It's an issue, but that would be about the reliability of a test. What Arkmedic discusses here is primarily the abuse of statistics. I use the example (probably stole it from someone else, I admit) of buying two lottery tickets instead of one. You have improved your chance to win by 100%. But your odds of winning are still many millions …
It's an issue, but that would be about the reliability of a test. What Arkmedic discusses here is primarily the abuse of statistics. I use the example (probably stole it from someone else, I admit) of buying two lottery tickets instead of one. You have improved your chance to win by 100%. But your odds of winning are still many millions against. In most medical contexts the rates are usually in the one per 1,000 to 10,000 range. Of several that I've looked at, the prize for "Hardly does anything" currently is held by aspirin therapy, which I computed provides an annual (absolute) risk reduction for healthy people of a whopping 0.02% (not a typo, that's "two one-hundredths of a percent", although the relative risk sounded better, at about 5.7% improvement.
If I recall one of Kendrick's rules of thumb: Generally speaking, if they are claiming a benefit, they will use a relative ratio. If dissimulating adverse data, use absolute.
1-thanks for the lottery analogy. very useful way to explain/understand RR concept.
2--the PCR test validity is very much an issue. first, it is useless test and not supposed to be used this way--as a diagnostic. second, it has a false positive rate that would be unacceptable in most realms of life--e.g., investing, DWI allegations etc. and of course, presence of antibodies doesn't actually mean you will feel sick/show symptoms.
I still can't believe how shallow the pharma research is apparently. 2-3 yrs ago I would have assumed arkmedic-type analysis was part of the vetting of new drugs. how naïve I was.
Agreed. I'm dismayed how half-assed the entire industry is, from research on down to the doctor who writes Rx for your pill. I no longer trust doctors of any type, and to the extent possible, will do my own research about proposed treatments before subjecting myself to them.
Small risk, but there is a risk of serious bleeding even with the low daily dose. Just conjecturing, but I'm sure the risk of death is low, but "serious bleeding" is, well, serious.
Based upon my admittedly limited research, the risk of bleeding exceeds the expected protective benefit.
It's an issue, but that would be about the reliability of a test. What Arkmedic discusses here is primarily the abuse of statistics. I use the example (probably stole it from someone else, I admit) of buying two lottery tickets instead of one. You have improved your chance to win by 100%. But your odds of winning are still many millions against. In most medical contexts the rates are usually in the one per 1,000 to 10,000 range. Of several that I've looked at, the prize for "Hardly does anything" currently is held by aspirin therapy, which I computed provides an annual (absolute) risk reduction for healthy people of a whopping 0.02% (not a typo, that's "two one-hundredths of a percent", although the relative risk sounded better, at about 5.7% improvement.
If I recall one of Kendrick's rules of thumb: Generally speaking, if they are claiming a benefit, they will use a relative ratio. If dissimulating adverse data, use absolute.
1-thanks for the lottery analogy. very useful way to explain/understand RR concept.
2--the PCR test validity is very much an issue. first, it is useless test and not supposed to be used this way--as a diagnostic. second, it has a false positive rate that would be unacceptable in most realms of life--e.g., investing, DWI allegations etc. and of course, presence of antibodies doesn't actually mean you will feel sick/show symptoms.
I still can't believe how shallow the pharma research is apparently. 2-3 yrs ago I would have assumed arkmedic-type analysis was part of the vetting of new drugs. how naïve I was.
Agreed. I'm dismayed how half-assed the entire industry is, from research on down to the doctor who writes Rx for your pill. I no longer trust doctors of any type, and to the extent possible, will do my own research about proposed treatments before subjecting myself to them.
At least a daily aspirin won't kill you?! Or will it?!
Small risk, but there is a risk of serious bleeding even with the low daily dose. Just conjecturing, but I'm sure the risk of death is low, but "serious bleeding" is, well, serious.
Based upon my admittedly limited research, the risk of bleeding exceeds the expected protective benefit.
I had a mini stroke after month or so of taking aspirin for a condition. I'll never take aspirin again.