There’s a lot of science stuff on Twitter, and I don’t intend to make a habit of highlighting it. However, one particular account I just came across this weekend deserves the highlight for the sterling work it can do: @justsaysinmice retweets various studies with the much-needed reminder that the results have only so far been found… well, guess.
Three cheers for a really good idea!
The problem with science in the media is not a new one: it’s much more exciting to talk about the potential implications of a study rather than what the study actually says. Take this one about reused cooking oil and breast cancer metastases: it opens with the bold statement that the study says people who consume food using “thermally abused” cooking oil are more likely to develop late-stage breast cancer. This sounds really solid, but if you dig into the study, it turns out that are several issues to consider. I have several quibbles with this article.
- The mice already have (simulated) late-stage breast cancer.
- THIS STUDY IS ONLY IN MICE.
#1 is right there in the study design: “The animals were then injected with 4T1 breast cancer cells to simulate late-stage breast cancer. The 4T1 cells are known to cause an aggressive version of cancer capable of metastasizing at different parts of the body spontaneously.” Thus, this study can tell you precisely nothing about the likelihood of developing late-stage breast cancer. I’m going nuts with my italics here because it’s difficult to figure out which part of these assumptions bother me more. Late-stage breast cancer is already present in this experiment, and any results are only going to show how late-stage breast cancer responds to the dietary inputs.
#2 is a notorious issue, and that’s presumably why @justsaysinmice was born. Don’t get me wrong: experiments in mice are extremely important. They’re mammals like us and we know their genomes extremely well. We have strains of mice dedicated to all kinds of research, and they’re used because they can be used to simulate the conditions in a human body. But it’s still a simulation. It’s not perfect. On average, we only share 85% of our genomes, and that’s an average: there are genes we don’t share at all, proteins which are wildly dissimilar. Moreover, these mice have not lived a human life — they don’t have our lifespan, so they can’t possibly be exposed to all the different factors that we can. This is both their advantage (for standardising the results of experiments) and their downside (are the results applicable?).
In many cases, mice used to study cancer are in fact meant to be more vulnerable to cancer than actual human beings, in order that we can get reliable results in the kind of timeframes we have to run experiments. We’ve selected them for that purpose.
It’s really important to stress the next part…
This does not mean the results are completely inapplicable to humans.
The results of a mouse study are certainly suggestive about what the results of a human study are likely to be. You just need to remember to take it with a very large grain of salt: there are so many factors that differ between the life of a laboratory mouse and the life of a human being. The experiment is highly artificial, restricting the complications of biology as much as possible in order to control the variables.
It makes sense that the media take the interesting implications and blow them up into a big deal. It’s misleading, and sometimes it’s unnecessarily terrifying.