-
Earlier it was pointed out that Drs. Isbister and Shander independently wrote criticism of Beliaev et al’s work.
I was surprised to see Isbister’s objection resting primarily on certain variable differentials when Beliaev’s work adjusted by logistic regression for each variable cited by Isbister. This is all the more surprising given this is expressly stated within Beliaev’s work. It’s as though Isbister didn’t thoroughly read the article.
Dr. Shander’s primary complaint focuses on the study failing to make a comparison between clinical benefit of transfusion versus another well-defined intervention, preferably one that meets what Shander presents as optimal management of JW patients. I will, though, point out that alternate treatment options applied to the JW patients is very well documented in Beliaev’s presentation. But these alternative treatments were not isolated to form a model for comparison.
I agree that moving forward the approach Shander suggests would have been more useful for the stated dual purpose of outcome and cost analysis. But for purposes of having a retrospective view of what has been occurring with JW patients the result of refusing blood, I see nothing from Shander that would argue the point.
As things are published, what Beliaev and colleagues did was present the facts as they found them, including establishing a criterion based comparison group where significant differentials were adjusted for by logistical regression.
Importantly, and for reason that should be obvious, Shander at no time suggests that the deaths that occurred among JWs did not occur, and it is these actual numbers of deaths that are important to my extrapolations. When it came to the numbers of deaths what Shander complained of was whether a higher number of deaths with lower Hb values skewed the picture of mortality among patients with higher Hb values (but still within the overall range used to define “severe anemia” within the article). But, again, it was the overall number of deaths used in my presentation which would not changed based on Hb values (again within the range).
Marvin Shilmer