Using Science to Magnify Risk

Author: Tim Platnich

Published: April 5, 2024

In this post, I will talk about three ways science is used to magnify the risks faced by humans.

Individuals face environmental risks every single day. These risks include health risks, risks of injury and even the risk of death. People make decisions based on their individual risk profiles. Some people are more risk averse than others. Before deciding whether or not to take a risk, people rely upon information concerning the contemplated risk. Science, and the reporting thereof, plays a role in risk assessment.

I will use cancer as an example. Cancer is a health risk. It is the leading cause of death in Canada. Cancer has both genetic and environmental components. Let's talk about environment components. The media often reports on various cancer-causing elements in our environment including: radiation; carcinogenic foods and beverages; pollution; chemicals and the like. It is suggested that any exposure to one or more of these items increases the risk of cancer.

This brings us to the first way that science is used to magnify risks. For simplicity, I will call anything thought to be carcinogenic a 'toxin'. In science, generally, there are two models for harm caused by toxins. One model is the threshold model. Under this model, a toxin only becomes harmful once it reaches an established threshold. There are different potential thresholds for different toxins. While under the applicable threshold, a toxin is considered 'safe'.

The other model is referred to as the linear non-threshold model or 'LNT' model. Under this model, no amount of a toxin is safe. "Even a single molecule of a chemical carcinogen could increase the risk of cancer." [Muller, HJ. 1946. The product of . Nobel Lecture, 1946 Nobleprize.org]

For a history of the LNT model and its evolution see: Edward J. Calabrese's article. This article is highly critical of the use of the LNT model in the context of cancer. Calabrese argues that the LNT model as used by Muller in the case of cancer was based on two false assumptions: 1) cell mutations are rare and therefore must be caused by something extraordinary like radiation; 2) once a mutation occurs, it is irreparable. Data now shows that cell mutations are as common as cell division and the body is masterful in detecting, repairing or destroying damaging mutated cells. But I digress.

One can imagine how the LNT model (in the wrong hands) magnifies risk! One single molecule can cause cancer! Or, at least increase the risk of cancer.

This brings me to the second way that science can amplify risk: failing to disclose clearly the difference between relative risk and absolute risk. Let's take a hypothetical example to demonstrate the difference. A scientific study identifies that Toxin A increases the risk of colon cancer by 100%. OMG! However, the risk of colon cancer is, say, .001 percent. Now the risk is .002 percent. Strike the OMG. These numbers are made up to show the increase of relative risk (100%) and of absolute risk (.002 percent).

For a concrete and recent example of relative v. absolute risk, the Canadian Centre on Substance Use and Addiction in its 2023 Final Report concluded that no amount of alcohol consumption is safe. According to the Final Report, for example, alcohol consumption by men increases the risk of many forms of cancer. The more the consumption, the higher the increased risk. For a review and analysis of this Final Report, see the Critique from the International Scientific Forum on Alcohol Research.

According to Table 2 of the Final Report, 14 standard drinks per week by a man increases the risk of liver cancer by 11.2%. Wow. That's a lot. Until one considers the absolute risk. It is estimated that 3,200 men in Canada will be diagnosed with liver cancer in 2023. According to Statistics Canada, as of that year, there were approximately 17 Million males in Canada over the age of 14. Approximately 80% of males over the age of 15 are estimated to consume alcohol. Doing the math, 3,200 over 13.6 Million equals .024 percent. So, by being a potential alcohol consuming male (without considering the amount of alcohol or any other factors) one's absolute risk of getting liver cancer is .024 percent or about 2.4 in 10,000. If one consumes 14 standard drinks per week, one's absolute risk of getting liver cancer increases from 2.4 in 10,000 to 2.62 in 10,000.

Another example is liver cirrhosis. It is estimated as of 2023 that one in ten Canadians have some form of liver disease from which about 5,000 will die. Liver cirrhosis is scarring of the liver caused by long term liver damage. It is sometime referred to as 'end-stage' liver disease. The most common forms of liver disease are viral hepatitis, fatty liver disease and liver cancer. Another form of liver disease is alcoholic liver disease. Statistics are somewhat uncertain as alcoholic liver disease overlaps with the other forms of liver disease. Some data suggests that approximately 50% of liver cirrhosis deaths are alcohol related. If we use the 5,000 number, all else being equal, one's chances of dying of liver cirrhosis in Canada is 5,000/40,000,000 or 1.25 in 10,000. Consuming alcohol increases the risk of cirrhosis. According to the Final Report, for men, 14 standard drinks per week increases the risk of liver cirrhosis by 113.6%. Oy! However, absolute risk goes from 1.25 in 10,000 to 1.42 in 10,000.

Other examples from Table 2 of the Canadian Centre on Substance Use and Addiction Final Report can be analyzed in the same fashion. On a positive note however, up to 14 standard drinks per week for men decreases the risk of ischemic heart disease and ischemic stroke by between 5 - 8 percent. This relative decrease in risk needs to be taken with the same grain of salt as with increased risks.

Let's look at absolute risk in another way. Absolute risk can be measured (or calculated) in terms of 'years of life lost' or 'YLL'. According to the Final Report, 6 standard drinks per week causes a 'moderate' risk of years of lost life. This is calculated to be 17.5 YLL in 100 lifetimes. According to the Critique, this equates to a risked loss of 64 days of life for someone who consumes 6 standard drinks per week.

The above analysis is not meant in any way to either support or criticize the consumption of alcohol. The analysis is merely intended to illustrate the differences between relative and absolute risk. The above analysis also assumes my math is correct.

The third way in which science is used to magnify risk is the 'Precautionary Principle'. If a scientific report identifies a risk, policy makers are quick to admonish people 'better to be safe than sorry'. The Report by the Canadian Centre on Substance Use and Addiction may be taken as an example of the use of the Precautionary Principle. Better not to consume any alcohol than to risk cancer or the whole host of other nasty health problems correlated to alcohol consumption. Depending on one's values and risk profile, the potential loss of 64 days of life is an unacceptable risk. For others, it is a risk worth taking. Taking the Precautionary Principle to an extreme may tempt central planners to ban alcohol consumption. Or, at the very least, use science to scare people into abstinence.