How to play the game with numbers!

How To Play The Game:

1. SCARDINO AND HIS FIGURES
2. GAME PLAYING AT ITS BEST
3. THE ACTUAL PROVEN NUMBERS

1. SCARDINO AND HIS FIGURES
In addition the only reason that Scardino figures are better than Walsh is the simple fact that he used a definition of failure that would give him better results. He has used 0.4 vs. Walsh 0.2. In addition he has used " Treatment failure was recorded when there was either clinical evidence of disease recurrence, a rising serum prostate-specific antigen level (two measurements of 0.4 ng/mL or greater and rising), or initiation of adjuvant therapy. "

But for discussion purposes let use just the difference in numbers between 0.2 and 0.4.

In the Mayo study:"DEFINING PROSTATE SPECIFIC ANTIGEN PROGRESSION AFTER RADICAL PROSTATECTOMY: WHAT IS THE MOST APPROPRIATE CUT POINT?"
Using the identical cohort of men they found that using a definition of failure of 0.2 they arrived at a 5 year figure of 62% and a ten year figure of 43% freedom from disease progression - with the same cohort they found that using 0.4 would give them a 76% freedom at 5 years and 61% at 10 years. Therefore by using this definition to equally the scores based on 0.2 as Walsh uses we would have to subtract 14 points off of Scardino scores at 5 years and 18 points at 10 years. I now suspect that Scardino is a long way from Walsh.

Now if we use his 2nd convoluted way of defining definition of failure as per the quote above - Mayo found that by that definition (using the first date as time of failure) he arrives at even higher figures 82% at 5 years and 75% at 10 years BUT lets go one step further if he uses the second date as the failure date the figures are 86% and 61%. Lets see if we can get that in a chart:

Modification Factor Disease Freedom 5 Year 10 Year.
1. Nadir of 0.2 or less 62% 43%
2. Nadir of 0.3 or less 72% 54%
3. Nadir of 0.4 or less 76% 58%
4. Nadir of 0.5 or less 78% 61%
Nadir of 0.4 with 2 rises
5. Using 1st date of rise 82% 75%
6. Using 2nd date of rise 86% 61%
ASTRO (Official)
7. Official date of failure (1) 78% 78%
8. Modified date of failure (2) 85% 59%

2. GAME PLAYING AT ITS BEST
Now lets play games:

Assume I am a surgeon and I am about to make a report of my results at 5 years. I look at the chart above and decide that at 5 years the best definition for me to use is number 6 above and I can show a disease freedom rate of 86%. Now I accumulate data for another five years and decide to make another study and I again look at the data above and decide that this time I can get the best figures out of the ASTRO definition number 7 or even use a nadir with 2 rises (used that before) and this time use the 1st date of rise - line 5.

This is the manipulation that doctors go through - except for Walsh/JH (and some others) who more correctly uses anything over 0.2 is by definition a failure. No Mickey Mousing around - pure and simple. If you are doing a study with a minimum of 5 years of follow-up if you have a PSA over 0.2 than you have failed.

And now - compare Scardino against Walsh using adjusted figures so that we are comparing apples and apples.

3. THE ACTUAL PROVEN NUMBERS
What is really amazing and something that all should understand when reading studies or being told by a doctor his results - what is the definition of failure. In the chart above using the identical cohort of men one can achieve a range of 61% to 85% at 5 years and a range of 43% to 78% at 10 years. One can even get a 10 year figure that is better than a number of five year figures. It is a travesty of the medical doctors and there playing of statistical games for their benefit - and not for the patients benefit.

It is a game of numbers and you must understand the game. Read http://www.cancer.prostate-help.org/download/studies1.pdf for clarification. (This is a paper I have written trying to clarify how these numbers are used and what they mean. It is a rather lengthy paper - be prepared to spend some time in reading and understanding.)

In my estimation any doctor that does not use what the leading medical institutions (Johns Hopkins) for surgery uses - they are simply playing the game to get the best results they can - even though it may prove deadly for the patient. One can be slipshod and get excellent results by picking the right definition of failure. But who cares the doctors get great numbers!!! Maybe the patient should care - I did and understood this in 1997. I see no reason for differing definitions of failures for differing treatments. They all should be the same in the long term (at least 5 years minimum follow-up or even longer). All treatments will continue to fail 5, 10 and more years out. But the continuation of failure should be about the same using the same definition when you are looking at 5 and 10 year minimum follow-up.

There is no surgeon that we are aware of through published peer reviewed studies that can match the results of Walsh when figures are adjusted to a common definition of failure.

This adjustments can be made through studies like Mayo above and several others that I put online in my PDF file above. They are not made up - they are real figures as published in leading peer reviewed medical journals.

NOTE from Don Cooley - May 27, 2007:
The above was written some years ago and needs to be updated. However an updating would not change the actual figures to any great extent - if at all. In addition the PDF paper referred to is also in need of some updating and corrections.