250,000 Jehovah's Witnesses have died refusing blood

by nicolaou 739 Replies latest watchtower medical

  • LisaRose
    LisaRose

    I wonder if is even possible they matched these patients in enough ways to make the results statistically meaningful. People are all unique individuals. Unless you are comparing identical twins, it's obvious that they are not matching in every thing, maybe not even then. You brush off the criticisms of Ibister as if they are nothing, but he did say they were comparing apples to oranges. I believe you do not want to accept the possibility that he could be right. With only 103 JW patients, it wouldn't be hard to totally throw off the results. That is why the results of only one study are suspect, especially when the control is so small.

    Do you not see the possibility that the study could be wrong or overstated?

  • Simon
    Simon

    Also, the data you’re using appears to come from cause of death data. Refusing blood is not usually cited as cause of death. For instance, cause of death from hemorrhage due to blunt trauma by a patient who refuses red cell transfusion would not be listed as “died due to refusing blood”. This is one of many reasons why these deaths fly under the radar and whose quantity is identified by retrospective statistical analysis.

    No Marvin, you have once again (I suspect deliberately) missed the point so you can gloss over it and get back to underlining things.

    The figures are approximations by applying the cause of death data to the JW population. Unless we have reason to believe they would be any different to the general population we can go with those. Actually, I suspect JWs would possibly to suffer from lung cancer and other things caused by smoking less than general but that would just confuse things further.

    The main thing is, by applying well known, accepted and pretty stable rates to the whole JW population we can come up with approximations for how many die from different causes each year. It doesn't matter if they died from refusing blood or not - that would only make the numbers higher so I guess you could say they are "conservative". But I won't, I'll just say they are a believable and good-enough approximation.

    The 'blood' column though is reversed as this is taking your 50,000 total and applying it pro-rata based on the number of JW deaths each year. So, basically if we take 2011 it would be roughly 1.8k JW deaths caused by refusal of blood transfusion. That is based on your number (it works out slightly different due to me working from the higher numbers downwards instead of vice-versa but it's close enough).

    Now, in 2011 the regular data tells is that 14k would die from Heart disease, 13.5k from Cancer, 1.9k from alzheimers and dementia, 1.6k from influenza and 900 suicides.

    So, you are saying that as many JWs will die due to refusing blood as will die from alzheimers and dementia and more than die from influenza, double how many commit suicide. That there will be someone dying from refusing blood for every 7 or so that die from heart disease, that die from cancer ...

    That is the implication of applying your numbers to more accepted and available numbers. I think of it as a sanity check.

    What I'm saying is that I find your numbers difficult to believe and the more likely explanation is not that death rates in general are wrong or the number of JWs is wrong, but that one minor study could have produced anomalous results and extrapolating them as you have in a simplistic way simply amplifies the anomaly.

  • Simon
    Simon

    Oh, and 50,000 means that 3 out of every 100 JW deaths would be due to refusing blood.

    250,000 would mean 15 out of every 100 JW deaths would be due to refusing blood.

    I think you are vastly overstating the issue.

    But I know, you are going to tell me you have hard numbers, that there are 19 of them etc...

  • Marvin Shilmer
    Marvin Shilmer

    -

    “The figures are approximations by applying the cause of death data to the JW population.”

    Yes. That is what I said before in other words.

    Refusing blood is almost never cited as cause of death. The deaths we’re talking about are buried in the numbers in your chart from one side to the other, sometimes including even suicide believe it or not.

    “Now, in 2011 the regular data tells is that 14k would die from Heart disease, 13.5k from Cancer, 1.9k from alzheimers and dementia, 1.6k from influenza and 900 suicides.”

    A deceased’s cause of death listed as “heart disease” may well be a result of refusing blood.

    Same with cancer.

    Etc.

    The numbers of those who’ve died prematurely are buried in the statistics you show, only the deaths we’re talking about add to the overall mortality. But, again, they’re buried under “cause of death” that does not identify the active consequent of refusing blood when it would have made a difference.

    I don’t have access to the patient records in Beliaev’s data set. But it would not surprise me to learn that “cause of death” was not listed as “refusing blood” in even one instance. Yet the reality is that these deaths were identified by retrospective statistical analysis with a common cause: refusing blood when it would have saved a life.

