It's done with mirrors

 Another article (on Gizmodo this time) about how "experts" are saying that cancer screening doesn't save lives.

For example, a 30 year follow-up to a colon cancer study showed that for every 10,000 patients who were screened for colon cancer, 128 died of the disease, while the rate of death due to colon cancer was 192 out of every 10,000.
 Those numbers indicate that screening is critical. It causes mortality due to colon cancer to plummet.
But another set of numbers is less encouraging. Of the group of unscreened patients, 7111 out of 10,000 had died by the time of the follow-up. Of the screened group 7109 had died—making for a difference of two people.

the article then goes on to blame the "stress" of treatment, the "stress" of false positive results.

Ah, but the dirty little secret is: These people are often older and have comorbidities (lots of things wrong with them)

So they die of something else.

When the anti cholesterol medicines came out, a similar finding was found: If you take it, your heart attack rate is slightly lower, but the death rate was the same.

In one study, the reason for the difference was death by violence/accident/suicide/homicide.

Another study was done, and the death rate was the same due to cancer of the colon deaths.

The Docs joked that since they'd rather die of a heart attack than cancer, why take the medicine?

Similar problem here.

Another dirty little secret is that the numbers sound impressive, but they are low risks for the actual patient.

then there is the "GIGO" problem: garbage in, garbage out.

What’s more, the researchers did a review of 10 meta-analyses of cancer screening studies. These meta-analyses combined the results from several studies to give an overall picture of the research done on screening for different cancers. Only three showed a reduction in mortality due to specific cancers. Not one showed a reduction in overall mortality.

a lot of these studies include those where there are a lot of drop outs, or the population was not representative of normal folks. In others, the numbers weren't significant.

Finally, not mentioned: that many cancers metastasize early, when it is too small to detect from screening. So early diagnosis of a young women with aggressive hormone dependent breast cancer might not make a difference, since it has already spread.

This might not be true of an elderly  person with slow growing cancer, who will die of something else.

And finally a dirty little secret: Malpractice suits are one reason that doctors were forced to order a lot of tests that didn't make a difference.

I was once sued for not ordering a test that wasn't being done because it was being done experimentally 500 miles away, the lawyer argueing that I should have known that the test was available.

Ironically, what saved me in that malpractice suit was sex: The lawyer was caught sleeping with the 16 year old daughter of one of his "expert" witnesses, and the sister of the deceased who knew her sister didn't keep follow up appointments, knew that the husband suing for "wrongful death" had a bimbo on the side, which made his wife ignore the growing cancer as a way of passive suicide...