• Our chronic disease care ain’t #1 – ctd.

    It’s frustrating, becaus whenever I show data on how we don’t do so well in caring for one disease or another, someone inevitably tries to “gotcha!” me by claiming that there’s a good explanation for it, other than the fact that our system is lacking. Today’s excuse seems to be “the prevalence must be the reason”.

    A number of you have emailed me, or commented, that it may be that the reason there are more amputations in the US is because diabetes is much more common here. More people sick = more amputations. So here’s the chart of amputations again:

    And here’s the prevalence of diabetes in those countries:

    Yes, we have a lot of diabetes in the US, but not so much more than everyone else, and they all do better than us. One in particular on the left (which gets a bad rap on “difficulty to get care”) has nearly the same prevalence in diabetes and less than one third the amputations.

    Someday, you’ll learn to trust me. Or at least, to trust Occam’s Razor.


    ARGH! 🙂

    1) I didn’t have these data “handy”. They take time to take down and chart. But now I’ve done it.

    2) Data on Germany were not available for amputations. Go look at the link to the original data I provided.

    3) Here’s Asthma. First hospital admissions again:


    And then prevalence. Data were not available for the Netherlands, and I used England for the UK. Scotland and Ireland were even worse:


    • Where’s Germany in the 1st graph?

    • OK, great. If you had the incidence numbers handy already, why even bother with the aggregate measure? I can’t speak for anyone else sending an email or commenting, but there was no “gotcha” in my comment. I just want to see meaningful measurements, whatever those measurements might say, including if they make the US look worse. There is an implied accustaion there that those of us who asked about this are not interested in understanding the data but only interested in defending the US health care system.

      Amputations/1000 without adjusting for incidence could potentially be very misleading. To anyone who spends time analyzing data it’s an obvious question to ask when someone shows aggregate numbers. If one of my staff came to me with a metric like that I’d chastise him/her for not making the most obvious of adjustments.

      I think it does a disservice to the debate to assign these motives to people rather than acknoweldge that sometimes people just want to most accurate and meaningful data.

      • I’m not accusing you. If you didn’t do it, don’t take offense. I got plenty of emails accusing me of deliberately hiding the data and trying to skew things just to make the US look bad. Others go through convoluted hoops and explanations to defend the US health care system against any and all “attacks”.

    • And like I said in my other comment, there are more things we need to understand other than incidence. Patient adherence is a big one, especially with something like diabetes. If we’re going to assess the quality of our health care system on metrics like this, we need to account for other factors. This could be help or hurt the US, but either way we need a more rigorous assessment of the data, removing things that are not a direct result of the health care system.

      This goes both ways. A lot of the cancer metrics are misleading in showing that the US does better at cancer survival. To properly understand those metrics you need to understand the differences between what is being measured, e.g. in the US we screen more often and catch things earlier and that makes 5-year survival rates look better than they really are. You alluded to this in the original post. And by that same principle we need more than amputations/1000 or asthma hospitalizations/1000 to really understand how the different health systems contribute to chronic disease outcomes.

    • These metrics ignore the elephant in the room–patient adherence and compliance.

      The underclass in America is both much larger than in England, Germany, etc. and much more heterogeneous. If you compare these outcomes controlling for race and demographics, you will likely find a much different picture (e.g., descendants of Swedish settlers in Minnesota will have similar outcomes for chronic disease as those in Sweden correcting for obesity, etc.). Essentially, all your graphs above ignore the proven impact of social capital, cultural factors, and race on health outcomes. This is in addition ignoring IQ,etc. which many like Sailer would so plays a role.

      In summary, I think the above post is interesting but not really informative on the quality of chronic care in the U.S.

    • One other fallacy in this post and the prior one:

      You assume that preventive care with chronic conditions saves money. The evidence for this is ambiguous and in some case, runs in the opposite direction.

      This is intuitively the case because dying of diseases such as heart disease, lung cancer, obesity, etc. early in life is still much cheaper than living till extreme old age (where different, more expensive threats like dementia exist).

      One can argue that preventive care should be a goal of the system but to assume that it saves money is naive and betrays a lack of understanding of the pathology of disease in humans.

    • V, please provide evidence to support your statements about race and social capital. While it is certainly true that the US has a larger underclass than the other nations in the graphs, I have never seen data to support the claim that this explains all or even half the difference in quality of care between the US and other nations.

      And even if it did, what would that show? That there are inferior groups so poorer outcomes don’t mean the system is deficient? Think hard before you answer.

    • Why is the prevalence of diabetes much lower in the UK than the US?

      The British are almost as fat as Americans.

    • I recreated these graphs with incidence adjusted data.

      Admits/Incidence for asthma shows US as the 3rd worst at 1.10, compared to 1.37 for Denmark and 1.30 for Switzerland (these numbers are mutiplied by 1000). The average is 0.68, with all other countries falling below it and US/DEN/SWI as outliers. Canada is the lowest at 0.13, less than half of the rate of any other country.

      Amputations/incidence for diabetes still shows the US as the worst, though the gap is nowhere near as large as the raw amputations/100,000 suggests. The varying levels of incidence dramatically skews those numbers. The US is at 0.29, compared to 0.27 for Denmark and 0.23 for Norway, average of 0.18. Canada is again an outlier at 0.09, next lowest is 0.14. Once again incidence plays a huge role here, and once again Canada is an extreme outlier. That seems logical to me given the way Canada controls costs in large part by restricting supply.

      So as expected, those original charts are pretty misleading as far as the impact of the health care system vs the impact of disease incidence. The US still compares poorly, but the story is nowhere near as extreme as those charts would suggest.

      • Until you remember that we spend like two to three times as much per person as many of those other countries for health care.

        See, the problem is that so many people spend so much time trying to prove we’re not the worst, instead of recognizing that – for the amount we spend – unless we are the best, it’s a travesty.

        So if you take comfort from the fact that we’re still the WORST in amputations/incidence of diabetes, but not by as much as the chart suggested, after spending more than 16% of our GDP on health care, more power to you.

        P.S. The dreaded “Canada” is kicking our ass in both.

    • We spend way too much money for middling quality, There are some things we do really well, but for the price we don’t get good value. I’ve never said otherwise.

      Of course I don’t take comfort in the fact that we’re still worse in some given health metric. But I do think it’s important to present the facts appropriately, and the fact remains that your original charts were missing a rudimentary but very relevant adjustment, rendering the story they told really misleading. It told a hyperbolic story that exaggerated the reality of us not comparing well on some health metrics despite how much we spend. This being the “evidence based blog” and all, I assumed we’d see the numbers in the proper context.

      Yes, for the money we spend we should be much better (part of the problem is that the marginal health care dollar is not very effective, and there are a lot of people chasing marginal health care dollars with little regard for the quality they’re adding), but no more do I like to see defenders of our system touting misleading cancer survival rates than I like to see our poor performance on quality metrics blown out of proportion, which is what both of those charts did.

    • Hi Aaron,

      Nice post, thank you for putting this mini-series together.

      If you assume that our outcomes in this case are due to our healthcare system, and not primarily because of any other factors, I was hoping you could elaborate on where in our US healthcare system you believe the failing to be (in this specific instance).

      Thank you,