Big numbers tell big stories, and I’ve got a hot-off-the-press big story about health, based on the latest big data from Google.
If you’ve ever been sick, or will get sick in the future, this story is about you.
It’s about the connection between your emotional well-being and your physical health.
Let’s begin with some impressive numbers.
Roughly 80% of visits to primary care doctors result from emotional distress, and a recent study demonstrated that workplace stress accounted for up to 38% of differences in longevity across workers in different professions.
In other words, stress kills.
Why is there such an enormous disconnect between the root cause of many illnesses and the way doctors treat patients?
There are probably many reasons, including lack of training in medical school treating stress, lack of time during rushed medical appointments, and sadly, in some cases, lack of financial incentives for doctors to prescribe stress reduction instead of surgery, drugs or other expensive medical procedures.
But, having worked and taught in a medical school, and being engaged to a physician who has explained to me how medical practices really work, I believe there’s a much deeper reason that physicians do not treat chronic stress:
Western medical science hasn’t fully embraced the role that emotional health plays in physical health.
Although some doctors understand the strong linkage between emotional and physical health, mind-body medicine still carries a taint of pseudo-science and “touchy-feely voodoo” in many quarters. One reason is that much of the research in the field is anecdotal or qualitative, lacking the “hard-science” edge of double blind clinical trials or laboratory bench research. To fully accept new clinical findings, many doctors want to see numbers: statistically significant test results, epidemiological data and so forth.
With this in mind, I have been using the new science of "Big Data" to learn if “numbers” show the importance of mental health to physical health. Not just any numbers, but really BIG numbers that are hard to ignore. In the parlance of health statistics, I seek an enormous “N” (N standing for number of samples). The higher the “N,” the stronger the inferences that can be drawn about the robustness of the findings.
I have also been searching for numbers that lend themselves to visual storytelling, since, for doctors as with anyone else, “seeing is believing.”
And Google trends, which tracks the frequency that people search different words or phrases, has proved to be a gold mine of very high N, very visual data, connecting mental well-being to physical well-being.
My Big Data analysis starts with a premise: If a mental condition, such as depression caused or triggered physical illness, you’d expect Google search interest in mental disorders to co-vary with interest in certain physical disorders. People who have an illness, or are close to someone with an illness, presumably “Google” that illness to get more information.
Consider, for example, the regular seasonal variation for searches on colds (blue) and flu (red). Over a ten year period from 2005 to 2015, searches for those respiratory illnesses peak in December and bottom-out in August, year after year. Not proof that people who are sick (or people caring for them) perform searches on diseases that afflict them, but a pretty strong indicator, nonetheless.
Now let’s compare searches for a common mental disorder, depression, with several physical maladies.
The following graphic, which shows Google search trends from 2005 through 2015, for four searches: “depressive disorder,” “fatigue,” “asthma,” and “migraine, ” reveals that there is indeed a strong correlation between a mental illness, depression, and certain disorders thought to have a strong psychological component.
The red lines show peaks of search interest during a year (typically in November and April) while the green lines line up with “troughs” of search interest (typically in August).
Notice that the peaks and troughs of the four different curves occur at the same time year after year, showing strong correlation of searches for depression and the other illness.
The numbers on which these search trends are based are much larger than for your average biomedical experiment. “Depressive disorder” averages about 6 million searches per month, “Migraine” 2.2 million, “Fatigue”, 1.7 million and “Asthma,” 1.2 million.
Those are just the monthly searches. Add up ten years worth of searches and you get some truly high “N’s!”
One interpretation of these mental-physical correlations is that people who are suffering from depression are simultaneously experiencing physical symptoms driven by their mental state.
OK. Correlations such as these do not prove that depression causes fatigue, asthma, or migraine. It could be the other way around: physical illness might cause people to get depressed. Or, perhaps some third factor causes depression and physical illness to co-vary.
But if physical illness were the primary cause of the blues, not vice versa, then you might expect that other mental conditions, such as bipolar disorder, that feature oscillating emotional high’s, might not correlate so strongly with physical illness.
But the numbers tell a different story, as shown below. Notice that the peaks and troughs for bipolar line up with the other maladies in exactly the same way as depressive disorder.
Still, physical illness might trigger even the up-swings in bipolar episodes.
But schizophrenia? Check out this graphic.
The correlations for peaks and troughs for schizophrenia with the other illnesses is again strong, although notice that the yearly up-down pattern for schizophrenia is even more pronounced than for bipolar.
Finally, getting back to where we started with the discussion of chronic stress, here’s an analysis showing how chronic stress searches relate to these same illnesses.
Yep. The peaks and troughs line up in the same way as before.
It turns out that search interest in many mental and physical disorders co-vary in this same pattern, although not always with the extreme peaks and valleys of depression. These illnesses include Cancer, Diabetes, Epilepsy…even Alzheimer’s disease.
Again, these data do not show that mental problems directly cause physical problems. Additional research is needed to explore, for example, the “co-morbidity” of mental and physical illness. One recent study, for example, found that that 65% of patients with mental disorders also suffer from one or more “medical” problems . But which symptoms emerge first in these patients, mental or physical? The answer could shed light on what typically causes what.
Even if it turns out that mental illness predominantly grows out of physical illness, not the other way around, such statistics suggest a very strong mind-body linkage, that, at a minimum, indicates that mental health should figure prominently in treating physical illness.
And vice versa. Surveys of mental health care are show that psychiatrists often under-appreciate and under-diagnose physical causes or triggers of emotional and behavioral problems. Dr. Ryan Hall of the University of Central Florida has found that as many as 47% of problems for which patients seek psychiatric treatment have a physical cause.
So, whether mental problems cause physical diseases, or physical diseases cause mental problems (or both) it appears likely that psychiatrists are underestimating the importance of medical illnesses, and primary care doctors are under-estimating the importance of mental problems.
This means if you’re seeking a physician for a medical problem, try to find one who is open to exploring (and treating) emotional distress. Conversely, if you seek help for an emotional or behavioral problem, get a thorough physical first!
From a health policy perspective, it also suggests that the U.S. might sharply reduce its health care costs if primary care doctors focused more on targeting the deep emotional causes of physical disease, rather than treating surface symptoms with expensive drugs, medical procedures and surgery.
One last point. To control for the possibility that all Google searches, health related or not, follow the cyclical peaks and valleys shown so far, I tried some random, non-health terms.
Here’s the result.
No real correlation of yearly peaks and troughs with generic search terms such as Dog, Banana and Music.
From these and other data, it does not appear that people increase and decrease their searches generally in the same pattern as for many illnesses. (Incidentally, searches for “cold” and “flu” also do not correlate strongly with searches for mental conditions.)
Now it’s entirely possible that Google search volume for mental and physical disease isn’t all that meaningful, because people search for all kinds of reasons, including term paper research, marketing analysis and idle curiosity.
But where there’s smoke, there’s often fire, and I predict that the field of mind-body medicine will heat up in a hurry!
I thank Chris Gilbert MD, PhD for her invaluable insights and assistance during both the research and writing phases of this project. She is the physician who called my attention to the importance of exploring emotional causes of “medical” problems.