Re: Delayed Interventions


"Leonard Evens" <len@xxxxxxxxxxxxxxxxxxxxx> wrote in message
news:uJqdnQoLPYjdu3vZnZ2dnUVZ_q2dnZ2d@xxxxxxxxxxxxxx
Ed Friedman wrote:
George Conklin wrote:

With news reports out that the Journal of the National Cancer
Institute has
found that delayed intervention for prostate cancer is better than
overtreatment, links to the August 16 journal site require you to PAY
for
the article. Has anyone posting here paid or has a university
subscription
to the journal? What about some details? News reports are pretty bad
on
the details.


George,

I have the complete article. What questions would you like answered?

The thing that struck me in the article is that they cited another
article which showed a significant drop in latent cases of prostate
cancer (PCa) discovered at autopsy since the introduction of PSA
testing. The implication is that less people are dying with latent PCa
because they are receiving early localized treatment. This would mean
that people who never would have died of PCa are going to show up as
being "cured" by early localized treatment.

Interesting. If you treat everyone who shows clinical evidence of
prostate cancer, you will remove a certain number of men from the pool
of men who otherwise would show cancer on autopsy. In principle that
would allow you to estimate the percentage of cases which could forego
treatment without any effect. But I do have some questions about that.
First, autopsy studies have in the past shown a wide range of results,
so you would have to be sure that the effect wasn't just in the normal
range of variation. Second, on autopsy, it should be easy to
distinguish clearly insignificant cancers from those which could be
detected clinically. It seems to me that it is only those that should
be counted in making the estimate. In other words, only cancers above a
certain size and showing a Gleason grade of at least 5 should be
counted.

Oh come on. Those are also being treated when discovered by PSA. That
is the point of the whole article.


.

Relevant Pages