If you happen to keep up with the media, you’ll no doubt encounter health stories focusing on behaviour and disease. Usually, the reports go along these lines…

“A new study has revealed that doing/not doing X can lead to a higher/lower risk for Y.”

The whole point is to make you, as the audience take a step back and think…


These studies are useful as they can help us determine whether certain actions can lead to subsequent health events. They can also help guide policy.

But, there is a rather unfortunate catch. Sometimes, the recommendations arising from these reports contradict each other. The epitome of this conundrum happens to centre around a drink we all know and many of us love…

coffee(A Cup o’ Joe…)

Over the last few years, we’ve been inundated with reports on the health benefits and risks associated with coffee consumption.

On the good side, two to four cups a day may reduce the risk of Parkinson’s disease, ovarian cancer, liver cancer, colon cancer and depression. If you’re a caffeine fiend (like me), six cups a day could reduce the chances for skin cancer.

Reading this list may make you want to say…

garfield(Never mess with an upset Garfield…)

But before you start purchasing those boxes of beans, there is a flip side. Other studies have revealed distinctly different outcomes. Some suggest more than four cups a day may decrease bone density and increase cholesterol. Then there are studies suggesting this much coffee may increase the risk of death by fifty-six percent.

Looking at these numbers might make change your answer to any java offering to…

lemur(I would rather not die…)

From a policy perspective, this information is useful in developing certain recommendations. You can drink two to four cups a day but don’t go any higher. Makes sense, right?

Yet the public perspective is more inclined to be based on an all or nothing principle.  The compilation of these reports can muddy the waters and lead you to drift from Hmm to…


As to why this type of study divergence happens, there is a simple answer. All these studies are observational. The researchers are only looking at links between certain activities in a population of people and what happens to their health over time. This means the results, no matter how robust, only show association, not cause.

We tend to use a specific saying to stress this point:


While this term is extremely useful in scientific realms, in the public forum, it may be considered yet another example of academic elitism.

To get a real grasp of the importance of this cautionary advice, a more appreciable example would provide a stronger perspective. I originally had thought about going through the coffee craziness, but a recent happening may be far more valuable.

This past weekend, I went to a concert for one of my favorite bands…

(Red Hot Chili Peppers…)

I’ve been a fan of their funky music for decades and have loved almost every album. Not to mention, they are perhaps the only group that regularly uses microbes in their lyrics.

Before the concert, I had reached out to the band on Twitter expressing my deep desire to hear a particular song, “Higher Ground.” Many of you might know it from the original singer…

wonder(Stevie Wonder…)

When the concert began however, the set list appeared to follow a path down which this song would not be included. While I was happy to be at the concert, I was feeling a little disappointed.

Then, about two-thirds into the 90 minute show, something magical happened. An incredible bass riff sounded through the concert hall. If you are not aware of it, just listen to the first few seconds…

(The whole song is worth a listen…)

I arose from my seat and starting singing – well screaming – along with the band. This was a singular moment of joy. But I also had another thought that made this experience even more special…

ididit(I caused the song to be played!)

I mean, I did send them a social media request and they did change their song list to suit me. I had to be the cause…right?!

Well, much to my chagrin, the answer is most likely…

goahead(and say NO…)

As hard as it is for me to accept, my one tweet most likely did not lead to a change in the set list. Sure, it might have happened but there was no evidence or mechanism to suggest I was the cause.

Now, if the member of the band I tweeted sent me a response, then there would be a mechanism in place…

  1. I tweeted.
  2. He saw it.
  3. He changed policy.
  4. I benefited.

But without proof, all I can do is surmise and apply conjecture similar to a…

fishing(You get the idea…)

Although the difference between my story and making recommendations for improved public health is an obvious stretch, I hope this at least provides some perspective on the limited power of observational studies. They are useful but need to be examined as a part of a larger picture. Changing one’s actions based on an individual report may lead to trouble.

I hope this example also will engage you to ask a very simple and yet incredibly important question the next time you see/hear/read one of these studies associating behaviour with health…

mechanism(Is there a mechanism?)

If the answer is no, then maybe it’s best to ignore the recommendations until better evidence comes around.