-
“Marvin, you say the patients were matched, but how closely? 103 patients is a very small sample. The paper that critiqued the study did mention they were comparing apples and oranges, as far as ages, and other criteria. it makes me wonder if it really is a valid comparison. I guess that is enough for the purposes of the study to say that, yes, taking blood results in less mortality, but is it enough to extrapolate from a sample with any degree of accuracy?There has to be a level of error in that study, and that level of error could result a huge difference, even if you are conservative on other areas.”
LisaRose,
The article “Comparing Apples with Oranges” by Isbister is based on a false premise that deviations such as age were not responsibly corrected for. Don’t ask me how Isbister overlooked this. But he did. You can see the corrections made in the tables, and it’s even pointed out in a footnote for those who didn’t take time to analyze the tables.
If we assume an extreme error (say, 50%) it would not begin to offset all the conservative inputs I assumed for sake of my extrapolation. I used an input value of 19 deaths. There are way more than just the 19 deaths due to refusing red cell transfusion. On top of this are additional deaths due to refusing platelets transfusion. On top of this are yet more deaths to add to the input due to refusing plasma exchange transfusion. Then we have the previous years’ greater dependency on blood transfusion prior to alternative techniques, methods and medicines we have today. This previous increased dependency on blood transfusion would add yet again to the input value.
Marvin Shilmer