    Marvin Shilmer

  • Marvin Shilmer
    Marvin Shilmer

    -

    “I wonder if is even possible they matched these patients in enough ways to make the results statistically meaningful. People are all unique individuals. Unless you are comparing identical twins, it's obvious that they are not matching in every thing, maybe not even then. You brush off the criticisms of Ibister as if they are nothing, but he did say they were comparing apples to oranges. I believe you do not want to accept the possibility that he could be right. With only 103 JW patients, it wouldn't be hard to totally throw off the results. That is why the results of only one study are suspect, especially when the control is so small.

    “Do you not see the possibility that the study could be wrong or overstated?”

    LisaRose,

    I’ve not brushed off Isbister’s criticism. I’ve pointed out a demonstrable false premise of his criticism. He overlooked the very thing he complained of. It’s right in the articles for everyone to see.

    To borrow a phrase, Do you not see the possibility that Isbister could be wrong?

    Otherwise, in a matched comparison the disparity between 2:103 and 21:103 is not likely to be freak happenstance.

    Marvin Shilmer

  • LisaRose
    LisaRose

    Sure, but I don't have the study. I just want you to admit that the study could be flawed, as most studies are.

  • Marvin Shilmer
    Marvin Shilmer

    -

    “I just want you to admit that the study could be flawed, as most studies are.”

    LisaRose,

    I don’t know that a particular presentation of information is flawed until a flaw is demonstrated one way or another.

    I don’t know that “most studies are” flawed. That’s a sweeping generalization.

    Marvin Shilmer

  • LisaRose
    LisaRose

    OK, I will concede that. But you know how the news reports are always saying "studies show that Elbonians who eat oat bran weigh 20% less. Then everyone rushes out to buy oat bran. Then six months later, another study come out saying Elbonians who eat oat bran weigh 20% more. I am sure you realize that not every study is 100% accurate or true. Whether the premise is flawed or the data was corrupt, or the wrong conclusion was drawn. It happens.

    And you must acknowledge that one study is less convincing than six that show the basic same thing. And you must also acknowledge that a smaller study has more room for variation than a larger one. If you can agree with that, we can move on.

  • adamah
    adamah

    Marvin said-

    Beliaev does not document how many 16-18-year olds died in the JW group.

    That's a problem, since that would be VERY IMPORTANT data to know, in order tighten up unknowns to obtain statistically-valid conclusions. Since it wasn't recorded/reported, we can only guess (similar to how the JWs don't track and report the # of deaths Worldwide). The problem is NZ's dangerously-low age of consent (16; in the rest of the World, it's generally 18).

    Now tell everyone why this could vastly inflate the number of my extrapolation. Can you please do that?

    Sure.

    Let's suppose ONLY ONE of those patients in Beliaev's study WAS aged 16-18 yrs, and hence would be allowed under NZ law to die, which wouldn't be the case if they had lived virtually anywhere else in the rest of the World (say, AU or US, where age of consent is 18). If they HAD, the State would step in and FORCE life-saving blood against their parent's or their wishes. In the U.S., we don't let 16 yr olds die, like they MUST do in NZ.

    (Here's a recent thread about a 17 y.o. in AU who's life was protected by the AU Court system, a safety net which wouldn't under him had the teen lived in NZ (since he'd be allowed to reject blood and die at 17 in NZ and no one could say boo). Here's a recent thread about one such boy who made the media outlets:)

    http://www.jehovahs-witness.net/watchtower/medical/262354/1/Jehovahs-Witness-boy-fights-court-for-right-to-die#.Ummm9SQhZLc

    So suppose only ONE participant was aged 16-18 in Beliaev's study. What would that do to your Worldwide extrapolation figure?

    Let's find out what happens if ONE pt wasn't allowed to die (like they would in NZ).

    From your website:

    This study found 103 Witness patients who suffered severe anemia and 20.4% 19.4% of these died (i.e., 21 20 of these patients died).

    During this period the death rate among patients who suffered severe anemia but accepted ARBC transfusion was 1.9%. The net difference in mortality rate is 17.5% 18.45%. Of the Witnesses who died, the number who died over and beyond the group that accepted ARBC transfusion (the non-Witness group) is 18 19, or 17.5% 18.45% of the 103.

    ● Over a 10 year period there were 18 19 Witnesses who died that shouldn’t have. That’s 1.8 1.9 per year.

