Why do some toxicologists make inadequate propositions or assumptions while interpreting quantitative data? And why do these same mistakes occur repeatedly? It all comes down to violation of three fundamental quantitative principles. This was presented by Dr. Wout Slob, the recipient of the 2019 EUROTOX Bo Holmstedt Memorial Award. The Bo Holmstedt Memorial Award is presented every year during the EUROTOX Congress to scientists who have made outstanding research contributions to the science of drug or chemical toxicology. This lecture also is presented during SOT Annual Meeting the following year. Unfortunately, the 2020 SOT Annual Meeting was canceled because of the pandemic. But keeping with the tradition, Dr. Slob presented his lecture during this year’s Annual Meeting. Dr. Slob works as an expert in quantitative risk assessment at Rijksinstituut voor Volksgezondheid en Milieu (RIVM), Netherlands.
Dr. Slob spoke about three fundamental quantitative principles. The first principle is “All experimental data contain sampling errors, which cannot be ignored.” For example: A group of 10 rats are dosed with a drug or chemical and no kidney lesions are observed in the experimental animals. We conclude that the applied dose does not evoke kidney lesions in rats. Dr. Slob called this conclusion a fallacy and coined the term NOMA—not observed means absent. If an additional 10 rats or more are dosed with the same chemical, it is possible that we may observe kidney lesions. Hence, the appropriate conclusion in such cases should be to calculate a confidence interval. In this case, “the true incidence in the population is between 0% and 26% (with 95% confidence).” The absence of effects can never be experimentally assessed. Therefore, we need to be aware of the NOMA fallacy and avoid it.
The second principle is “Most things are not dichotomous,” but we tend to think they are. For example: In a toxicological experiment, a drug or chemical may have “an effect” or “no effect” and there are no other options; either there is an effect or not. Consider a chemical that causes an effect—for example, a large decrease in body weight (30%), a moderate decrease (20%), or a small decrease (10%). This conclusion is actually based on the sample size and the percentage of effect may decrease with increasing sample size. So where does “no effect” start? This is not possible to answer because effect is not a dichotomous thing.
The third principle is “The distance between two values of a continuous variable is measured by a ratio, not a difference.” Dr. Slob provided an example of a data plot representing the comparison of body weight changes as a function of dose between rats and mice. The graph provided an illusion that rats are more sensitive to the chemical than mice. When the same data were plotted on a log scale, it was observed that there was no significant difference in body weight changes between rats and mice. Even points that appear as outliers in the normal graph may not be so when data are plotted on a log scale. So not plotting the data on a log scale may lead to visual illusion.
Dr. Slob summarized that any argument/concept/proposition used in science needs to obey the three fundamental principles:
- All data contain sampling errors, which must not be ignored
- Things are not dichotomous and must not be treated as if they are
- Continuous variables have a ratio-scale, and must be treated as such (e.g., by plotting data on log scale)
Dr. Slob recommended reading the book Thinking, Fast and Slow by Daniel Kahneman, a Nobel Laureate. Dr. Kahneman, a professor of psychology, observed that people use two ways of thinking, which he denoted into “System 1” and “System 2” thinking. System 1 thinking is fast, effortless, on intuition, and nonstatistical, and helps in multitasking and making quick decisions. In contrast, System 2 thinking is slow and strenuous, requires statistical analysis, is focused on one thing only, and is useful in non-urgent decisions. Dr. Slob concluded that the reason why the three fundamental principles are violated repeatedly is because of System 1 thinking by toxicologists. Cognitive errors can be avoided by checking each scientific proposition by using System 2 thinking.
In the last part of his lecture, Dr. Slob provided examples of how the three fundamental principles can be evaluated in toxicological studies. He also explained why the benchmark dose is more precise than the no-observed-adverse-effect level while deriving the point of departure.
Dr. Wout Slob’s lecture was simple yet powerful and thought-provoking. I recommend SOT members and those interested who were not able to attend this lecture on March 22 do so on the Virtual Meeting Platform.
This blog was prepared by an SOT Reporter and represents the views of the author. SOT Reporters are SOT members who volunteer to write about sessions and events in which they participate during the SOT Annual Meeting and ToxExpo. SOT does not propose or endorse any position by posting this article. If you are interested in participating in the SOT Reporter program in the future, please email Giuliana Macaluso.
Sessions delivered during the 2021 SOT Annual Meeting and ToxExpo will be available on the Virtual Meeting Platform on demand to registrants through May 31, 2021.#SOTReporter#Communique:AnnualMeeting#2021AnnualMeeting#Communique:SOTNews#Communique:ScienceNews