Just One Cornetto

It’s a helluva week for cancer stories!  This one’s from the Guardian:

Statins could halve risk of dying from cancer, says major study.  “Dr Ange Wang of the Stanford University School of Medicine, who led the women’s study, said the findings suggest that statins could play an important part in cancer treatment. “We’re definitely very excited by these results,” she said.”

The media just loves this kind of cancer story.

How many “Major Breakthrough!!” stories have you come across?  Or “Major Failings Found!!”?  This story falls into the same trap.  The important word in both the headline and the quote from Dr. Wang is “could”.  Statins could halve the risk…Statins could play an important part…

The thing is, the reported study was not a proper randomised trial.  Instead, the researchers looked at the outcomes for patients with a variety of different diseases, and then looked to see if there was a difference in the death rates of patients who taken statins compared to those who hadn’t.  What they found, was a correlation between taking statins and increased survival rates.  So, not definite proof, but a correlation.

This type of story appears all the time.  Some newspapers are worse than others.  The Daily Mail, in particular seems determined to separate the entire world into things which cause cancer, and things which cure cancer.  Or both:

Daily Mail:  Drinking wine causes cancer

Daily Mail:  Drinking wine prevents cancer

Now, these stories in the Mail cannot both be correct.  Or can they?  Well, in one sense they can.  Taking the Mail stories above, there will be some individuals, with certain types of cancer, where there is a positive correlation between wine consumption and risk of cancer (ie, the more they drink, the more the risk) and others where there is a negative correlation (ie, the less they drink, the less the risk).  But, alternatively, there will be other individuals, with different types of cancer, where there is an inverse correlation between wine consumption and risk of cancer (ie, the more they drink, the less the risk).  So, it is possible for wine to correlate with both an increase and a decrease in the risk of cancer, depending on the patients and depending on the cancer.

But, the thing is, just like in the Guardian’s statins story, all we are talking about here is a correlation.  And an important thing to remember in science is that correlation does not imply causation.  I’m just going to repeat that for you.

Correlation Does Not Imply Causation.

Just because two events happen at the same time, it doesn’t mean that they are connected.  At all.  So, in the examples above, while some studies have found a statistical association between wine consumption and cancer (good or bad), this doesn’t mean that the wine is causing these effects.  Likewise, just because some studies have found a correlation between taking statins and increased survival, this doesn’t prove that there is a causal link.  It may be a contributing factor, but it’s just as likely to be a coincidence.

Let me clarify by way of an example.  In the UK, the amount of ice cream that gets eaten increases dramatically during June-August.  And, also, the amount of professional football games that get played drop significantly during June-August.  Now, how do we explain this remarkable correlation?  What is the link between ice cream and football?  Well, if Correlation = Causation, then there would be an easy, simple explanation:  consumption of ice cream prevents football!  There must be some ingredient in ice cream…some strange, powerful factor…which takes away a footballer’s abilities.  That must be the answer!  When it’s hot and sunny, a footballer is more likely to eat ice cream, therefore his playing ability goes down.  That must be it!  Otherwise, how can you explain the amazing correlation between increased ice cream consumption and fewer professional football games?  Eh?  EH??

But of course, in reality, there is no link.  Correlation is NOT Causation.  It is a coincidence that the football season shuts down during the summer months**, when ice cream consumption is highest.  So, as tempting as it might be to attribute Rangers’ Play-off performances last week to Just One Cornetto too many, I’m afraid that just won’t fly.  Sorry Mr. McCall.

And when it comes to drinking wine, it might increase your risk of cancer … it might decrease your risk.  Or it might be just a big coincidence.  And taking statins might increase the survival rates in some cancer patients.  Or it might not.  The only way to tell for sure is a proper, randomised, double-blinded trial.

(** As an aside, I heard somewhere that the reason that the football season shuts down in the summer is that, at least in England, football was originally linked with cricket.  Cricket games can last, not just for hours at a time, but for days (and days and looong, interminable days).  So, sports clubs played cricket during the summer months, when you get the most daylight.  As the year progressed, and days became shorter, it became harder to play cricket, so the sports clubs would switch to football, which was easier to play in autumn & winter as the games were shorter.  Then, the following year, they would switch back to cricket again in late spring.  So, cricket was played during summer, and football was played from autumn through to the following spring.  And this has remained the tradition, even though the cricket-football link has not existed for decades.  This idea certainly seems logical in England, where cricket is bafflingly popular.  I’m not sure how it would work in Scotland though, where cricket is pretty much ignored.)

ResearchBlogging.org

AG McCluskey (2015). Just One Cornetto Zongo’s Cancer Diaries

Advertisements

7 thoughts on “Just One Cornetto

  1. Pingback: Doobie-ous conclusions | zongo's cancer diaries

  2. Pingback: Tough luck, Stretch! | zongo's cancer diaries

  3. Pingback: Tough luck, Stretch! | zongo's cancer diaries

  4. Pingback: Marginal Gains | zongo's cancer diaries

  5. Pingback: Marginal Gains | zongo's cancer diaries

  6. Pingback: A Meaty Subject… | zongo's cancer diaries

  7. Pingback: Damn Statistics! | zongo's cancer diaries

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s