-
“This is the part I have trouble with. The study included four hospitals. There are more than four hospitals in that part of New Zealand. If there were Jehovah's Witnesses in that part of New Zealand being treated for anemia in other hospitals they would not have been included in the statistics. That would mean your statistics are low, which you feel is OK, because it is a more conservative number.
“Yes, or no response please, you can then add an explanation.”
The answer is not a simple yes or no because we’re not dealing strictly with incidents of “severe anemia”. The statistical deaths of “severe anemia” JW patients due to refusing blood is the count used in my extrapolation, and in this case I’ve intentionally based this count on the localized data set for the 2 regions
When we do a patient-to-patient matched comparison between patients in the same institutions with the same comorbidities et al so the sole difference is accepting or rejecting red cell transfusion we end up with a differential based strictly on accepting or rejecting red cell transfusion.
When this patient-to-patient matched comparison is performed across 4 separate medical institutions there is a degree of confidence that institutional variation has been accounted for within the spectrum of like facilities, which in this case is an array of trauma service facilities. Regardless, the matched comparison still gives us a statistical hard-count of units (statistically preventable deaths in this case) for those 4 institutions.
My extrapolation assumes a valid matched comparison and does not include any more deaths other than the ones indicated by the above differential. In other words, my extrapolation only uses the hard-count of deaths over and above the norm from the 4 institutions against the population of JWs in the 2 regions. I assume there are no additional deaths to count in the ratio of deaths per capita of JWs in New Zealand.
If another institution had as many as 10,000 or as little as 1 JW patient with “severe anemia” it would not change the statistical hard-count of deaths found by the differential in Beliaev’s study of patient data from the 4 institutions.
If another institution in the 2 region data set had even 1 death of a JW patient with “severe anemia” and refusing blood this would increase the ratio of deaths per capita of JWs in New Zealand. This is probably the case. But my extrapolation does not assume this. It assumes no more than the hard-count derived from the matched patient comparison in the 4 hospitals for the 2 regions.
Does this answer your question?
Marvin Shilmer