Thursday, January 26, 2006

Banckruptcy -- Medical and journalistic

Posted by Craig Westover | 11:53 AM |  

I am but a humble hobby columnist. Were I a better wordsmith, perhaps if I knew “more stuff,” or maybe the answer is as simple as a drowsy “gatekeeper,” then I too would be able to startle and frighten my readers with such sentences hinting of Bruckheimerian devastation --

“As we learned last fall when Congress tightened the bankruptcy laws for the benefit of credit card companies, about half of the people who file for bankruptcy do so because of illness or medical costs.”
Laura, Laura, Laura (shaking my head sadly).

Let’s take a closer look at that statement from today’s Laura Billings column in the Pioneer Press and see if we might liberate it from the statistical Abu Ghraib where she tortures it to confess the need for universal health care.

The full report on which this statement is based is found here. The Reuters news story found here provides the fodder for a fisk of the little chip of sky that fell on sister Laura.

WASHINGTON - Half of all U.S. bankruptcies are caused by soaring medical bills and most people sent into debt by illness are middle-class workers with health insurance, researchers said on Wednesday.

The study, published in the journal Health Affairs, estimated that medical bankruptcies affect about 2 million Americans every year, if both debtors and their dependents, including about 700,000 children, are counted.
Okay, Let’s apply some logic and calculations to those statistics. Two million is NOT the number of medical bankruptcies. In 2004 there were approximately 1.5 million non-commercial bankruptcy filings in the United States or, according to estimates in this study, about 750,000 medical bankruptcies that AFFECTED two million people, only if one includes dependents of those filing bankruptcy.

That’s an important distinction because it enables us to put the 2 million number in perspective. There are approximately 298 million people in the United States. Thus, medical bankruptcy AFFECTS about 0.7 percent of the population. That’s seven-tenths of one percent of the population.

Certainly bankruptcy is a real problem for that seven-tenths of one percent of the population, but Billings is raising a question of public policy -- specifically she notes that the United States is “the only major industrialized nation that has no universal health care,” fully intending that we, the readers, make a connection here. Looking at the population affected by the problem -- seven-tenths of one percent -- and the proposed solution -- government health care for 298 million people -- is not exactly a split the baby conundrum. The solution significantly exceeds the extent of the problem.

Billings, however, is not the only one with a chunk of sky at her feet.

"Our study is frightening. Unless you're Bill Gates you're just one serious illness away from bankruptcy," said Dr. David Himmelstein, an associate professor of medicine at Harvard Medical School who led the study.
[A gratuitous aside for a friend -- or one misconceived government policy away from bankruptcy.]

Bill Gates is the only person safe from medical bankruptcy? Let’s look at figures from Himmelstein‘s own study.

Among those whose illnesses led to bankruptcy, out-of-pocket costs averaged $11,854 since the start of illness; 75.7 percent had insurance at the onset of illness."


The average bankrupt person surveyed had spent $13,460 on co-payments, deductibles and uncovered services if they had private insurance. People with no insurance spent an average of $10,893 for such out-of-pocket expenses.

Again, not to minimize the problem for the seven-tenths of one percent of the population affected by medical bankruptcy, but there are a lot more people than Bill Gates that can absorb medical costs that are roughly equivalent to the cost of a mid-range mini-van.

Now grant it, these are costs at the time bankruptcy was declared and the debt could run higher. But let’s not forget that one consequence of declaring bankruptcy is protecting assets. In other words, bankruptcy protects individuals from paying debts to others that cannot be paid because of medical costs. The Reuters article quotes George Cauthen, a lawyer at Columbia-based law firm Nelson Mullins Riley & Scarborough LLP.

Cauthen said he was not surprised to hear that so many of the bankrupt people in the study were middle-class.

"Usually people who have something to protect file bankruptcy," he said.
In other words, while certainly no fun, especially in the context of serious medical problems, bankruptcy under the circumstances this study covers does not leave people homeless, destitute, out in the street, with no hope of ever having a roof over their heads again.

