On October 19, ron posted:

I'd like to read the original article, but am unable to find it using the
reference ron provided. I'd very much appreciate further info.

Frex, what is the exact title of the publication? Is it online? It
apparently is not the Journal of Urology nor the Johns Hopkins publication
entitled, "Urology."

Thanks!

Regards,

Steve J

Interesting.

I wonder why the sample was 3003 cases when their database, according to
the CAPSure website includes 11,000 patients.

Another suprise is that the typical patient is college educated.  Since
the typical American man of that age is not college educated, what is
happening to all those other men?   Is it possible that they are not
being treated at all and hence contributing significantly to the death
rate from prostate cancer?  Alternately, of course, it may be that the
CAPSure database is strongly biased towards college educated men because
of the selection of participating clinics.

I went and looked at the paper, and I must say I am unimpressed.   The
data is interesting, but instead of trying to simplify it by broad
statements which are not meaningful, they should just have reported what
they found.

Since they just examined the patients in their database, they don't know
whether or not results flow from bias in the selection process or have
something to do with prostate cancer.   Consider the education issue.
They state that over 60 percent of the patients had some college
education compared to 52 percent nationally.  But maybe the clinics
participating in the study were fed from a population which had higher
than average college education.  That could result for example from
concentration in urban areas or for other reasons.  Or it could result
from a bias towards insurance coverage.  Later they do seem to note that
college education seemed somewhat associated with early detection in the
database, and that would seem to be an interesting, though hardly
surprising, conclusion.

Also, consider the overweight conclusion.   It is not clear from the
article what they are comparing to.  Perhaps older men as a group are
just overweight and this has nothing to do with prostate cancer.
Comparing 1997-99 to 2000-2003 in their database, they find a decrease
of men in the BMI category 'normal weight' of about 4 percent, no
signficnat change in the category 'overweight' (a decrease of 1
percent), and an increase in the category 'obese' of about 5 percent.
This does seem to show a small increase in increased weight of patients,
but what it has to do with prostate cancer is not clear.  And it might
just represent a change in the population being sampled.

One figure they report suggests something very fishy to me.  In 1997-99
the percentage of African American patients was 11 percent, but in
2000-2003 it dropped to 5 percent.   The 11 percent figure suggests a
bias in the sample to begin with, since African American men are more
likely to get prostate cancer in the first place.  But it is not
believable that incidence of prostate cancer in African American men
actually dropped in half in that period of time.   This suggests that
their database is highly skewed in important ways, and some of their
demographic conclusions are measures of the populations from which the
participating clinics selected patients than they are of prostate cancer
patients.