Whose life is it anyway?

Via Roger L. Simon, good news for those suffering from insomnia -- they get to suffer more! That's right, the more sleep you lose, the longer you live:

the refrain that Americans are sleep deprived originates largely from people funded by the drug industry or with financial interests in sleep research clinics.

"They think that scaring people about sleep increases their income," Kripke told LiveScience.

Thanks to the marketing of less addictive drugs directly to consumers, sleeping pills have become a hot commodity, especially in the past five years. People worldwide spent $2 billion on the most popular sleeping pill, Ambien (zolpidem), in 2004, according to the BioMarket, a biotech research company.

Nightly sleeping pill use is about as dangerous as cigarette smoking:
A six-year study Kripke headed up of more than a million adults ages 30 to 102 showed that people who get only 6 to 7 hours a night have a lower death rate than those who get 8 hours of sleep. The risk from taking sleeping pills 30 times or more a month was not much less than the risk of smoking a pack of cigarettes a day, he says.

Those who took sleeping pills nightly had a greater risk of death than those who took them occasionally, but the latter risk was still 10 to 15 percent higher than it was among people who never took sleeping pills. Sleeping pills appear unsafe in any amount, Kripke writes in his online book, "The Dark Side of Sleeping Pills."

"There is really no evidence that the average 8-hour sleeper functions better than the average 6- or 7-hour sleeper," Kripke says, on the basis of his ongoing psychiatric practice with patients along with research, including the large study of a million adults (called the Cancer Prevention Study II).

And he suspects that people who sleep less than average make more money and are more successful.

The Cancer Prevention Study II even showed that people with serious insomnia or who only get 3.5 hours of sleep per night, live longer than people who get more than 7.5 hours.

One of the reasons I don't trust statistics is that they depend on who compiles them. The drug companies come up with statistics showing that "sleep deprivation" kills, then another study comes along and says too much sleep kills.

The New England Journal of Medicine recently reported a study showing that black cigarette smokers developed cancer at a higher rate than whites. Yet the actual numbers were only of interest as statistics, because they weren't that dramatic. Nevertheless, the study was widely written up (this one in the Philadelphia Inquirer was typical). Most of the stories scrubbed the actual numbers, and while I can't prove my suspicions, I suspect that there's a policy decision involved (i.e. they don't want people knowing the actual statistical risks from smoking). Here's the New England Journal of Medicine:

Among participants who smoked no more than 30 cigarettes per day, African Americans and Native Hawaiians had significantly greater risks of lung cancer than did the other groups. Among those who smoked no more than 10 and those who smoked 11 to 20 cigarettes per day, relative risks ranged from 0.21 to 0.39 (P<0.001) among Japanese Americans and Latinos and from 0.45 to 0.57 (P<0.001) among whites, as compared with African Americans. However, at levels exceeding 30 cigarettes per day, these differences were not significant.
Here's a table which MSNBC was honest enough to reprint, showing lung cancer totals for smokers:

LUNG_CANCER_RISK.gif


(Shhhhhh! We wouldn't want the kids know that the chance of getting lung cancer average out to slightly more than one in a thousand. If they thought that 99.8% of smokers don't get cancer, they might wonder what the crackdown on smoking in "public places" was all about. And we just can't have that!) Had it not been for the racial differences which made the story newsworthy, I doubt the total numbers would have appeared in very many places.

I may be old-fashioned, but I don't think that either corporate profits or public policy should influence statistical compilations or news reporting, and I'm always distrustful of statistics.

If I've learned one thing, it's that that if you rely on them, they'll change!

It's always nice to remember that individuals are not statistics.

The statisticians didn't tell me when I was going to be born, so what business have they telling me when I'm going to die?

(Sheesh. You'd almost think the world was being run by fatalists trying to run the world according to some neo-Calvinistic theories of predestination.)

posted by Eric on 03.26.06 at 09:34 PM










Comments

Just a quibble: I think incidence as used in the article means something different from "the number of smokers who will get lung cancer". Incidence typically refers to the number of diagnoses in a given year. Prevalence might refer to the number of people in a given population with the diagnosed condition. Thus, the incidence of death from all causes in a given year might be only 1000 per 100,000, for example.

Incidentally, prevalence might not be appreciably higher than incidence for lung cancer because it isn't a very survivable cancer (wether you get it from smoking or just bad luck).

Of course, since it's a journalist writing this, the chances of them misusing a term are pretty high, so my point could well be moot.

If they correctly quoted incidence rates, your point might more accurately be "if they thought 99.8% of smokers don't get lung cancer in a given year...".

Cheers,

Sebastian

Anonymous   ·  March 27, 2006 4:44 PM

Thanks Sebastian. There's no data provided on the age of the smokers, nor on the length of time they'd been smoking. For that matter, it's unclear whether "smokers" is limited to mean present smokers who developed lung cancer, or includes former smokers.

Eric Scheie   ·  March 27, 2006 5:46 PM

April 2011
Sun Mon Tue Wed Thu Fri Sat
          1 2
3 4 5 6 7 8 9
10 11 12 13 14 15 16
17 18 19 20 21 22 23
24 25 26 27 28 29 30

ANCIENT (AND MODERN)
WORLD-WIDE CALENDAR


Search the Site


E-mail



Classics To Go

Classical Values PDA Link



Archives



Recent Entries



Links



Site Credits