The Importance of Outcomes

Last week, Bill Moyers interviewed Jim Yong Kim, a distinguished medical professor and leader of nonprofit organizations and the new president of Dartmouth College. A lot of Kim’s work is dedicated to improving health in the developing world, so you might think he is some sort of soft-hearted lefty. But one of his main points about our health care problems was that our health care delivery system is not sufficiently tough-minded and calculating, and that health care providers can learn a lot from the business world. For example:

“JIM YONG KIM: So a patient comes into the hospital. There’s a judgment made the minute that patient walks into the emergency room about how sick that person is. And then there are relays of information from the triage nurse to the physician, from the physician to the other physician, who comes on the shift.

“From them to the ward team, that takes over that patient. There’s so many just transfers of information. You know, we haven’t looked at that transfer of information the way that, for example, Southwest Airlines has. Apparently they do it better than any other company in the world.

“BILL MOYERS: Computers?

“DR. JIM YONG KIM: No, they have taken seriously the human science of how you transfer simple information from one person to the next. And in medical school, and in the hospitals that I’ve worked in, we’ve done it ad hoc. Sometimes we do it well. Sometimes we don’t do it well. But what we know is that transfer of information is critical. Now to me, again, that’s the rocket science. That’s the human rocket science of how you make health care systems work well.”

In other words, to take just this example, we know that the way information is transmitted can have a major impact on the quality of health care, yet most people in the field neither study it or pay attention to it. Why is this? Kim points to a few reasons.

First, there is the myth of technology: “I think for many, many years, we’ve been working under the fantasy that if we come up with new drugs and new treatments, we’re done.”

Second, there is the peculiar nature of the health care delivery market: “What we’ve learned about organizations is that it is very difficult to get a complex organization, a group of people, to work consistently toward a goal. In the business world, if you don’t do it well, the market gets rid of you. You go out of business. But many hospitals executing very poorly persist for a very, very long time.”

And third, there is a cultural aspect: “I’ve noticed over the years that when it comes to our most cherished social goals, not only do we tolerate poor execution, sometimes we celebrate poor execution. Sometimes it’s part of the culture. You know, these folks are trying to solve this terrible problem. They can’t keep their books straight. They really don’t know what they’re getting. They don’t measure anything. But they’re on the right side, so that’s okay.”

Kim’s solution is something that the Dartmouth Institute has been doing: study outcomes, and study them scientifically, so you can identify what inputs lead to what outputs. I don’t think this is a sufficient solution, because even if you know exactly how to deliver high-quality, low-cost health care, if physicians are still being paid more to conduct unnecessary tests, they will conduct those tests. But if you don’t know how to deliver high-quality, low-cost health care, then you are just wandering around in the dark.

By James Kwak

30 responses to “The Importance of Outcomes

  1. My experiences in our local hospital’s outpatient services are excellent; the ER is a complete disaster where you literally take your life into your own hands just entering there. One would think running the ER might be considered slightly more important, but I suspect it is not a good “profit center”.

  2. One of the first rules of information transmission is that the fewer links, the better. Huge efficiencies could be gained by having competent doctors examine people entering emergency rooms when they arrive, followed by immediate referral to specialists when necessary. It would be even better if their complete medical history was immediately available as well in electronic form, and if the doctor performing the examination had immediate access to most major testing equipment.

    if you don’t know how to deliver high-quality, low-cost health care, then you are just wandering around in the dark.

    CBS from the West has made some excellent comments about this in earlier posts. There seem to be political barriers to studying how to do this.

  3. The fact that everyone must be “triaged” and then wait anywhere from 10 minutes to 10 hours in a true emergency means that staffing is insufficient.

  4. The problem is in the definition of “unnecessary”. You see me for headaches-I could:
    1) tell you it is probably migraine, no further testing- probability that I am correct ~95%
    2) tell you there is a small chance you have a brain tumor and get an MRI- probability of brain tumor as the cause of headaches <1% ( these probabilities would vary depending on the specific case. Approximate numbers are used for illustrative purposes.)

    From a population perspective #1 is the most cost-effective option. From a population-based outcomes perspective #1 is even better, since many of the brain tumors won't have effective treatments anyway. From an individual perspective though, finite costs for incremental benefits are not always "unnecessary". Even when the benefits are very small.

  5. His comments about information transfer were very interesting. Healthcare is filled with high volume of very small details, possibly inconsequential at the moment of the info transaction, but adds up over time (in volume and significance). So for example, a handwritten paper script can be quicker and more efficient at the moment of making that decision, rather than fiddling with a database that involves scrolling and verification of patient id’s, etc. Interacting with technology can increase cognitive overload and be added burden in an already fast-paced, demanding environment. So the frustration factor of transmitting a high volume of small bits of information needs to be managed or eased. Some of this is improving the technology but that’s only part of it. Fostering dependency or over-reliance on automation is also undesirable. Decisions need to be active and within thoughtful context but technology can foster passivity/mindless compliance, especially because it often only works properly under a certain set of conditions. So if you have an exception to the usual routine, that can throw the whole thing and require a cascade of hassle-filled interventions or work-arounds. I’m very curious about what he’s going to come up with.

