Infant Mortality and Fukushima

I am prompted to write this post after a commenter on another post ( See Here ) referred to two studies on the medical effects of radiation.

One of these (the KiKK report) was legitimate science. I will come back to this in another post, because there are still some well known issues with it.

But today I want to make some observations on the other report, which proposes that the Fukushima meltdowns caused a large number of deaths in the USA in the fourteen week period immediately after the Tsunami.

The report ( Press Release Here) claims that there have been 14000 deaths in the USA in the first 14 weeks  after the Japanese Tsunami including 822 infants. It attributes these deaths to Nuclear meltdowns at Fukushima in Japan.

Scary stuff - if it were true.

Rather than having me, a mere bog standard engineer moaning about the dismal and probably dishonest methodology of the paper, you would be better off reading ( This Link ) by the illustrious Mike Moyer (Editor in charge of technology coverage at Scientific American).

If you want a complete demolition of the pseudo-statistics that forms the base of this report try This Post.. The is some more incisive  Informed Commentary Here.
But what I am going to do is make some observations on the consequences if this claim is true. We can then see how much this proposed increase in infant mortality has on the annual USA infant mortality rate and then see how this fits in with the infant mortality trends over the last 50 years.

The report "calculates" 14000 people have died from radiation effects in the USA from the Fukushima nuclear meltdown. They also state that infants have been particularly badly affected with an additional 822 dying in the 14 week period.  So if this was maintained for a year the excess number of infant deaths would be over 3000. So I am going to assume a conservative figure of an increase mortality of 2000 children between 0 -1 year over the full year. The figure is not critical. Use you own if you like.

The following maths is simple enough.

The CIA fact book (Here) tells us that there are about 4.35 million births in the USA per year. The world bank graph of USA infant mortality rate shows 2010 as 6.5 deaths per 1000 births.

That means there are 28275 deaths of infants in the USA in a year.
(4,350,000 /1000) x 6.5 = 28275

So using this study as a base, infant mortality in the USA has, from 2010 to 2011 jumped from 28275 to about 30275. All in one year.

How does our 2000 extra deaths affect the annual infant death mortality rate?

we know  28275 deaths equates to 6.5 deaths per 1000 live births

So our new rate is  (30275/28275) x 6.5 = (as near as dammit) 7 deaths per 1000 live births

Even if the 822 deaths was the maximum and the radiological mortality stopped mysteriously at the 14 week boundary this would still push the annual infant mortality rate up to 6.7 per 1000 live births.

Do you find this plausible?

Why have the thousands of front line medical staff in the USA not noticed what amounts to a huge jump in infant mortality? Especially when we are talking about the first 800 or so casualties occurring in the such a short time.

Are all USA paediatricians asleep on the job?

Seriously, do you think it even remotely realistic that nobody has noticed the undoing nearly ten years of improvement? Except for the two anti-nuclear campaigners who discovered it in their cherry picked data?

This is the World Bank graph of infant mortality in the USA over the last few decades. This Link takes you to an interactive version so you can explore the rate year by year. Notice how the infant mortality rate has continuously declined, following a roughly exponential decay from the mid 1960's to the late 1980's .

Notice how it then heads on down (post Chernobyl) at a slightly increased decline rate. 

Let us look for some events.

Notice 1979 - Three Mile Island.

Can you see anything? To me it looks like there is no discontinuity let alone a prolonged spike.

Now look to 1986 - Chernobyl.

The only noticeable thing on the graph around this time is that the rate of decrease (i.e. improvement) accelerates towards the 1990's. Maybe the extra vigilance due to the fear of Chernobyl led to this small acceleration in decline?

Over the whole fifty year period, at no time does the infant mortality rate increase.

Of course, there will be variations month by month, day by day and location by location over the year. But this only turns into an increase and epidemic if you cherry pick your data and have an agenda.

If this report is correct we should expect the first increase in the annual infant mortality rate the USA in the last fifty years, and a large increase at that.

Mike Moyer of Scientific American wrote this about the study, it's methodology and the authors:
... a check reveals that the authors’ statistical claims are critically flawed—if not deliberate mistruths.” The authors appeared to start from a conclusion—babies are dying because of Fukushima radiation—and work backwards, torturing the data to fit their claims.

This report is appallingly flawed. But it also puts politics before science. It is dishonest, immoral and cruel. Such a distortion of the truth to suite political ends is a terrible indictment of the depths that sections of anti-nuclear lobby will sink to.


Anonymous said...

Report says 14000 deaths within few weeks and it is literally terrible if the number is not fabricated otherwise.

I recently discovered the following dispersion model, which someone had linked to Berkeley’s discussion page. It uses TEPCO emission data to model possible dispersion patterns for Neptunium and Plutonium

If this model is accurate, it is very disturbing. Where are all of the so-called experts who claimed these elements were too heavy to travel far from the plant site?

BilloTheWisp said...


Thanks for the link(s).

I'm afraid I have only skimmed the modelling so far. I will try and take a closer look at it soon.

But I am a little concerned at the way you are looking at this:

First - your opening statement concerning how terrible the 14000 deaths would be if the report was not fraudulent.

I could make a fraudulent statement saying 10 million have died - Is that more terrible?.

Fraudulent is fraudulent.

Bad science is bad science, whatever the fear factor that has been invoked.

Worse still, bad science undermines good science.

For example, assuming this modelling stands up up to public expert scrutiny, it never-the-less gets undermined by the false claims made by this report.

Secondly, with your closing statement - the so called experts are just that - experts.

The modelling is unproven.

You should not assume the modelling is correct simply because it meets your preconceptions.

I look forward to seeing a peer reviewed and published version of the modellers work - and then the expert critical analysis of it.