Scientific progress is driven in part by failures and accidental discoveries.
In the early 1990s, for instance, sildenafil was performing badly as a treatment for angina when researchers at Pfizer noticed something odd: Male volunteers kept reporting erections.
Pfizer soon abandoned its original plans for sildenafil. Instead, they marketed it as Viagra, and what had been a failed heart medicine became the first effective ED drug.
Accidents and failures can be useful, but what about bad science? What value could there be in studies so poorly executed that other scientists ignore them (or subject them to ridicule)?
In this newsletter I'll be describing a new study claiming to show that depressed people transfer oral bacteria to their spouses and cause them to become depressed too.
In essence, the claim is that emotional disturbances are contagious. Not metaphorically, but in the literal, biological sense. Kissing a depressed person puts you at risk of depression. Sharing spoons with an anxious person might increase your anxiety.
This study is deeply flawed – the researchers misinterpret their own data – but, as I'll explain, the findings are valuable nonetheless.
Emotional contagion
There's nothing new about the concept of "catching" feelings. In one of the most memorable scenes from the Iliad, Priam sneaks into Achilles' tent one night and, as he grieves over the loss of his sons, Achilles is quickly moved to tears.
Nor is it a new idea that people in close relationships "catch" emotional problems from each other. Contagious depression has been shown among spouses, couples who are dating, and even college roommates.
What's special about the new study is the claim that oral bacteria are the mechanism of contagion.
(Experts on the topic, including Elaine Hatfield, the University of Hawaii psychologist who coined the term "emotional contagion" back in the 1990s, only speak of contagiousness metaphorically.)
Emotional contagion is often desirable. If someone close to us is suffering, we would want to empathize. If they're joyful, we'd want to share in their joy.
The downside is that when the other person is constantly depressed, or anxious, or otherwise miserable in our presence, we're susceptible to developing similar feelings.
If the new study can be trusted, this sort of emotional contagion might be preventable by simply adjusting the oral microbiome. Prescription mouthwash might be all we need.
The new study
This study, published in Exploratory Research and Hypothesis in Medicine three weeks ago, was led by Reza Rastmanesh and an international team of researchers from Iran, India, Italy, and the UK.
The study focused on heterosexual newlyweds – 268 couples who'd gotten married at some point during the previous 6 months.
These weren't just any couples. They were chosen because screening tests indicated that one member of each couple was experiencing moderate levels of depression, anxiety, and insomnia, while the other member was not.
This part of the study seems fine. To screen participants, researchers used three validated, widely-respected measures – the Beck Depression Inventory (BDI-II), the Beck Anxiety Inventory (BAI), and the Pittsburgh Sleep Quality Index (PSQI).
The researchers also obtained oral samples from the mouths and tonsils of each participant at the time of screening.
180 days later, all measurements were repeated.
So far, so good...
Main findings
The researchers compared measurements taken on the first day to those taken on day 180. Here are the two main changes they observed:
1. The healthy members of each couple became significantly more depressed and anxious, and reported significantly poorer sleep quality.
In spite of these changes, the healthy individuals remained significantly better than their spouses in all three outcomes.
In other words, they "caught" emotional and sleep disturbances from their spouses, but their symptoms weren't as severe.
2. The healthy members of each couple became more similar to their spouses in oral microbiome.
The healthy individuals more closely mirrored their spouses in the bacterial phyla obtained from oral samples.
A fatal flaw
At first glance, this is an amazing study. It seems to show that depression, anxiety, and insomnia can spread via oral bacteria.
The "fatal flaw" is that it fails to link the psychological changes to the biological ones.
We know that after 180 days, the healthy individuals had become more like their spouses in emotional state and quality of sleep.
We also know that after 180 days, the healthy individuals had become more like their spouses in oral microbiome.
However, there's no evidence that one caused the other.
If you're a newlywed and your spouse is constantly depressed, you're at risk of becoming more gloomy yourself. No surprise there. Daily interactions with a morose, unenthusiastic person are known to be depressing.
Nor is it surprising that your oral microbiome might increasingly resemble that of your new spouse. You're kissing, drinking from the same cups, sharing utensils, etc.
In short, the changes in mood and oral microbiome are correlated – they happen around the same time – but there's no reason to believe one causes the other.
I would call this "bad science" for two reasons:
The researchers present no good evidence for their central claim.
The findings can be easily explained without assuming that oral bacteria transmit mental health problems.
The researchers could have presented evidence bearing on their central claim.
The researchers had all the data they needed to directly test their main hypothesis. For instance, if that hypothesis were correct, we might expect that at day 180, more similar spousal microbiomes would predict greater similarity in depression scores. Analyses like that could've been set up and run in 20 minutes.
(Side note for those with a background in stats: In spite of the study's title, there were no tests of mediation.)
Why this study is valuable
1. Bad science isn't necessarily all bad.
The new study reinforces the idea that a person who's experiencing emotional distress can undermine the mental health of those around them.
Coincidentally, a Virginia Tech study published exactly two weeks ago showed that even dogs can "catch" stress from their owners.
Here the researchers didn't blame oral bacteria, even though some folks do let their dogs kiss them (assuming you call getting licked all over your face being "kissed").
Rather, the claim is that stressed people come home from work still ruminating about their day, they're consequently less attentive to Fido, and Fido notices. This makes sense, though I'm not sure a study was needed to make the case.
In any event, the new study reminds us that concerns about mental health should not be limited to those who experience problems. Support may also be needed for the people (and pets) who are closest to them. (Additional information and resources can be found here.)
2. Bad science helps illustrate what better science looks like.
"Bad" studies provide instructive examples for students, as well as useful reminders for researchers, policymakers, and the general public about best practices in research.
This particular study illustrates the old adage that correlation doesn't prove causation. That's not just an academic point.
Consider, for instance, the claim that MMR vaccines cause Autism Spectrum Disorder (ASD). The conspiracy theorists who push this claim argue that ASD is first detected within a few months after vaccination.
Setting aside data showing that ASD can be detected earlier, the fact that X comes before Y isn't good evidence that X causes Y. Babies often begin walking shortly before the first signs of ASD, but we all know that walking isn't the culprit.
All the same, the CDC is currently planning new research on whether the MMR vaccine causes ASD, in spite of dozens of studies with hundreds of thousands of kids showing no connection.
Your tax dollars would be better spent if the folks behind the CDC's plan, including HHS director RFK Jr., were clearer about the notion that correlation doesn't prove causation – and more receptive to well-established science.
More on bad science, junk science, and good science in a future newsletter.
Thanks for reading!
Thank you for this clearly illustrative example of junk science and faulty reasoning. Articles such as yours that dissect flawed studies can help to promote scientific literacy. Education that fosters critical thinking and scientific reasoning has been inadequate, and that is now reflected in how people are making their political choices.
Association or correlation is not causation. Establishing causation is another animal. However, they can be guides to defining research. Thanks for this thoughtful look at junk/bad science.
As a footnote, I should mention that the gut-brain axis has been connected to gut microbiota. I'll just keep it short here.