  6. You’re right of course (I’m not sure I will ever post that again about your stuff, so take a bow) but Dr. Johnson is also right about the incentive issues. I don’t want to say competent, but certainly experienced doctors don’t work the ER, they “graduated” out of that tedium and work in private practice or specialties. Only the greenest Doctors have to work the ER, it’s great experience for them, but they soon leave. Because insurance doesn’t cover many ER patients, we don’t pay the doctors enough to stay at that job – meanwhile the dermatologist that spends 90% of his / her day examining moles and while certainly saving lives never sees anyone in imminent danger, makes mid-six digits.

  7. “But if you don’t know how to deliver high-quality, low-cost health care, then you are just wandering around in the dark.”

    Well said. If you don’t know how to deliver any product or service at a high-quality, find something else to do

  8. Yeah, I was kind of hoping this distortion where people can’t get routine care, but have to wait until they are practically dead to get care in an ER will be solved soon.

  9. There are certainly no engineering barriers to having the same equipment that takes the readings feed the results into a patient’s records. From there it can be processed into whatever type of helpful format is needed by the doctor.

    There are also no significant engineering barriers to making reliable enough equipment. And there is no reason why a system couldn’t be designed that obviated any objections, and that incorporated some hand written notes (you’d be surprised at the number of harmful mistakes caused by bad handwriting), and an appropriate amount of oral communication.

  10. The history of medicine would have been very different if Doctors (and Barbers!) down through the ages took your advice.

  11. Also mk, I’m sure you have the good sense not to agree with me, but to agree with facts because they are facts.
    Most of what I say here is purely polemical in an effort to jolt people out of intellectual inertia.

  12. “their complete medical history was immediately available as well in electronic form,”
    I can’t stop being terribly frightened by this new feature being the next improvement of health care

  13. I spent the bulk of my life designing and building single family homes. The most profitable aspect of my trade is quality. Customers will pay 20-30% more for quality. What’s more, quality, when the team has a good communication language, is less expensive to deliver than is shoddy work.

    The key to getting more money for your work that costs less to deliver is the control language that the community of contributors uses to deliver a complex product. Double-entry bookkeeping is that control language, and has been for 670 years.

    I learned the industrial data model over 50 years ago at a GE plant that built steam turbines, generators, and ship propulsion systems. Very complex products that are huge in size, heavy in weight, and must be manufactured with deadly precision. Their data model was run on the double-entry bookkeeping framework of rules.

    In 1979, when owning ones own computer was made practical, I had returned from industry and gone back into building design and construction. I assumed that bookkeeping was bookkeeping and so I bought a package with my computer. However, what came with the package was a shadow of the system industry had used before computers. Bookkeeping software is worse today than it was then.

    I have studied the problem of getting the bookkeeping correct for 30 years. I have built a prototype to prove that it can be programmed into a computer. There is no substitute for what this ancient language is designed to accomplish. What’s more, when it is computerized it is magnitudes more powerful than it ever was before.

    Healthcare needs a full double-entry for the same reason I needed it in building better houses. The sad part of this story is that the bookkeeping framework itself is lost on today’s culture. A person coming out of the best business schools does not have a clue how the framework manages data. It is clear to me that neither James nor Simon understand what our culture has lost in losing this essential business discipline.

  14. Assuming it’s only you, the doctors who are specifically treating you, the billing departments, and epidemiologists (with redacted personal data), who get to see them, what are specifically worried about?

  15. During their time they may have been performing at the highest level possible.

  16. Yakkis
    that data are completely safe and private is always promised but as I am from a country that traditionally has registered peoples’ place of residence and made “good” use of those data from 1933 on I can’t help being afraid of any data collections, especially electronic ones because they have the big “advantage” over all previous ones that you may connect data from different sources on your screen quite easily serving beneficial as well as sinister purposes.

    My favourite example is: how are you ever going to hide a second person in your apartment when a computer is quietly and effortlessly monitoring your electricity and water and sounding the alarm if you use more assuming there is a leakage.

    As to health: with the old system I can easily figure out sending somebody to a doctor for treatment having her pose as me. How am I going to pull that off with a central recording and my card all of a sudden showing a disease I plainly do not have? Resistance to government is not only possessing guns it is also preserving the ability to fly under the radar. And yes I am aware that the smart cards have the potential of preventing quite some misery. But I personally living in this country and coming from a bureaucratic profession vote for preserving the misery. But I know it’s no use and I have already sent in my photograph as demanded as demanded by the insurer …

  17. Why do we expect that hospitals will run better than your average Burger King? Both have the same perverse incentives.

