When Good Data Goes Bad

Michael Koumjian, a heart surgeon for nearly three decades, said he considered treating the sickest patients a badge of honor. The San Diego doctor was frequently called upon to operate on those who had multiple illnesses or who’d undergone CPR before arriving at the hospital.
Recently, however, Koumjian received some unwelcome recognition: He was identified in a public database of California heart surgeons as one of seven with a higher-than-average death rate for patients who underwent a common bypass procedure.
“If you are willing to give people a shot and their only chance is surgery, then you are going to have more deaths and be criticized,” said Koumjian, whose risk-adjusted death rate was 7.5 per 100 surgeries in 2014-15. “The surgeons that worry about their stats just don’t take those cases.”
Now, Koumjian said he is reconsidering taking such complicated cases because he can’t afford to continue being labeled a “bad surgeon.”
[…]
Consumers Union, which sponsored the bill that led to the cardiac surgeon reports, supports expanding doctor-specific reporting to include a variety of other procedures — for example, birth outcomes, which could be valuable for expectant parents as they look for a doctor.

What could possibly go wrong?

13 Replies to “When Good Data Goes Bad”

  1. I have a jaded view of the medical profession due to a number of questionable things doctors have done with me in the past, ie drop me as a patient without my knowledge when the missus asked for a 2nd opinion related to her pregnancy.
    but, the guts and cool-headedness it takes to meet the challenge of those seemingly hopeless and intractable lost causes, give them your exceptional a truly best effort in the face of such dreadful odds . . . . .
    I couldn’t. I stuck to mainframe computers and came up with the adage (used on occasion) ‘you really know your stuff when you can hide your mistakes’. I might add this included falsifying the mainframe log file when I decided to try an experiment . . . . I was hacking people’s directories and interoffice mail before the term existed.

  2. So the best doctor has the worst outcome. California you say. That’s the place where every restaurant tells you that eating there will cause cancer and birth defects – more really meaningful information.

  3. It just means that doctors will choose to not take any case with a low probability of success. What is “low” you ask? Why a subjective term of course. For some people, the only acceptable number would be zero. So, surgeons with perfect records become the highest paid. So, to get a raise, every surgeon starts to take only zero risk cases.
    No one has to do your heart surgery… Welcome to the death panel.

  4. It’s that dihydrogen monoxide, man – everybody who lets it enter their body, dies…

  5. Once you are at the point where only zero risk patients can get treatment, there may be a niche for those one or two guys who will still take the tough cases. They will probably want a premium though, and since they are the last resort, you will have to suck that up buttercup.

  6. You guys all realize that doctors avoiding risk by turning down patients is -exactly- in line with socialized medicine, right? Care is expensive, the less that gets provided the better.

  7. Either you police yourselves or we will do it for you. (signed, People of California)
    “It’s a long shot bud. You’re morbidly obese, your liver is shot, your kidneys are failing, you have a cardiac aneurysm and you’re 64 years old. We can fix that aneurysm. Medical care is free in California.” (You might live 6 months more if I’m successful though much of that will be in intensive care.) (I’ll still get paid though.) (I really need that new Tesla).
    “You’re 84 years old. You have colon cancer. We can cut out part of your gut and give you a colostomy bag. If you survive the surgery you might live another 8 months but you’ll need extensive home care and a good supply of bags. Oh, and your diet will be very restricted.” (I’ll still get paid though).

  8. Hello dear admin,
    You have a great blog. With the BloggerContest.com you are a candidate to win with your website or blog, promote your web page, win prizes and promotions. With the support of our sponsors via we organize our blog competition for the first time in 2017. You can learn more about participation from our website.
    You can join, the applications are ongoing.
    Web : http://www.bloggercontest.com
    Mail : contact@bloggercontest.com
    Pbx : +440694802050
    Pousada arden

  9. From the linked article:
    State officials say it has worked here as well: The [death] rate [due to cardiac surgery] declined from 2.91 to 1.97 deaths per 100 surgeries from 2003 to 2014.
    I suppose it’s too much to expect that somebody somewhere asked if the extra 0.94 deaths/100 surgeries might have been due to delayed care because nobody wanted to operate on them for fear of skewing their career stats?

  10. Just another prime example of misleading statistics. Statistics without detail or context are a very dangerous thing that is harming, not helping, our society. Just like “gun death” statistics in the USA … without controlling for “urban” gang slaughter … or hell …just control for Chicago …

  11. the only reason Detoilet Mich lost 1st place is they ran into a shortage of targets when the economy tanked and the huge exodus began. looks good on them.

  12. *
    no good deed, as they say, goes unpunished…
    “The continued criticism and lash out finally broke him down. By 1865,
    Ignaz Semmelweis was suffering from depression, forgetfulness and
    other neural complaints and was eventually committed to an asylum.
    He only lasted there for two weeks and died on August 13, 1865 at the
    age of 47.”

    *

Navigation