Let’s look at yet another angle -- bad debt in general, not just from medical bankruptcy, is a cost of doing business and factored into the prices we all pay for goods. But thinking logically, what percentage of the allocation for bad debt can seven-tenths of one percent of the population generate? Especially if we are talking about individuals declaring bankruptcy over unexpected expenses of $13,000. Again, I’m not minimizing the impact of bankruptcy on these individuals, I’m calling into question the connection between the problem, medically-caused bankruptcy, and the solution, universal health care.

Medical bankruptcies hardly justify the call for Universal Health Care. Some quick calculations indicate that the total cost of the bankruptcy problem for individuals ($13,460 average outlay in medical bankruptcy filings times approximately 750,000 medical bankruptcies) means we're looking at about a $10 billion (corrected 1/270 bankruptcy problem. That’s a far cry from the many (added 1/27) billions of dollars that Universal Health Care would cost, not to mention the inflationary effect on medical costs of “free” service.

As a country, we’d be better off cutting checks to seven-tenths of one percent of the population that are (a big leap here for the sake or argument) victims of market failure than we would be to turn everyone’s health care over to the same government implementing a square-wheel rollout of the prescription drug program.

Just to make sure I was on both the right and correct track, I ran my numbers by King Banaian, chairman of the Economics Department at St. Cloud State. King declared my points “dead on,” and managed with a quick read to scope out other flaws in the study.

It’s not a representative sample.
Sampling strategy. We used cluster sampling to assemble a cohort of households filing for personal bankruptcy in five (of the seventy-seven total) federal judicial districts.12 We collected 250 questionnaires in each district, representative of the proportion of Chapters 7 and 13 filings in that district. These 1,250 cases constitute our “core sample.” For planned studies on housing, we collected identical data from an additional 521 homeowners filing for bankruptcy. We based our analyses on all 1,771 bankruptcies with responses weighted to maintain the representativeness of the sample.
This looks OK, but the clusters were selected only because those were the judges who opened their records; the other 72 did not. The districts they used were in CA, IL, PA, TN, and TX. They also had problems reaching almost half of the 1,771 for follow-up questionnaires, and the questions on their actual illnesses were done on a sub-sample under 400. They’re using that truncated sample to extrapolate the other 72.

It’s post hoc fallacious. (CW -- King, King, King. For those of us whose Latin extends only to having lusted after Deborah Kerr in “Quo Vadis,” a post hoc fallacy is an argument that attempts to establish a causal connection based on a false assumption.)
Under the rubric “Major Medical Bankruptcy” we included debtors who either (1) cited illness or injury as a specific reason for bankruptcy, or (2) reported uncovered medical bills exceeding $1,000 in the past years, or (3) lost at least two weeks of work-related income because of illness/injury, or (4) mortgaged a home to pay medical bills.
Lost two weeks?? Suppose I am out of work for two weeks because I went on a bender and during that time I bet the ponies badly. Is that a “major medical bankruptcy”? I can see (1) and (4), and I might quibble about the low threshold in (2), but there’s no way to determine whether those who missed two weeks of work because of injury or illness went into bankruptcy because of that. That’s 21.3% of the bankruptcies. Maybe the problem is low pay, and maybe it’s absenteeism. Note exhibit 2 in the study. And if medical bankruptcies are caused by unpaid debts at $10,000 or more, why set the limit for major medical bankruptcies at $1,000?
King raises a good question, but perhaps the answer can be found here -- two of the study’s authors are on the board of Physicians for a National Health Program.

The point is statistics that use percentages without providing a baseline that indicate the significance of the percentage in terms of real numbers lead to the kind of jumps in logic that Himmelstein intends gullible journalists and drowsy gatekeepers to swallow, report, and scare the crap out of the rest of us.

Laura, Laura, Laura.

A tip of the Sou’wester to King Banaian for his contributions to this post.

Update: Oops, my "gatekeeper" was drowsy -- 750,000 time $13,460 is $10 billion not $10, million. That's a hefty price tag, to be sure, but still not one that justifies inflicting universal health care on the economy and government mandated health care on every individual in the United States. King Banaian adds some additional thoughts.

Category: Journalism, Health Care, Public Health, Local Politics NationalPolitics