Wednesday, April 8, 2015

The Goldilocks Principle

The so-called Goldilocks Principle can be paraphrased as: not too little, not too much, just right.  The health care version of it, though, goes something more like this: if some is good, more must be better.

People have been complaining about overutilization/waste/fraud and abuse for, well, all the time I've been in health care.  A few recent stories reminded me how topical these concerns remain.

One study in Health Affairs, by Ong and Mandl, postulates that false positives in mammograms and breast cancer diagnoses cost the U.S. an estimated $4b each year.  Given that total spending on breast cancer is around $16b, that's a lot of "excess" spending.

Of course, the American Cancer Society was quick to respond.  Richard Wender of ACS told CBS News: "There was no attempt to balance the costs with the benefits."  Of course, as Dr. Mandl pointed out, by definition there are no medical benefits to false positives.

Dr. Wender believes the authors "made a very selective choice of estimates" in coming up with their results, although why they might have chosen ones that would produce artificially higher estimates is unclear.  He went on to assert: "There is no debate about the benefits of mammography."  Dr. Wender must not have been paying attention these past few years, such as to the recent study that concluded they don't really reduce the number of women dying from breast cancer.

Pretending there is no debate about, nor any negative consequences of, testing is simply irrational.

That Health Affairs issue was devoted to cancer.  One study found the greatest mortality improvements between 1995 and 2007 did come in countries that spent the most.  Those results are balanced by another study that attempted to look at costs versus "quality-adjusted" costs, taking into account the value of survival gains.  Their three examples "provide starkly different answers to the question of whether society got what it paid for."

In other words, more spending may be good...but it may not be.

The second article that caught my attention was a piece Jenny Gold did for Kaiser Health News.  The title pretty much says it all: "Ritual, Not Science, Keeps the Annual Physical Alive."

Despite the fact that 92% of Americans think it is important to get an annual physical, Ms. Gold quotes Dr. Ateev Mehrotra of Harvard Medical School, "I would argue that we should move forward with the elimination of the annual physical."  He further noted that randomized trials over the last 30 years have not supported their value, despite costing an estimated $10b annually.

Ms. Gold brings up the same problem Ong and Mandl found; the direct costs are bad enough, but the cost of all the false positive is what really drives up spending.

The third story actually revolves around some advice Mark Cuban, of all people, gave on Twitter last week.   He tweeted:
1)If you can afford to have your blood tested for everything available, do it quarterly so you have a baseline of your own personal health 
2) create your own personal health profile and history. It will help you and create a base of knowledge for your children, their children, etc
3) a big failing of medicine = we wait till we are sick to have our blood tested and compare the results to "comparable demographics" 
 Mark Cuban stirring up controversy is nothing unusual, of course.  His interest in health is new, to me anyway.  His suggestions drew widespread responses, mostly discouraging people from taking his advice.

For example, Charles Ornstein, a reporter for ProPublica, had a lengthy back-and-forth with Cuban.  Cuban maintained it is important to have the data, and Ornstein brought up the issues about false positives (including how annual physicals are overrated).

Saurabh Jha, a radiologist at HUP, had a similar exchange, which he elaborated on at The Health Care Blog.  Dr. Jha uses the thought experiment of Lucky Mark, whose repeated tests do pick up a pancreatic cancer at an early stage and thus save his life.  He contrasts this experience with all the Unlucky Marks, whose preemptive testing produce lots of false positives, which end up killing some of them through complications from biopsies and other interventions.

Dr. Jha claims that the repeated tests are like asking to "prove" that you are healthy.  Mr. Cuban argues that it's always better to have data than to not have data.  They both have points.

I am deeply sympathetic with Mr. Cuban's point-of-view that consumers should have their own data and proactively use it to guide their care, such as in discussions with their physicians.  Still, as Dr. Jha points out, if we're trying to "prove" we're still healthy we could get scans every two weeks and biopsies every other month, but is that really what we want to do?

We talk about data in health care as though medicine was a science, and now worship at the altar of Big Data, but what we don't talk enough about is that much of that data is, for lack of a better word, fuzzy.  We have no shortage of data in health care, but it often doesn't tell us what we want to know.  And we have a disgraceful shortage of data that is either accessible or meaningful to patients.

I've already referenced false positives several times.  If you've read economists like Steven Levitt (Freakonomics) or statisticians like Nate Silver, you've probably seen examples of how easily false positives can impact far more people than true positives.  Given that we are a nation of virtual math illiterates, these examples don't resonate with most of us.

A 2009 study showed that, for a variety of common cancer screenings, in as few as four tests 37% of men and 26% of women will have had false positives.  By 14 tests 60% of men and 49% of women will have had false positives -- resulting in invasive diagnostic procedures for 29% of men and 22% of women.

False positives aren't the exception; they become the norm over time.

If the above examples aren't enough, we could look at our overuse of CT scans or surgeries, but the poster child for "more must be better" has to be antibiotics.  A study last year by Consumers Union found that 97% of physicians were concerned about their overuse.  Eighty-five percent (85%) had a patient(s) with multi-drug resistant bacteria infection within the last year -- and 35% had patients who died or suffered significant complications as a result.  There are a number of other known or suspected consequences of antibiotic overuse as well.

Yes, penicillin was truly a wonder drug, but its success came too quickly and too easily for us to figure out when to most effectively use it.  Like much of what we do in health care.

Too bad our health care system doesn't follow the real Goldilocks Principle.

No comments:

Post a Comment