    ● Over the same 10 year period the number of Witnesses in New Zealand averaged 12,700. 1.8 1.9 deaths per year is 0.014% .015% of the Witness population.

    ● During year 2011 the Watchtower organization presents the number of Jehovah’s Witnesses at 7,400,000.[3] 0.014% 0.015% of this number is 1,049 1109.

    ● During year 2011 there were 1,049 1109 premature deaths among Jehovah's Witnesses due to Watchtower's blood doctrine.

    I'll leave you to do the rest of the math (noting that only ONE pt less in the JW death count results in an over-estimattion for 2011 by over 5%; a similiar error overall would drop your 50k figure to 45k, if applying the same difference found in 2011 numbers; not a valid assumption, but I'm not taking the time to do it for EVERY YEAR and then add up the results).

    That demonstrates the importance of what Simon has repeatedly stated about the so-called "butterfly effect" in statistics, where small changes in the input side of the calculator can have a dramatic impact on the final results (especially when you multiply those seemingly minor changes by 7.4 x 10 6 )!

    But again, the actual number of such teen patients who participated in the study is unknown, and hence that's a MAJOR OVERSIGHT in the study for the purposes of extrapolation to the various legal climates around the World which don't let teens die. It would've been nice for the study authors to have reported the ages of the 21 JWs who died, so as to support a campaign of RAISING NZ's relatively-low age of consent, or to support extrapolation Worldwide as you'd like to do in a responsible manner manner supported by valid accepted methodology of statistical analysis.

    BTW, you said this earlier to LisaRose:

    Don’t forget that the Beliaev study was a matched comparison. This means patient profiles and comorbidities were accounted for. So, for instance, in the 103 JW patient pool there were 8 with cancer and in the non-JW pool there were 15 patients with cancer. But regardless of condition, if treatment options were exhausted so a patient was in palliative care they were not included in the study because inclusion would skew the result.

    I suspect that factor alone would skew the results, but not in the way you're suggesting, since excluding those end-of-life palliative care patients from BOTH groups would actually INCREASE the effect of the more easily-treatable young patients who died, hence amplifying the effects of those in this otherwise-protected group who live in other Countries (the exclusion would tend to bias the data set away from the older patients, and skewing towards the younger, more-easily saved, thus skewing the results to over-exaggerate those deaths).

    Marvin said-

    There has been much bantered around in this discussion about whether New Zealanders have a higher rate of anemia (in this case Hb =/< 8 grams dL). This criticism boils down to: If New Zealanders have a higher rate of anemia compared to the rest of the world on average, this could mean my extrapolation at issue is inflated.

    Here’s what wrong with that notion: There is a difference between rate of anemia and rate of mortality due to anemia, and there is a gargantuan difference between rate of anemia and rate of mortality due to refusing blood with anemia.

    Because anemia is a largely treatable condition then rate of anemia is not so much a determinant of mortality due to anemia. What counts in my extrapolation is not rate of anemia (or severe anemia) but, rather, known deaths attributed solely to lack of red cell transfusion in patients suffering severe anemia. Rate of anemia does not change this mortality statistic against the population.

    And that's what you're missing:

    Not only does the sample population have a higher rate of anemia, but you admit that in the sub-population it has a higher mortality rate if left untreated, and those two factors explain WHY the study is unfit for extrapolation purposes: it's not truly a representative sample of the entire population Worldwide. Hence, the results would likely VASTLY AMPLIFY the mortality due to REFUSING BT, and hence the results wouldn't extrapolate to OTHER groups. Why? The subjects of the study weren't an accurate sampling of the population you're trying to extrapolate to!

    Take the incidence of diabetes amongst the Navajo Nation: four times higher than the national average, diabetes is also more fatal if left untreated (the life expectancy of a male Navajo is shocking low, like in the low 50's). Hence it's more important for Navajo pts with diabetes to take their meds regularly, good nutrition, exercise, don't drink/smoke, etc, since diabetes will cause vastly-more problems, and at a younger age

    (As part of my clinical rotations in my last year of training, I spent a quarter at Shiprock Indian Health Service Hospital, NM,dealing with the management of pts with dbb. I remember a few pts in their 30's who were on dialysis, losing limbs in wheelchairs, etc but they continued to not take their condition seriously, with a fatalistic "Oh, well" attitude. Sad, but people will do what they will do, and you can only advise, since you cannot force anyone to care if they don't.)

