads to more safety over the long run? Doesn’t that require an argument?

      • Jamie Olson

        I would think that historical evidence based on reduction in injuries and mortalities after substantial regulations were introduced (e.g. seatbelts) would be more than sufficient evidence.

  • Jordan

    The comparison of the raw number of people killed is a flawed one. As you point out, we’ve invested a great deal of resources into reducing deaths due to car accidents and much less so into preventing catheter-related infections. Presumably the number of accident-related deaths would be much higher if we did not take such measures to prevent them.

    • colinnwn

      I wouldn’t presume so much. Money spent in making roadways safer (e.g. TxDot and CalDot along with federal motor regulations) are almost certainly very effective as almost every change is studied for impact either by the gov. organization or safety researchers after.

      I am unaware of any studies that try to demonstrate the effectiveness of traffic patrols and traffic courts over the broad community, compared to the cost of those regulatory structures. This doesn’t mean that traffic patrols at a specific point) like a dangerous intersection) don’t reduce traffic accidents there.

      • Jamie Olson

        I don’t understand. Are you claiming that despite each specific intervention causing a reduction in accidents and mortality, the overall effect might not exist? That does not seem particularly plausible to me. There might be more cost-effective solutions, but I can’t imagine that things like stop signs and traffic lights and well-maintained roads don’t reduce fatalities.

      • colinnwn

        @Jamie Olson

        No. Read my initial comment again. I said specifically I think there are few to no studies that show police and traffic courts are effective at reducing accidents. I’ve never read any. I have read studies comparing physical infrastructure improvements like lights vs. roundabouts, energy deforming devices, and traffic calming techniques.

        People rarely think about the administrative consequences when they are acting unsafely, so I personally don’t think police and traffic courts do much good for road safety.

        I said I do think much of the infrastructure improvements are probably effective. A good example of this is the cable barriers on highways that TxDOT pioneered. They prevent highway crossover accidents that happen head on at high relative speed, by softly bouncing errant cars back in the following flow of traffic that the car was in.

      • I constantly think of the ~$500+ cost of missing a red light by a fraction of a second and the $300+ cost of a speeding ticket if I get ticketed. This follows getting tickets. I am also under the impression over the years that increasing police ticketing in locations where there are high accident rates does reduce the accident rates, but I can’t cite a link for that.

  • Ultimately, this problem is caused by the doctors, not the hospitals, since the central venous catheters that are responsible for most of the fatal infections you describe are almost invariably inserted by doctors rather than nursing personnel. All the preventive measures described relate to things the doctor should do at the time of catheter insertion, or the decisions doctors make about catheter persistence. Here is a documented account of how a doctor’s attitude about this was changed. Perhaps you can substitute this true story about non-handwashing doctors killing patients for the undocumented anecdote you use in your econ classes.

  • We are accepting of paternalism when it applies to the common slob, but we resent it when applied to high-status professionals.

    Also, it is comforting for us to think of medical professionals as faultless, and policing them forces us to confront the discomforting reality that medical care is risky.

    • Grant

      Vinnie, while your comment sounds plausible, I think its much more likely that our government simply isn’t all that interested in saving lives. After all, there are a number of other regulated, lower-status industries (traffic was mentioned) where the marginal cost of saving a life is very low.

      So why don’t health insurers push for these procedures?

      • Jess Riedel

        > So why don’t health insurers push for these procedures?

        This is an excellent question. The only thing I can think of is that in a frictionless market, across the board savings don’t help any insurers (either individually or as a group). But since heath care premiums are somewhat sticky, it seems like the industry could make money by reducing health care costs.

  • Really? The procedure that cannot legally be done by anyone but a professional with at least 4* years of training is the unregulated one? Badly regulated, maybe. But hardly unregulated.

    *I’m assuming RNs can do catheterization; do LPNs do it as well? That would be 2 years.

