5 deaths from acute leukemia in Soma City, Japan last year.

Fukushima Diary is reporting that a Japanese lawyer stated that 5 people in Soma City, Japan died of acute leukemia last year. This lawyer added that information like this is being concealed by the Japanese government.

Soma City is about 30 miles north of the destroyed Fukushima Daiichi nuclear plant. It has a population of 36,891 according to Wikipedia. The website World Life Expectancy, Japan has an age-standardized leukemia death rate of 3.1 per 100,000.

So I investigated this to see if this leukemia mortality incidence might have something to do with Fuku. The age-standardized death rate is not really the same as the raw death rate, but typically they are very close, so we may use this rate to compute an expected number of leukemia deaths in 2012:

36891 x .000031 = 1.143621

The observed deaths were 5, with a rate of 5 / 36891 = .000135 . Note that all the leukemia deaths were from the acute variety, the info did not include deaths from chronic or other types of leukemia. So this analysis may be conservative.

Next these two proportions were compared using exact binomial statistics. The one-tailed probability of the event of death by leukemia occurring 5 or more times was P < .007 . This value is below the typical threshold of .05 (95% confidence), and also the more stringent threshold of .01 (99% confidence). But, since this a discrete statistic, this value is excessively conservative. Since our data may also be conservative to begin with, we don’t want to compound the problem, and reduce the statistical power of a dataset which already has small values. A solution generally recommended by mathematical statisticians to overcome this issue is to use mid-P probability. Here, we would take the probability of the event occurring 5 or more times, and the probability of the event occurring more than 5 times, and average them. This yields a value of P < .004, a stronger result.

It was mentioned that this is a one-tailed probability. What we are doing is setting up a null hypothesis, testing this hypothesis, and in rejecting it, we accept the alternative hypothesis:

One-tailed hypotheses:
Null hypothesis – “There is no association of leukemia deaths with radiation”
Alternative hypothesis – “Leukemia deaths are positively associated with radiation”

It has been known for many decades that radiation causes leukemia, there is the atomic bomb study and many many other studies that show this. But there is a theory called “hormesis” that says radiation protects people from cancer and leukemia. Even though there is not a shred of evidence for this ridiculous theory, sociopathic entities like the nuclear industry and the US Department of Energy continue to fund research that make this claim. Scientists see this and think, “Hmmm, maybe we better allow for the possibility that radiation is good for you”. This entails the use of two-tailed hypotheses:

Two-tailed hypotheses:
Null hypothesis – “There is no association of leukemia deaths with radiation”
Alternative hypothesis – “Leukemia deaths are positively or negatively associated with radiation”

The value we get for this dataset using two-tailed binomial statistics is P < .008 . It still rejects the null hypothesis at the strict .01 (99% confidence) level. But you can see how in other experiments, this might make the difference between a statistically significant result and a nonsignificant one. Hormesis is a trick used to reduce statistical power. It is especially egregiously bad in small datasets like this one, typically seen for thyroid cancer and leukemia deaths.

But there’s more. The epidemiology protocol usually doesn’t use simple exact binomial or other statistics. What they do is fit the data to a normal distribution and use a z-test or other tests based on this distribution. The problem is that 5 deaths are too small to make a reliable fit. And typically, nobody actually does any tests, or provides any evidence that the distribution is normal anyway. Plugging our data into the z-test yields P < .061 (one-tail) and P < .121 (two-tail). This does not qualify as significant according to the .05 level, and the two-tail doesn’t even work at the loose criterion of .10 . So we have:

Correct P < .004
Likely found P < .121

You have to wonder how many studies are like this. Instead of having a powerful result, that satisfies the strict criterion, we would see “We found no evidence that radiation caused leukemia deaths in Soma City”. Or even worse, we might see “There is no evidence that radiation causes leukemia deaths”, which is a non sequitur. It is a short hop and skip from this to “We expect no health impacts in the United States from Fukushima radiation”.

I wanted to add some recent measurements of iodine-131 in Japanese sewage sludge. We can see that there was a massive release of I-131 in August and September. This has decayed by now, and any iodine-131 floating around originated after this event. Of course, there is probably iodine-129 too, which is rarely reported, and hangs around for 157 million years.

7 thoughts on “5 deaths from acute leukemia in Soma City, Japan last year.

  1. Incredible work Bobby1!!!
    I actually understood most of that,….so I’m a little scared? 🙂
    But pleasantly so!

    But scary! I think we all know that the REAL numbers will never be known,…but are HIGH!

    The saddest thing to me,…is how slow and painful these deaths are!
    That sucks on ice!

    Great piece!

    Love you guys!
    Jilly

  2. When I was a child I had a friend that died of leukemia, he was the oldest and only healthy child the parents had, the other 2 boys had Muscular Dystrophy, I remember the first day our family saw the mother walking down the road in front of our house, later we befriended them an we played together daily as they deteriorated and were not able to walk at all, so they went everywhere with us in wagons as we pulled them everywhere playing !
    Later when I was in training in the Army and came home I was told and went to see him before I was sent to Vietnam, reminding you the other two brothers had already died from their ailments !
    When I came home, the parents were only left, but I heard a few years later the mother succumbed to a cancer !

Leave a Reply

Your email address will not be published. Required fields are marked *