    Now, given that we know the Navajo tribe is not representative of the U.S. population (300 mil), it would be ENTIRELY INAPPROPRIATE to try to extrapolate a figure on the prevalence and mortality from diabetes in the US based on a study conducted on the Navajo reservation, since we KNOW it isn't representative! Sampling a population to make sure it truly is representative of what you're trying to sample is a subject with many courses dedicated to the task.

    Invalid sampling makes extrapolation unstable and unreliable, as attempts to draw conclusions become unreliable when the sample actually isn't representative of the larger group you're trying to draw conclusions about. That's why ALL studies can only reliably tell us anything for the environment they were designed to sample (and YES, some study results MAY lend themselves to extrapolation, AS LONG AS the population is fairly similar in characteristics to that used in the study; you refer to this in matched groups, but you cannot be sure the groups used in the study are similar to other populations, and you cannot just assume they are similar).

    In this case, Beliaev's study was looking at the clinical outcomes and cost comparisons within NZ for refusing blood, and anything beyond that causes the confidence interval (margin of error) to get unacceptably large.

    Adam

  • Marvin Shilmer
    Marvin Shilmer

    -

    “And you must acknowledge that one study is less convincing than six that show the basic same thing. And you must also acknowledge that a smaller study has more room for variation than a larger one. If you can agree with that, we can move on.”

    LisaRose,

    More documented information is always more convincing that less documented information. In other words I’ve expressed this several times during this discussion. I haven’t changed my mind.

    Adamah,

    When composing my extrapolation I considered the issue of teens and a higher incidence ratio unrelated to refusal of blood product (e.g., Maori ethnicity).

    Here’s why I opted not to mathematically account for the issue of teen deaths due to red cell transfusion refusal:

    1. I don’t know that there is anything to account for because I don’t know that any of the teens are among the 21 deaths in the 103 JW patients.

    2. I’m familiar with laws and customs related to refusal of treatment of a broad spectrum of healthcare service areas around the world. I disagree with you about the potential for inclusion of teens in Beliaev’s study by comparison for at least the following reasons: 1) In third-world areas refusal of blood transfusion by adults alone would more than offset the potential disparity you cite because these service areas are more dependant on red cell transfusion to treat anemia than in developed service areas. 2) In developed healthcare service areas (e.g., the USA) there is a medical ethics concept known as “mature minor”. Under this ethical legal construct doctors and institutions have allowed themselves to perform high risk procedures on minors who achieved a rather subjective competency threshold value and who refused blood. At the very least this would mitigate the effect you cite. 3) The advent of alternatives to red cell transfusion to treat anemia (particularly in the Hb range of =/<8 g dL to >7 g dL) further mitigates the effect of teen inclusion because the threshold value of Hb =/<8 g dL does not necessarily demand transfusion of red cells when alternative treatment is available. In many instances there is actually a good medical argument not to give red cell transfusion to treat anemia Hb =/>6-7 g dL. Hence had a JW teen with Hb =/>6-7 g dL not been treated with red cells transfusion it could easily be the case the transfusion was avoided because of implementing an alternative therapy that is arguably better and that this was done regardless of patient preference of no red cell transfusion. In other words, in these cases alternative treatment effectivly trumped the question of red cell transfusion or no red cell transfusion.

    3. In addition to the preceding, there are additional deaths due to JWs refusing blood products other than red cells, and the Beliaev study examined outcomes related strictly to red cell transfusion. Alone this single factor would likely (if not surely) outweigh the effect you cite of teen inclusion, should it be real.

    “Not only does the sample population have a higher rate of anemia, but you admit that in the sub-population it has a higher mortality rate if left untreated, and those two factors explain WHY the study is unfit for extrapolation purposes: it's not truly a representative sample of the entire population Worldwide.”

    I do not compare results of the Beliaev study with the entire population worldwide.

    I compare the result of the Beliaev study with the population of JWs worldwide.

    JW make up the “the sample population” you speak of and the “sub-population” you speak of, and it's mortality among the JW population that is at issue.

    If I’ve misunderstood you please feel free to restate yourself or otherwise clarify.

    Marvin Shilmer

Share this

Google+
Pinterest
Reddit