[Vision2020] A Crash Course in Playing the Numbers

[Art Deco]

  [image: The New York Times] <http://www.nytimes.com/>

------------------------------
January 28, 2013
A Crash Course in Playing the Numbers By ABIGAIL ZUGER, M.D.

The chances are that you turn to this part of the newspaper in search of
some reliable tools for optimizing your health. The chances are that you
periodically visit a doctor for the same reason.

Alas, what you seek cannot be found in either place, not if it’s certitude
you’re after. Whether you are healthy, moribund or traversing the stages of
decrepitude in between, every morsel of medical advice you receive is pure
conjecture — educated guesswork perhaps, but guesswork nonetheless. Your
health care provider and your favorite columnist are both mere croupiers,
enablers for your health
gambling<http://health.nytimes.com/health/guides/disease/pathological-gambling/overview.html?inline=nyt-classifier>habit.

Staying well is all about probability and risk. So is the interpretation of
medical tests, and so are all treatments for all illnesses, dire and
trivial alike. Health has nothing in common with the laws of physics and
everything in common with lottery cards, mutual funds and tomorrow’s
weather forecast.

Thus, no matter how many vitamin-based, colon-cleansing, fat-busting diet
and exercise books show up in 2013, the most important health book of the
year is likely to remain Charles Wheelan’s sparkling and intensely readable
“Naked Statistics,” even though it’s not primarily about health.

A professor of public policy and economics at Dartmouth, Mr. Wheelan earned
journalism credentials writing for The Economist and has previously drawn
on both careers to produce “Naked
Economics”<http://6thfloor.blogs.nytimes.com/2012/09/28/how-to-read-like-an-economics-columnist/>(2002),
an accessible guide for the lay reader. “Naked Statistics” is
similar, a riff on basic statistics that is neither textbook nor essay but
a happy amalgam of the two.

It is not the place to learn for the first time about medians and means,
but definitely the place to remember what you were once supposed to know
about these and other key concepts — and, more important, why you were
supposed to know them.

And that means you. While a great measure of the book’s appeal comes from
Mr. Wheelan’s fluent style — a natural comedian, he is truly the Dave Barry
of the coin toss set — the rest comes from his multiple real world examples
illustrating exactly why even the most reluctant mathophobe is well advised
to achieve a personal understanding of the statistical underpinnings of
life, whether that individual is watching football on the couch, picking a
school for the children or jiggling anxiously in a hospital admitting
office.

Are you a fan of those handy ranking systems based on performance data,
guaranteed to steer you to the best surgeons in town? If so, you are up to
your armpits in descriptive statistics, and Mr. Wheelan has some advice for
you: beware. The easiest way for doctors to game those numbers is by
avoiding the sickest patients.

Has a diagnosis of
cancer<http://health.nytimes.com/health/guides/disease/cancer/overview.html?inline=nyt-classifier>in
the family sent you fishing around for information on average
survival?
You would be well advised to figure out if it is mean or median survival
you’re looking at.

Mr. Wheelan pays homage to Stephen Jay Gould’s classic essay on prognosis, “The
Median Isn’t the Message,”
<http://www.cancerguide.org/median_not_msg.html>in which Dr. Gould
uses his own cancer diagnosis to illustrate how
misleading median-survival data can be. (Told he had a malignancy with a
median survival of eight months, Dr. Gould died of an unrelated cancer two
decades later.)

Are you impressed with studies showing that people who take Vitamin X or
perform Exercise Y live longer? Remember, correlation does not imply
causation. Do you obsess over studies claiming to show that various dietary
patterns cause cancer? In fact, Mr. Wheelan points out, this kind of
research examines not so much how diet affects the likelihood of cancer as
how getting cancer affects people’s memory of what they used to eat.

And while we’re talking about bias, let’s not forget publication bias:
studies that show a drug works get published, but those showing a drug does
nothing tend to disappear.

Do you think you need a yearly PET scan to check for rare diseases? Mr.
Wheelan has you contemplate the spam filter on your e-mail program. Set too
low, it feeds a lot of garbage into your in-box; set too high, it will make
you miss the important messages. The same trade-off applies to the
interpretation of medical tests. Unproven disease screens are likely to do
little but feed lots of costly, anxiety-producing garbage into your medical
record.

Mr. Wheelan deftly whisks you through mean distributions and standard
errors, and before you know it you have plunged into the dense thickets of
regression analysis, a tool for separating out many different strands of
cause and effect. With the aid of a long shaggy-dog story involving a group
of whippet-thin marathon runners and a handful of considerably plumper
delegates to an International Festival of Sausage, you will undoubtedly get
closer than ever to understanding the powerful technique he calls “the
hydrogen bomb of statistics.”

Regression analysis is the engine that drives the giant randomized
controlled studies that increasingly inform every medical decision anyone
makes for you these days. It can be a force for good or for ill, and Mr.
Wheelan aptly demonstrates how it can succeed and how it can fail (note,
for instance, the recent gigantic flip-flops on
estrogen<http://topics.nytimes.com/top/news/health/diseasesconditionsandhealthtopics/estrogen/index.html?inline=nyt-classifier>prescribing).

If you want to eat sausage and survive, you should know what goes on in the
factory. That dictum — one of only a few certainties in an uncertain world
— most definitely applies to the statistical sausage factory where medical
data is ground into advice. Mr. Wheelan has propped the factory gates wide
open. Take his tour.


-- 
Art Deco (Wayne A. Fox)
art.deco.studios at gmail.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.fsr.com/pipermail/vision2020/attachments/20130129/9e5d205a/attachment-0001.html>