    This seems more like a story of surprisingly poor practice within medicine in a particular area of low-hanging-fruit benefits, but I believe you’ll find very similar tales within the realm of cars and traffic. Here’s one, in a quote from Tom Vanderbilt’s excellent book “Traffic”:

    For pedestrians, a seemingly trivial variance in a car’s speed can be the difference between life and death. A Florida study [see National Highway Safety Administration, “Literature Review on Vehicle Traffic Speeds and Pedestrian Injuries,” DOT HS 809 021, October 1999] found that a pedestrian struck by a car moving 36 to 45 miles per hour was almost twice as likely to be killed than one struck by a car moving 31 to 35 miles per hour, and almost four times as likely as one struck by a car moving 26 to 30 miles per hour.”

    But returning to the examples above, I’d ask a semi-rhetorical question: is one a regulatory failure, or are they both symptomatic of the poor job regulations do in many places of regulating what really matters?

    There are many places where similar issues crop up in both traffic studies and medicine: medical professionals don’t wash their hands enough (doctors being worse than nurses, but surgical practice being an exception with generally excellent sterility standards and adherence); traffic engineering has not been following current best-practices regarding urban street safety very aggressively; we’re still not really sure if SSRIs work against depression, and the movement to universal self-driving cars (which the DARPA challenges and other tests suggest are nearly practical) is still the cause of cranks, not the goal of a national Moon-Mission agenda to transform the roads.

    I’m saying this as someone who isn’t terribly interested in increasing government regulation. But I would plead for those who would regulate the world to at least do so effectively :).

  • Pingback: Tweets that mention Overcoming Bias : Catheter Infection Law --

  • Douglas Knight

    There’s a common complaint in the comments which I think would be better phrased by comparing car accidents not to catheter infections, but to all hospital morbidity or to all medical problems. People offer panaceas for driving all the time and most of them are nonsense. Even when it should be trivial to verify the claims, it takes years for the knowledge to percolate and years more to turn into regulation. Does moving high school start time from 7:30 to 8:30 really reduce teen accidents by 15%?

    A big difference is that cars and drivers are regulated by fiat, while doctors are regulated by a common law standard of prevailing practices, so the fact that no one else washes their hands is a pretty good excuse. I think that is in agreement with the heart of the post.

  • Alex

    Not only are bad doctors not punished, but following best practice is often illegal. For example, doctors used to lose their medical licenses if they washed their hands before surgery. And similarly, the AMA has long history of lobbying against the pharmaceutical industry having to prove that their medicines are safe and effective.

    • “For example, doctors used to lose their medical licenses if they washed their hands before surgery”
      That doesn’t sound plausible.

      Somewhat related: Is Anarchy on the Streets a Good Thing?

      • Alex

        “That doesn’t sound plausible.”

        I can’t find the exact cite, but I definitely read that after Semmelweis’s book, U.S. doctors were threatened with losing their licenses if they washed their hands. This was basically during the civil war.

      • Jayson Virissimo

        The video claims that there are no rules. I don’t think that is quite right.
        What it should say is that, there is no legislation of rules. The rules are conventional (Nash equilibria).

  • ikmar

    The death rate from car accidents is probably very sensitive with respect to relaxations of the legal apparatus….. as for catheters, it seems that this same relaxation sensitivity is zero.

  • Bruna

    Have you read Better by Atul Gawande? There’s a chapter on how hospital staff’s collective commitment to washing their hands would prevent a lot of illness and death. He also writes about the repeated failure of programs to encourage doctors and nurses to wash their hands to lead to lasting change, and gives a few explanations. The biggest problem seems to be time constraints.

  • Drewfus

    Are there two types of trust – assumed trust and learned trust?

    We learn to trust certain individuals or groups. We assume the trustworthy status of certain other individuals and groups.
    Those assumed to be trusty might be a simple matter of convenience. If you can’t trust doctors to “do the right thing”, who can you trust? We all have to trust someone. IT admins have to be trusted – what other option is there? Also in the case of doctors, surely their relatively high social status plays a part in our implicit trust, which again is an example of having to trust the top of any hierarchy.

    Another reason might be that we have found that certain modes of behavior are good candidates for regulation, others aren’t. So its a learning exercise, like learning what goods and services can be effectively banned.

    I notice you didn’t give an answer to your own question, Robin.

  • Don’t you think it’s a bit misleading to compare car accident death’s and catheter deaths? Surely the loss of QALYs from an average catheter related death is substantially lower than that from a car accident.

    QALYs=Quality Adjusted Life Years in case the abbreviation isn’t well known.

