It is easy to see numbers in almost any way you want. ODAC was
looking for negatives. I'm trained to think that way also. So, I'll
show you how it can be done. If I wanted to present these data in a
poor light for DEPO here's what I could do. First, I could take the
table I presented:
Table 4: Severity of Chemical Arachnoiditis Across Cycles; Percent of Episodes.
Number of Cycles DepoCyt (n=102) Methotrexate (n=69.5)
with DEX w/out DEX with DEX w/out DEX
Severity n = 92 n = 10 n = 60 n = 9.5
Mild 9% 20% 3% 11%
Moderate 5% 30% 5% 53%
Severe 2% 10% 3% 0%
Life-Threatening 2% 0% 0% 0%
TOTAL 19% 60% 12% 63%
This table presents the data as percentages, which is the fairest way
to present the data, as it corrects for the total numbers of observations. However, if I convert it from percentages to number
of headaches, it looks like the following:
Table 4a: Severity of Chemical Arachnoiditis Across Cycles; # of Episodes.
Number of Cycles DepoCyt (n=102) Methotrexate (n=69.5)
with DEX w/out DEX with DEX w/out DEX
Severity n = 92 n = 10 n = 60 n = 9.5
Mild 8 2 2 1
Moderate 5 3 3 5
Severe 2 1 2 0
Life-Threatening 2 0 0 0
TOTAL 17 6 7 6
Total # of Patient Headaches DepoCyt = 23 Methotrexate = 13
Presenting it in this way DepoCyt produced 77% more headaches than
Methotrexate. This is not fair to DEPO since there were wore cycles
with DepoCyt, because the patients lived longer. But, if you want
to make DEPO look bad, this will do it.
Now I can make them look even worse. Each two week cycle consists
of 1 injection of DepoCyt and 4 of Methotrexate. So, if we recast the
table into headaches per injection we get the following:
Table 4b: Severity of Chemical Arachnoiditis per Injection; # of Episodes.
Number of Cycles DepoCyt (n=102) Methotrexate (n=69.5)
with DEX w/out DEX with DEX w/out DEX
Severity n = 92 n = 10 n = 60 n = 9.5
Mild 8 2 .5 .25
Moderate 5 3 .75 1.25
Severe 2 1 .5 0
Life-Threatening 2 0 0 0
TOTAL 17 6 1.75 1.5
Total # of Patient Headaches DepoCyt = 23 Methotrexate = 3.25
Presenting the data in this way DepoCyt could be seen as producing
7 times more headaches. Of course this is totally unfair and penalizes
DepoCyt for the fact that it lasts 4 times longer. But, again, if
you want to make the data look bad, this will do quite well.
I prefer the original presentation, as I believe that it presents
the fairest overall view of the data. It corrects for the number
of observations and it corrects for the longer duration of DepoCyt.
However, if you want to view it another way, I've supplied you with
plenty of ammunition.
John de C |
