Yes we’ve been through it — you said the principle was the same, and I said the principle was not the same. Clinical trials do not curve fit data from a sample population, and use extrapolation to predict the magnitude of effect on another population selected specifically for being at the extreme range. Their conclusions are typically binary and high level, like “it looks like it worksâ€, and “it looks safeâ€.
While clinical trials allow us to determine a drug is generally safe, and/or more effective than previous treatments, no single clinical trial in the history of clinical trials, has developed a performance model from the data of mediocre athletes, for the purpose of allowing to predict changes in elite athlete performances by using the error-prone method of extrapolation.
There is nothing about any principle of clinical trials that is the same as extrapolating data from the trial to the extremes.
You seem obsessed with “directionâ€, while not addressing the magnitude, or in this case “diminishing magnitudeâ€.
I make the extraordinary claim that the effect gets smaller “at the top†— perhaps eventually so small that it gets lost in the noise.
Take the EPO study with the Scottish runners, and the Kenyan runners.
We saw the 10-minute runners improve by 38 seconds, or 5.7%.
We saw the 9-minute runners improve by 25 seconds, or 4.46%.
Here’s some “counter-evidence†supporting the “extraordinary†claim that “EPO works LESS†on faster runners.
This suggests a relation exists between the level performance and how much “EPO worksâ€.
Yet, from these studies, the researchers don’t propose any curve fit that would allow us to estimate how fast “EPO works†diminishes extrapolate.
What can we reasonably expect for the 8:00 runner? 7:40 runner? 7:25 runner?
In these studies, I didn’t see information about a control group, showing how much non-EPO subjects would improve, with comparable training.
We have no real way to assess how much of the observed improvement was due to EPO.
Take them out then. Repeat pre- and post- 1985 observations exclusively on non-Africans, and the observations remain unchanged.
If EPO works, we cannot determine how much “EPO works†on the non-African elite male population based on a progression that has virtually stagnated for the two decades of the EPO-era.
Those are your words. I said “African genetic AND ENVIRONMENTAL "confoundersâ€â€.
While “genetic diversity†is a solid enough reason enough to separate observations from the African population and from non-African populations (see for example “www.geneticliteracyproject.org/2015/06/01/does-africas-rich-genetic-diversity-explain-dominance-of-elite-african-runners/“), I also wanted to acknowledge and eliminate confounding factors in environmental differences, such as high altitude, economic situation, personal motivation, etc., that separate Africans from richer, first world nations.
The pre-EPO 1985 baseline is a reference point to help put the magnitude of changes, from all sources, into perspective.
When a curve is flat between 1985 and 2009, it is hard to argue that, inside that flat curve, “EPO worksâ€.
If we include post-1985 steroid users, we cannot say the change was due to EPO. We have to claim something else — like EPO plus steroids resulted in little change.
But the more important observation, is that, even including steroids, we have seen global stagnation of performances, for non-African elite males, through 2009, from all sources.
I didn’t actually accuse anyone, but simply stated a no-nonsense criteria that we must eliminate the risk of concluding “EPO works†from runners who improved when taking steroids. It is not required to name them.
By your standard, the Chinese women did not test positive for EPO either, so they should not be used to support an “EPO works†thesis.
Am I really conducting a “population study� It would seem to me that “observational study†is more accurate.
From wikipedia “an observational study draws inferences from a sample to a population where the independent variable is not under the control of the researcher because of ethical concerns or logistical constraints.â€
I’m not making “extra-ordinary claims†contradicting “EPO worksâ€, but asking the questions, how much does EPO work at the top?, and can we actually observe EPO-like magnitude of changes at the top? and what part of the observed changes can we attribute to EPO?
If EPO studies promise 25-38 second, or 4.5%-5.7%, improvements for 3K, in the lab, and other “professional†estimates are ~20 seconds for 5K, ~40 seconds for 10K, and 3-4 minutes for marathon, I think is a fair to assume that, over a period of decades, some new significant performances would emerge, in many populations, and it is reasonable to attempt to look for and identify any such impact.
This would especially be true in longer events, like the marathon, where steroids would seem to have less impact.
When we look at non-African elite males globally for two decades, the clear answer is — we cannot see any real change “at the topâ€, from all sources, as a result of the introduction of EPO, in pretty much every distance event.