spells and ringing in his ears. Audiology tests
showed acute hearing loss; an MRI showed a discrete mass tucked into a corner of his skull. An
acoustic neuroma—benign, except it required
neurosurgery. “I knew it,” he said when I explained
all this. He was looking, well, annoyed.
“You did?” I said, a bit surprised.
“Yeah. I’m on my cellphone all the time.” He
looked up at me, perhaps seeing my puzzled
expression. “That has to be the cause,” he said.
Sometimes you need an explanation more
than you need the truth.
HOW DOES ANYBODY KNOW WHAT’S
hazardous to our health? We don’t, at least not
with any certainty. Medical research, especially
when it involves risks in the environment, has
a very hard time pinning down cause and effect.
You’d think the question would be simple: “If
you’ve been exposed to this, does it increase
your risk of developing that?” But we don’t get
to ask the question that way. Experiments that
expose people to risk are generally unethical,
so we have to go at it in reverse: “You have that;
now what on earth caused it?” The problem
with this approach is that it’s often hard to pick
out the actual culprit from the much larger
crowd of innocent bystanders.
A classic instance of this kind of confusion
was a series of studies that seemed to find that
people with oral cancer were more likely to use
mouthwash. This association was translated
into “mouthwash causes cancer.” But association isn’t the same as cause and effect; often,
studies that find such associations omit some
crucial intermediate step that the researchers
This Is Going to
Be the Death of You
OR IS IT? STATS HAVE A WAY OF MAKING EVEN
THE MOST UNLIKELY DEMISE SEEM POSSIBLE.
Heart disease 1 in every 6 people
Cancer 1 in 7
Stroke 1 in 29
Motor vehicle accident 1 in 98
Suicide 1 in 109
Fall 1 in 163
Accidental drowning 1 in 1, 103
Fire 1 in 1,344
Air transportation accident 1 in 7,178
Cataclysmic storm 1 in 29,196
Bee, hornet, or wasp sting 1 in 79,842
Earthquake 1 in 97,807
Legal execution 1 in 111,779
Lightning 1 in 134,906
Dog bite 1 in 144,899
Flood 1 in 558,896
Source: Injury Facts 2012 Edition (National Safety Council)
just hadn’t considered but that can turn cause
and effect 180 degrees. What if people with cancer develop bad breath? (They can.) In that
case, oral cancer becomes a risk factor for using
mouthwash—not a very useful finding.
When it comes to human health risks, causality is hard to prove. The best we can do is keep
accumulating associations. If after a while the
“weight of evidence” keeps pointing to the same
culprit, then we may have uncovered the actual
cause. But as long as all we have is the weight of
evidence, there will always be a chance that
something random will tip the scale. That’s why
much of what you read about environmental
health risks so rarely pans out: What seems
dangerous is really just something an investigator found at the scene of the crime, and more
often than not it’s simply an innocent bystander.
The environment is, after all, a very big place.
One of the most difficult aspects of finding
the real killer has to do with the role of random
chance, and the way it can play into the well-known human tendency to find patterns where
none exist. You can demonstrate the latter for
yourself right now. Flip this page, hold the magazine at arm’s length, and squint at the collection of random lines in the box at the bottom:
See how the spaces between the lines jump out
at you? It doesn’t even need to be lines. Stare at
any page of print this way for more than a few
seconds and the spaces between the words will
start forming little patterns—chains, zigzags,
stars, you name it. Those patterns result from
your brain doing what human brains do, which
is impose some kind of order on the world,
even if your brain has to invent it. You can do
the same thing with white noise, stars in the
sky, or the Mayan calendar.
Randomness has a peculiar feature that plays
into this aspect of human perception, and it’s
another concept you can demonstrate for yourself. Take a handful of loose change and let it fall
to your desktop: Study the arrangement of
coins and you’ll find that they are not evenly distributed. Some of them cluster together. If you
think about it, this makes perfect sense, because
one way of looking at the randomness of those
coins is that the distance between them varies
randomly. The average distance between pennies may be an inch, but a few will be farther
apart—and a few will be closer together, giving
you what looks like a cluster.
This phenomenon gives researchers and
other people fits, especially when, say, three
children in one neighborhood all develop leukemia the same year. People immediately start
looking for a cause—it’s in the water, it’s in the
soil, it’s coming from that power line. All of
which may be so, but it’s also entirely possible
that you’ve just encountered a tragic example of
random clustering. And it’s next to impossible
to prove or disprove any one of those theories.
It’s not that we’re entirely helpless to know
what’s dangerous: It’s just that this knowledge
is accessible only once you think you’ve identi-
fied a specific risk and then do something to
reduce it. There’s no ethical barrier to studying
risk reduction, such as lowering your blood
pressure or quitting smoking. If such an experi-
ment reduces the incidence of disease, it can
tell you something about what causes the dis-
ease. But still not as much as you’d think.
HE’S AN ABSURDLY HEALTHY GUY IN HIS
30s who was in my office just last week for his flu
shot. Now he’s back. As soon as I open the door I
can see he’s in a panic.
“I’m numb,” he says, holding out his hands.
His story explains both the numbness and
the panic. After seeing me the week before, he’d
mentioned to a friend that he had received his
flu shot. The friend had then told him this was a
bad idea—didn’t he know the flu shot caused
He didn’t start to worry about it until he went
online to check out his friend’s claim. Turned out,
there were thousands of websites that seemed to
confirm the danger, many of them offering up
agonizing personal stories about the horrors of
Guillain-Barré. And sure enough, not long after
that he began to notice pins-and-needles sensations in his hands. Which is, of course, one of the
symptoms of Guillain-Barré syndrome, a rare
neurologic condition marked by a creeping paralysis that climbs up the victim’s arms and legs. In
very severe cases, it can be fatal.
I know the statistics: The average individual’s
chance of dying of Guillain-Barré in any given
year is about 1 in 10 million; death from flu is a
thousand times Continued on page 174