5 ms·

The study cited by your CNBC article is simply flawed and doesn't measure what innumerate reporters think it measures. To measure the number of bankruptcies du

The study cited by your CNBC article is simply flawed and doesn't measure what innumerate reporters think it measures. To measure the number of bankruptcies due to medical causes, you compute P(bankruptcy|medical cause) - P(bankruptcy|no medical cause). The study measures P(medical cause|bankruptcy). Here, "medical cause" = "spent $1001 or more on medicine out of pocket". The classic example of a medical bankruptcy is Michael Vick, the NFL quarterback who went bankrupt after going to jail for dogfighting (due I'm sure to his medical bills). See also McCardle's deeper debunking of it: http://www.theatlantic.com/business/archive/2009/06/elizabeth-warren-and-the-terrible-horrible-no-good-very-bad-utterly-misleading-bankruptcy-study/18826/ http://www.theatlantic.com/business/archive/2009/06/elizabet...

wk_end 10y ago
The study cited by that CNBC article is not the PNHP study criticized by McCardle [0], although its bankruptcy numbers are derived from it (it also uses numbers from the CDC and the Commonwealth Fund). I encourage you take a look at the original press release [1] (also linked to in the CNBC article), which has any number of other damning statistics besides the bankruptcy numbers that you and McCardle take issue with. Without getting into it here, I don't think McCardle's "debunking" is very persuasive, but you're right that the PNHP study (and thus at least the bankruptcy numbers in the NerdWallet Health report) has serious problems; I hadn't read it, and just as you point out, the PNHP authors conflate numbers to get them as high as possible: around 62%, whereas "when asked about problems that contributed very much or somewhat to their bankruptcy [...] 54.9% cited medical or drug costs". But while 54.9% is lower, it's...not a whole lot lower. Even if the numbers were halved, say, by taking the 29% of people who wholly attributed their bankruptcy to their own medical bills (not their family's), that's nowhere near an order of magnitude different: a million people going through bankruptcy due to medical bills is better than two million but is just as disgraceful, and simply not going to happen in countries with socialized medicine...so I think my original point still stands. [0] http://www.pnhp.org/new_bankruptcy_study/Bankruptcy-2009.pdf http://www.pnhp.org/new_bankruptcy_study/Bankruptcy-2009.pdf [1] https://www.nerdwallet.com/blog/health/managing-medical-bills/nerdwallet-health-study-estimates-56-million-americans-65-struggle-medical-bills-2013/ https://www.nerdwallet.com/blog/health/managing-medical-bill...
- yummyfajitas 10y ago
  The nerdwallet article directly copies their medical bankruptcy numbers from the article McCardle cites. Since they added no value to this, criticizing the original source is adequate. whereas "when asked about problems that contributed very much or somewhat to their bankruptcy [...] 54.9% cited medical or drug costs". But while 54.9% is lower, it's...not a whole lot lower. So what? This number is again not the true figure. The true figure is #bankruptcies in real life - #bankruptcies there would be absent medical causes. The study doesn't compute this. Nearly 100% of people who went bankrupt could cite that they paid at least $1001 in taxes and could attribute their bankruptcy to that in a subjective survey if they wanted. Does this mean that it's a "damning statistic" and "disgraceful" that most people pay $1001 in taxes? Furthermore, bankruptcy for medical reasons (like being unable to work - which the study does include) does happen in countries with socialized medicine. These numbers - and the implications innumerate reporters drew from them - are simply indefensible.
  - wk_end 10y ago
    Yes, the bankruptcy statistics are from the PNHS study. What about the other numbers from the CDC and the Commonwealth Fund? They also damn the US private healthcare system, and so far you've been ignoring them. It's impossible to determine "#bankruptcies there would be absent medical causes", because it's impossible to determine a number that only exists in a counterfactual. Absent this, subjective surveys give us an approximation that's broadly useful; literally anyone can lie or misattribute just about anything on any subjective survey, but that hardly makes them useless. There's no sense in being interested in "the true figure" exactly unless we have reason to believe that it varies dramatically (say, by an order of magnitude) from our best approximations, and outliers like a criminal millionaire football player are hardly that; what matters is an idea of the impact of medical costs on people's financial security, and the study absolutely provides that. Exactly what realistically obtainable data could persuade you that medical bills contribute to financial instability and bankruptcy in the US? The PNHS study showed that among medical bankruptcies, people on average paid ~$17k out-of-pocket on medical bills. Would those people have gone bankrupt even without having to drop that much money on bills? Maybe! It's impossible to know for sure. But here's a study that says 76% of Americans live paycheque to paycheque [0]; what happens when you drop an unexpected $17000 bill on them? You're fighting an uphill battle against common sense, here. And from the other direction: a blog post with some numbers showing that people do indeed go bankrupt in countries with single-payer or socialized medicine [1]. Based on the blog title and the writing, this is, I suppose, is supposed to be a rebuttal against arguments for the public system, but the actual numbers - exactly as imprecise and subjective as the American PNHS numbers, mind - are 5%-15% instead of 30%-60%. That's still too high and still disgraceful, but a dramatic improvement nevertheless. [0] http://money.cnn.com/2013/06/24/pf/emergency-savings/ http://money.cnn.com/2013/06/24/pf/emergency-savings/ [1] http://www.conservativeblog.org/amyridenour/2013/5/28/yes-people-do-go-bankrupt-from-medical-bills-in-single-payer.html http://www.conservativeblog.org/amyridenour/2013/5/28/yes-pe...

