Known knowns and Unknown Unknowns

On several occasions during scientific presentations, I have come across citations of Donald Rumsfeld's statement. Donald Rumsfeld was one of the most famous US Secretary of Defense.

"There are known knowns; there are things we know that we know. 

There are known unknowns; that is to say, there are things we now know we don't know.

But there are also unknown unknowns - there are things we do not know we don't know."

Wikipedia includes an entry on the phrase 'there are unknown unknowns,' a term popularized by Donald Rumsfeld. He famously used it in response to a question about the absence of evidence linking the Iraqi government to the supply of weapons of mass destruction to terrorist groups.

With respect to awareness and understanding, unknown unknowns can be compared to other types of problems in the following matrix:

In clinical trials, comparing an experimental therapy to a control group is often complicated by confounding factors—both known and unknown. Randomization is a key method for addressing these challenges, as it helps balance these factors across treatment groups. By randomly assigning participants to different groups, randomization ensures that potential confounders are evenly distributed, enabling a more accurate comparison of treatment effects.

For known confounding factors, stratified randomization can be employed. This approach involves dividing participants into strata based on specific factors and then randomizing them within each stratum, ensuring an equal probability of assignment to either treatment group within each category. For unknown known or unknown unknown confounding factors, the only way to minimize the impact is to utilize the randomization. 

Randomization is regarded as the cornerstone of causal inference in randomized controlled trials (RCTs). It enables researchers to attribute differences in outcomes between groups to the treatment under investigation, rather than to pre-existing differences among participants, thereby strengthening the validity of the findings.

The awareness-understanding matrix, which includes concepts like 'known unknowns' and 'unknown unknowns,' can be applied to scenarios such as xenotransplantation—for instance, the transplantation of porcine organs into humans. In the context of xenotransplantation, there is always a potential risk of zoonotic infections, where pathogens may be transmitted from animals to humans. There are known pathogens (viruses) and there are unknown pathogens. As Dr Jay Fisherman discussed the issue in his paper "Xenotransplantation-associated infectious risk: a WHO consultation":

"In xenotransplantation, there is the unique potential risk for the transmission of both known and unknown zoonotic infectious agents of animal origin into human recipients and into the wider human population."

"Xenotransplantation will necessitate the development of surveillance programs to detect known infectious agents as well as previously unknown or unexpected pathogens in the absence of recognizable clinical syndromes. This may include assays for known infectious agents, probes for classes of infectious agents (e.g., common genes or antigens of herpesviruses), and assays for unknown pathogens in a variety of tissues."

 "Unknown pathogens: Organisms not known to be human pathogens, not known to be present in the source animals, or for which clinical syndromes and microbiologic assays are poorly described or unknown"

The awareness-understanding matrix is dynamic. With advancements in science, today's unknown unknowns may eventually evolve into known unknowns or even known knowns.