Comment by irons | Hacker News Reader

Comment by irons | Hacker News Reader

The study found at a general level that people with meningioma were more than twice as likely as people without the brain tumor to have had a bitewing X-ray sometime in their life.

Running a story like this using relative percentages, but no absolute values, is absurd and contemptible.

I'm not finding any absolute values for meningioma rates in any other stories about this particular study, but old pal wikipedia says:

http://en.wikipedia.org/wiki/Meningioma

Many individuals have meningiomas but remain asymptomatic (no symptoms) for their entire life, so the meningiomas are not discovered until after an autopsy. 1-2% of all autopsies reveal meningiomas which were unknown to the individuals during their lifetime, since there were never any symptoms. In the 1970s, tumors causing symptoms were discovered in 2 out of 100,000 people, while tumors discovered without causing symptoms occurred in 5.7 out of 100,000, for a total incidence of 7.7/100,000. With the advent of modern sophisticated imaging systems such as CT scans, the discovery of asymptomatic meningiomas has tripled.

I think I'll worry about something else.

Thanks for this; I was about to say pretty much the same thing.

Gerd Gigerenzer has a nice book about the way risk is communicated.

(http://en.wikipedia.org/wiki/Gerd_Gigerenzer)

(http://www.amazon.co.uk/Reckoning-Risk-Learning-Uncertainty-...)

Otis Brawley's "How We Do Harm" is also good reading.

You don't feel 40 to 90 percent higher risk from bitewings is statistically significant?

If a tumour is only discovered in an autopsy, it's not clinically significant... and the benefits of the x-rays may well outweigh the risk of such a (in this case, usually benign) tumour.

Definition per http://www.ncbi.nlm.nih.gov/books/NBK10468/ :

"Clinically significant: A result that is large enough to affect a patient’s disease state in a manner that is noticeable to the patient and/or caregiver."

But "autopsy" means the person died but not when they died. So an autopsy of a person killed by gunshot in their 20's doesn't alters the average lifetime risk from the xrays.

You added the word "statistically". I know people tend to kvetch about statistical significance almost reflexively, but that's not being done here. I believe the implication here is that it just plain isn't worth worrying about.

Not to mention, cavities and other dental problems pose a measurable threat to your health and longevity (not to mention quality of life), so I would bet a certain amount of radiation risk can actually be offset by the resulting improved dental health.

That's the wrong question. The question is if the X-rays are actually necessary in order to have good dental health, and apparently the answer is no.

Dentists use them because "what's the harm", but now that a harm is demonstrated they should use them only when actually necessary (when "the benefit outweighs the risks").

Um, a dentist's time isn't infinite, and other methods of examination can be more annoying for the patient than an x-ray. I guarantee you that if we don't use x-rays then in practice people are going to suffer more from dental issues.

Statistically significant yes. But 90 percent more of 0.00002 is 0.000038. It would be like avoiding conductive jewelry because you're worried about lightning.

If it has the capacity to produce one kind of cancer then it has the capacity to produce a kind of cancer that was not part of the study.

What is absurd and contemptible is an attitude that "if such and such medical intervention was not proven to cause harm, then we should act as if it does cause no harm", ignoring the fact that the law of unintended consequences trumps what medical science has not yet proven.

http://en.wikipedia.org/wiki/Thalidomide

Reply on news.ycombinator.com