wk_end 10y ago

The study cited by that CNBC article is not the PNHP study criticized by McCardle [0], although its bankruptcy numbers are derived from it (it also uses numbers from the CDC and the Commonwealth Fund). I encourage you take a look at the original press release [1] (also linked to in the CNBC article), which has any number of other damning statistics besides the bankruptcy numbers that you and McCardle take issue with. Without getting into it here, I don't think McCardle's "debunking" is very persuasive, but you're right that the PNHP study (and thus at least the bankruptcy numbers in the NerdWallet Health report) has serious problems; I hadn't read it, and just as you point out, the PNHP authors conflate numbers to get them as high as possible: around 62%, whereas "when asked about problems that contributed very much or somewhat to their bankruptcy [...] 54.9% cited medical or drug costs". But while 54.9% is lower, it's...not a whole lot lower. Even if the numbers were halved, say, by taking the 29% of people who wholly attributed their bankruptcy to their own medical bills (not their family's), that's nowhere near an order of magnitude different: a million people going through bankruptcy due to medical bills is better than two million but is just as disgraceful, and simply not going to happen in countries with socialized medicine...so I think my original point still stands. [0] http://www.pnhp.org/new_bankruptcy_study/Bankruptcy-2009.pdf http://www.pnhp.org/new_bankruptcy_study/Bankruptcy-2009.pdf [1] https://www.nerdwallet.com/blog/health/managing-medical-bills/nerdwallet-health-study-estimates-56-million-americans-65-struggle-medical-bills-2013/ https://www.nerdwallet.com/blog/health/managing-medical-bill...

yummyfajitas 10y ago
The nerdwallet article directly copies their medical bankruptcy numbers from the article McCardle cites. Since they added no value to this, criticizing the original source is adequate. whereas "when asked about problems that contributed very much or somewhat to their bankruptcy [...] 54.9% cited medical or drug costs". But while 54.9% is lower, it's...not a whole lot lower. So what? This number is again not the true figure. The true figure is #bankruptcies in real life - #bankruptcies there would be absent medical causes. The study doesn't compute this. Nearly 100% of people who went bankrupt could cite that they paid at least $1001 in taxes and could attribute their bankruptcy to that in a subjective survey if they wanted. Does this mean that it's a "damning statistic" and "disgraceful" that most people pay $1001 in taxes? Furthermore, bankruptcy for medical reasons (like being unable to work - which the study does include) does happen in countries with socialized medicine. These numbers - and the implications innumerate reporters drew from them - are simply indefensible.
- wk_end 10y ago
  Yes, the bankruptcy statistics are from the PNHS study. What about the other numbers from the CDC and the Commonwealth Fund? They also damn the US private healthcare system, and so far you've been ignoring them. It's impossible to determine "#bankruptcies there would be absent medical causes", because it's impossible to determine a number that only exists in a counterfactual. Absent this, subjective surveys give us an approximation that's broadly useful; literally anyone can lie or misattribute just about anything on any subjective survey, but that hardly makes them useless. There's no sense in being interested in "the true figure" exactly unless we have reason to believe that it varies dramatically (say, by an order of magnitude) from our best approximations, and outliers like a criminal millionaire football player are hardly that; what matters is an idea of the impact of medical costs on people's financial security, and the study absolutely provides that. Exactly what realistically obtainable data could persuade you that medical bills contribute to financial instability and bankruptcy in the US? The PNHS study showed that among medical bankruptcies, people on average paid ~$17k out-of-pocket on medical bills. Would those people have gone bankrupt even without having to drop that much money on bills? Maybe! It's impossible to know for sure. But here's a study that says 76% of Americans live paycheque to paycheque [0]; what happens when you drop an unexpected $17000 bill on them? You're fighting an uphill battle against common sense, here. And from the other direction: a blog post with some numbers showing that people do indeed go bankrupt in countries with single-payer or socialized medicine [1]. Based on the blog title and the writing, this is, I suppose, is supposed to be a rebuttal against arguments for the public system, but the actual numbers - exactly as imprecise and subjective as the American PNHS numbers, mind - are 5%-15% instead of 30%-60%. That's still too high and still disgraceful, but a dramatic improvement nevertheless. [0] http://money.cnn.com/2013/06/24/pf/emergency-savings/ http://money.cnn.com/2013/06/24/pf/emergency-savings/ [1] http://www.conservativeblog.org/amyridenour/2013/5/28/yes-people-do-go-bankrupt-from-medical-bills-in-single-payer.html http://www.conservativeblog.org/amyridenour/2013/5/28/yes-pe...