Why do people who have cats also seem to be more likely to have depression? Is it because cats are more likely to bite a depressed person? Or is it because of some sort of toxic parasite?
Perhaps because depressed people like cats. Or cats and dogs. Or maybe, really, there’s no link there at all.
But don’t worry… We researchers will describe data and make suppositions that don’t take into account the most obvious of explanation of them all.
The researchers did some data mining on the electronic health records drawn from a population of 1.3 million patients. This should give us a wealth of information, right?
Well, demonstrating that if you pick two random variables (out of hundreds) and find a relationship between them, it may not be telling you anything, the researchers nonetheless stretched themselves thin to write about this relationship:
Overall there were 750 patients with cat bites, 1,108 with dog bites, and approximately 117,000 patients with depression.
The highest depression rate was for patients who had both a dog bite and a cat bite, with nearly half (47.8%) having depression, all of them women.
Depression was found in 41.3% of patients with cat bites and 28.7% of those with dog bites.
Furthermore, 85.5% of those with both cat bites and depression were women, compared to 64.5% of those with dog bites and depression.
Other media outlets have tried their best to turn this into some odd finding… Something that echos the myriad of hypotheses the researchers threw out there to explain this relationship.
But the researchers’ own words sum up the findings nicely:
While the total number of patients with cat bites in our study was relatively small, the consequences of untreated depression can be large. It may be that the relationship between cat bites and human depression is spurious and no true cause-and-effect exists […]
Which is likely exactly the case. Since the researchers only had the electronic health records themselves to draw data from, they were limited to only looking at variables in that data. Dozens of alternative variables and confounding factors that might also explain the relationship were not examined.
Let’s look at an example of why two variables that share a relationship may still tell you very little about either variable. If you examined the purchases of M&Ms at a movie theater, and found out that women with cats were 3x more likely to buy M&Ms than women without cats, would that tell you anything about why — or if — cats influence M&M purchases? The two variables might have a relationship, but still be unrelated and have no direct impact on one another. (What if, for example, most of the women who had cats also had children, and the women were buying the M&Ms not for themselves, but for their children?)
Science is full of these kinds of meaningless relationships. Relationships that we want desperately to make sense in some way, but are probably nothing more that “data coincidences,” affected by third variables not being measured.
In fact, recognizing their data really didn’t shed much new light on this topic, the researchers spent most of their paper discussing the fact that their data really didn’t shed much new light on the topic. In a nearly 5,000 word research paper, over 3,000 words were devoted to the “Discussion” section — an unusually large amount.
In a “Well, maybe it’s this way, or maybe it’s that way” back-and-forth, the researchers note many studies have found that pet owners’ health benefits from pet ownership. Only that it doesn’t: “But not all studies have reached similar conclusions, and the role of pets and human health remains controversial with multiple studies reporting inconclusive result.”
Which, at the end of the day, is a fancy way of admitting that we don’t know much about this relationship — other than to report that in this one study, women who owned cats and had a serious cat bite were also more likely to report depression.
Which to me is nothing more than a data artifact of the high prevalence of (a) cat owners who are women and the fact that (b) more women suffer from depression than men (who are far less likely to be cat owners, and therefore, far less likely to obtain a serious cat bite).
Hanauer, DA, Ramakrishnan, N., Seyfried, LS. (2013). Describing the Relationship between Cat Bites and Human Depression Using Data from an Electronic Health Record . PLOS One. DOI: 10.1371/journal.pone.0070585