    Besides, it’s just not true that avoiding death is that important of a societal goal. I mean certainly form a utilitarian/happiness maximization standpoint it’s quite strange to think that average (or time integrated) happiness keeps increasing as you extend people’s life span. More likely is that after a certain average lifespan further extension either has no effect on happiness or lowers it (people become more risk averse, greater calcification as the population becomes almost exclusively composed of established figures invested in the current system/theory).

    Given this background it seems totally reasonable for the government to radically prioritize certain kinds of deaths, those that we find particularly scary/sad/tragic or those that preferentially target those with the largest expectation of future societal contribution (young adults who haven’t yet yielded many dividends on society’s investment in them).

  • Helix

    The Post article says the cost of an infection is around $30K. But this is not a cost to the hospital, it’s income. If insurers, especially Medicare, refused to pay the $30K, then it would actually be a cost and the incidence of infection would probably fall.

    • I just returned from my board of directors meeting at a heart hospital in which I am an investor. The risk control manager tells me that the hospital is not reimbursed at all for any patient who develops a methicillin resistant staph aureus (MSRA) infection while in the hospital, regardless of the total bill, and regardless of private or Medicare insurance. The hospital eats the entire expense.

  • Robert Wiblin

    The effectiveness of many kinds of safety regulation we have varies over three orders of magnitude (so I was taught). Good to collect a long list of these regulations (and potential regulations), and work out what makes some appealing and others not. In this case I bet it’s that doctors don’t want to be monitored for fear they’ll be caught screwing up.

    Atul Gawande talked about the issue of lists in medicine at length in the New Yorker a while back:

    He’s always a good read.

  • Michael Price

    Get nursing matrons back in charge instead of administrators. A good matron will make sure that staff understand the negative consequences of bad catheter procedure. Specifically that the matron will tear strips of them, whether or not they have “M.D.” after their name. This sort of thing was a hell of a lot less common when career nurses were in charge of nursing.

  • Great post!

    We certainly don’t do a great job of picking low hanging fruit, which is to say, of regulating from the biggest-bang-for-buck down to smaller bangs-for-bucks. You don’t need to know the cause of a problem to recognize that problems with similar costs (i.e. deaths per year, QALYs adjust this result but don’t change it by a large factor) deserve similar expense at addressing.

  • Kevin

    If it’s a law, how do you enforce it? We have a police force who already monitor roads. We’d have to create another police force for hospitals. The Dept. of Health (DOH) periodically checks restaurants for cleanliness, I don’t know but maybe they already check hospitals too. We could now have them do random checks to monitor the catheter process. Maybe that wouldn’t be too difficult.

    I’d also recommend having hand-washing laws for all citizens. That would save a lot more lives than catheter laws. Then the DOH could walk around all public places and workplaces to monitor whether or not people are washing their hands before a meal, after using the bathroom, etc. This would be a huge success in reducing deaths. Or maybe that would be a little insane…

    I guess the difference would maybe be in the government protecting customers from others who do them harm, as opposed to protecting them from themselves. But then germs do spread and affect others. Just a little food for thought.

    I’m all for regulation, I just want to make it effective. Whether regulation or recommendations or institutions, etc. they need to be both practical and effective. If we had public healthcare, would those same employees follow procedures? Or if we had less healthcare red-tape in other areas (either by institutionalizing it or liberating it), maybe administrators could focus on other outcomes besides just navigating financial/legal nightmares.

  • GU

    Consider the vast legal apparatus we maintain to reduce the US’s ~30,000 annual car crash deaths. This includes a vast complex of traffic laws, such as re speeding and stop signs, auto safety rules and tests, and huge police forces and courts devoted to enforcing all this.

    And we actually take it easy on drivers that kill people. Most state legislatures have invented something called “vehicular manslaughter” which is the same as manslaughter except (1) the perpetrator killed the person with their car, and (2) the punishment is much less than normal manslaughter.

    I see no justification for the differentiation. Manslaughter is basically non-intentional killing due to reckless (a standard higher than negligence) action. Perhaps it is due to the fact that most people could imagine themselves engaging in reckless driving, but not in other similarly reckless behavior.

  • Sorry, but I don’t understand this comparison either.