  18. Take a look at Sara Robinson’s account of an ER visit in Canada. My impression is that most of this is, in fact, already in progress in countries with modern health care system.

  19. If that is your desire, then I must say you mostly accomplish the task.

  20. Americans really need to get out more, and see how the rest of the world is leaving them in the dust while they debate whether something is “socialist” or not.

  21. First of all – if a doctor has to see an inordinate amount of patients – and hands them off to someone else – it doesn’t matter how good the information transfer is – you still are churning more than processing information.

    There may be some incremental benefit – but this looks like silver bullet syndrome.

    The problem is that the information hand off has to happen in the first place – if you have an assembly line operation – which is great for insurance companies to have lots of check points (why there are so many tests, for example) and data, but terrible for doctors to understand the context of a patient’s condition, then efficient information hand off is a band aid solution to a much bigger problem.

  22. This is a big topic, so big it deserves a whole book. Wait! Someone wrote a book about it …

    The Innovator’s Prescription by Clayton M. Christensen (The Innovator’s Dilemma), Jerome H. Grossman MD and Jason Hwang MD.

    It includes pretty much what was said in this post (in greater detail) plus lots more.

  23. Remove the noun “healthcare” or “hospitals” from this discussion and replace it with any large-scale organization. It all still holds true.

    The comment about sometimes celebrating poor execution at an institutional level is dead-on. It’s one of the ways that large organizations (corporations, universities, etc.) maintain themselves.

  24. CBS from the West

    Suppose you are traveling somewhere in the Bible belt, and you fall and break your arm. You go to the ER to have it assessed and get the fracture set. Do you really want everybody in that ER to know that you are gay, or that you once had an abortion, or something like that? And it’s tricky, because I can’t see any obvious way to restrict the information made available to just “what’s needed for current treatment,” because that is often an extremely vague qualifier.

    My opinion is that on balance we’re better off with fully available medical records, but I can certainly understand why people have qualms about it.

  25. When my father was ill we made it a point to be available during nursing changes to make sure that information buried deep within a 2 inch thick file made it to the top and assumptions weren’t made by simply looking at the person in the bed. We discovered the need for this after the first exchange out of critical care. The nurse assumed by age and appearance that this was an end of life case instead of a sudden onset and recoverable event, naturally making it appear that attention would be better spent on other patients. Not an indictment, simply a natural human tendency – current orders are all there is time for usually and so much information is lost every shift.

  26. “Transfer of information” (when I was a little kid we called it “communication”) is a problem in health care, but a very very small part in the ballooning costs of health care. Let’s pick our battles. After we stop the parasitical insurance corporations from bilking us, then we can fine tune how people communicate in E.R.

  27. How to deliver high-quality, low-cost health care:

    This is where the decades of reading economics combined with the years of dealing with health insurance to help me produce some very specific market and market-like mechanisms designed to pay for outcomes over time, and handle various complexities:

    http://findingourdream.blogspot.com/2009/06/new-way-to-hold-down-health-care-costs.html

  28. Not just the insurance companies. A whole new class of doctors that hire nurse practitioners and “expert billing companies” and go hog wild. Bill the local governments for full time hospital employment and open up a few clinics elsewhere. Hey, who’s checking? We had an entire monster hospital company (Columbia/HCA) which took billing to an almost religious level. Wish I had numbers, but just based on personal experience and reading the news, the fraud must amount to many billions of dollars.

    *Not* a slam at the profession. The professionals surely have a feel for how bad it is. Why don’t they blow the whistle? Or do the bad guys and the good guys in medicine just operate in parallel, non-intersecting worlds?

  29. “Or do the bad guys and the good guys in medicine just operate in parallel, non-intersecting worlds?”

    if you happen to have met a bad guy before you have found out that he is one you quite often form the opinion that he is more than usual likeable and decent – it takes quite an effort to “harmonize” these two pictures of the same person – the facts and the gut reaction you once had

    maybe this probably quite common experience is a great help to those wanting to seem honest crooks

  30. Hospitals are organizations (how’s that for insight). Meaning they are amenable to the same improvement regimens that the Dr. Jim Long Kim mentions using Southwest as an example. Goldratt’s The Goal and his theory of constraints seems esp appropriate. but organizations change on their own terms and the so-called cultural/behavioral variables are the toughest to change. that’s why “computers” is anther dumb Moyers comment. Many doctors get their jollies by subverting new IT systems. They might however adpot them if they play a role in designing the system. some enlightened facilities actually manage change well–treating organizations as human/technical systems. A lot of modern medicine uses essentially a Taylorian management system–break things into small tasks, specialize every worker, pass things along the assembly line.
    That can work in very simple tasks but in the complexity of some medicine that is a bad joke.

    Let’s hope Dr. Jim Kim can have an impact. He seems to understand the issues more deeply than most.