SNT is too conservative

Some members of the choir have wondered if SNT might be too aggressive, might under-estimate the cancer incidence associated with a dose rate profile. This piece argues that the opposite is true. It’s a bit technical. The prerequisite is a little knowledge of the LNT and SNT radiation harm models.

Detecting Detectable Harm

In 1980, Lauriston Taylor, founding chairman of the NCRP, said.

Collectively, there exists a vast array of facts and general knowledge about ionizing radiation effects on animal and man. It cannot be disputed that the depth and extent of this knowledge is unmatched by that for most of the myriads of other toxic agents know to man. No one has been identifiably injured by radiation while working within the first numerical standards first set by the NCRP and then the ICRP in 1934. [2 mSv/day]

The Green Table shows that this is still the case. In the top portion of the table are populations which have received their dose in less than a day. In this case, we start to see significant increase in cancer somewhere between 100 and 300 mSv. The bottom portion shows populations which have received their dose more or less evenly over periods of years or decades. For these people, as long as the dose rate is below 20 mSv/d, we see no increase in cancer regardless of the total dose which in some cases is over 100,000 mSv. To see detectable harm --- get out of the green --- requires dose rates of 20 mSv/d or more. Taylor's 2 mSv/d provides a factor of 10 margin. This is consistent with everything we know about our ability to repair radiation damage. It's a good rule.

At the same time, the Gordian Knot Group proposes to replace LNT with SNT as the basis for our radiation protection regulation, and the basis for radiation exposure compensation. This creates a possible contradiction. Suppose someone is exposed to 2 mSv/day for every day of an 80 year life. According to SNT, this dose rate profile increases his cancer mortality by 2.8%. His Lost Life Expectancy is 122 days. If somebody experienced this dose rate profile as the result of a nuclear power plant release, under UCert she would receive compensation of $42,700.

A 2.8% increase in cancer would be detectable if we had a large enough sample population. This might seem to contradict Taylor's claim of no detectable harm. But Taylor himself was careful to point out that the fact that we had not detected any harm, does not imply that we would not be able to detect harm if the tolerance dose is imposed for a long enough period.

The adoption of tolerance or permissible doses did not depend upon the assumed existence of a threshold. The setting of the tolerance level was on the basis of whether or not effects could be observed, and this has been the common approach to similar problems for all other toxicological agents.\cite{taylor-1971}[page 13]

The only dose, therefore, that could be harmless would be one that would have no effect whatever on the particular tissue concerned and, hence, be administered ad libitum.\cite{taylor-1971}[page 12]

If there is any harm at all, one can dream up situations in which you would be able to detect that harm, at least theoretically. In this case, a life spent in High Earth Orbit might do it.¹

According to Cohen, Figure 1, a 122 day Lost Life Expectancy is a risk on the same level as drinking alcohol and less than dying in a car crash. In other words, it is equivalent to risks that most of us except with little or no thought. It's a risk that an astronaut might willingly assume in order to obtain the benefits of being an astronaut.

Figure 1 Lost Life Expectancies.\cite{cohen-1991}}

Is SNT Too Conservative?

On the other hand, a glance at the Green Table shows that the 2 mSv/day rule is highly and consistently conservative in relation to all the dose rate profiles we earth bound humans will experience. And SNT is conservative relative to the 2 mSv/day rule. There are at least a half dozen examples in the Green Table. Perhaps the most compelling is the radium dial painters, Figure 2.²

Figure 2 LNT and SNT versus Dial Painter Bone Cancer

SNT claims we should see nearly no cancer up to about 2 mSv/day; but then the prediction turns sharply upward. The actual jump upward takes place around 20 mSv/day. Moreover, at 20 mSv/d, SNT predicts a 99% incidence rate while the observed rate never gets above 40%. This leveling off of cancer incidence at the high end at well below 1.0 is something we see in other extremely high dose populations, such as the bomb survivors, and children who received very high thyroid doses after Chernobyl. It is not a statistical fluke.

S-shaped models have a procedurely easy fix to the high end. Simply change the top of the high end hook from 1.0 to say 0.4. There is a possible biological justification for this. The idea is that cancer results from viable misrepairs of the cell's DNA, mainly from misrejoining the ends of closely spaced DSB's. The cell then has to survive and replicate to start the path to cancer. If the dose to the cell is too high, either a viable misrepair becomes less likely or the cell is unable to replicate due to damage outside the DNA. A cell dying is not a problem for the organism as a whole unless it loses so many cells that we begin to see Acute Radiation Syndrome. It's quite possible there is a range of high dose rates where we see no increase in cancer incidence and no clinical symptoms of ARS.

The orange line in Figure 3 shows a Five Parameter Logistic fit to the dial painter data that does a reasonable job of regurgitating the dial painter cancer incidence.

Figure 3. LNT and SNT versus Dial Painter Bone Cancer

Unfortunately, this curve results in a poor fit to the bomb survivor data. The dial painter data is in my view much more reliable than the bomb survivor. But the dial painter exposure was internal due to radium, an alpha emitter. The bomb survivor doses were mainly external photons, which is much closer to the type of exposure experienced in a power plant release. Barring the emergence of the liquidator dose-rate profiles, I’m afraid we have to stick with the bomb survivors.

We can reluctantly live with this conservatism. Reasonable worst case releases are still insurable which is not the case for LNT. Over-compensating people exposed to radiation will decrease overall social welfare; but perhaps this loss can be justified on the pragmatic grounds that without this over-compensation nuclear will never have a chance to achieve its promethean promise. But don't let anybody tell you, SNT is too aggressive.

1. Astronauts in Low Earth Orbit get 0.5 to 1 mSv/day, more during a big solar flare.

2. It’s actually worse than it looks. To convert the dial painter total doses to dose rates, I had to assume the dose was received evenly. In fact, their dose rate profiles were uneven and included long periods in which the dose rate was above average. For any non-linear model, tan uneven dose rate profile will produce more cancer than a flat profile at the average dose rate.