18 February 2014

Attaining national happiness through chemistry

Apparently, Scandinavia is big in the UK right now.  (Something "foreign" is always big in the UK, it seems.)  And when something is big, the backlash will be along very soon, as exemplified by this Guardian article that I came across last week.  So far, so predictable.  But while I was reading the introductory section of that article, before getting to the dissection of the dark side of each individual Nordic country, this stood out:
[T]he Danes ... claim to be the happiest people in the world, but why no mention of the fact they are second only to Iceland when it comes to consuming anti-depressants?
Hold on a minute.  When those international happiness/wellbeing surveys come out each year, Denmark is always pretty close to the top.  So I checked a couple of surveys of happiness (well-being, etc.).  The United Nations World Happiness Report 2013 ranks Denmark #1.  The OECD Better Life Index (BLI) 2013 rather coyly does not give an immediate overall ranking, but by taking the default option of weighting all the various factors equally, you get a list headed by Australia, Sweden, and Canada, with Denmark in seventh place (still not too shabby).  The OECD report adds helpful commentary ; for example, here you can read that "Denmark, Iceland and Japan feel the most positive in the OECD area, while Turkey, Estonia and Hungary show lower levels of happiness."

So what's with the antidepressants?  Well, it turns out that the OECD has been researching that as well.  Here is a chart listing antidepressant consumption in 2011, in standardised doses per head of population per day, for 23 OECD member states.  Let's see.  Denmark is pretty near the top.  (It's not second, as mentioned in the Guardian article above, because the author of that piece was using the 2010 chart.)  And the other top consumers?  Iceland first (remember, Iceland is among the "most positive [countries] in the OECD area"), followed by Australia, Canada, and (after Denmark) Sweden.  That's right: the top three countries for happiness according to the UN are among the top five consumers of antidepressants in the OECD's survey(*).  And those countries showing "lower levels of happiness"?  Two of the three (Estonia and Hungary) are on the antidepressant list - right near the bottom.  Perhaps they'd be happier if they just took some more pills?

I decided to see if I could apply a little science here.  I wanted to examine more closely the relationship between antidepressant consumption and happiness/wellbeing/etc. So I built a dataset (available on request) with a rank-order number for each of the 23 countries in the antidepressant survey, on each of several measures.  Then I asked my trusty computer to give me the correlation (Spearman's rho) between consumption of antidepressants and each of these measures.  Here are the results:
Measure Correlation with antidepressant consumption (Spearman's rho) p
UN World Happiness Report 2013 .590 .003
OECD BLI 2013 - Life Satisfaction .621 .002
OECD BLI 2013 - Self-reported health .730 .000
OECD BLI 2013 - Educational attainment -.077 .727
OECD BLI 2013 - Leisure time .106 .631
OECD BLI 2013 - Overall (equal weighting) .653 .001
For the uninitiated: the first three lines, and the last line, show substantial and statistically significant correlations between the item concerned and antidepressant usage. The fourth and fifth lines show no significant correlation. Overall summary: countries with high antidepressant usage also report high happiness/life satisfaction/health.

Note that this is not a case of "everything being correlated with everything else" (as the great Paul Meehl put it).  Only certain measures from the OECD survey are significantly correlated with antidepressant consumption.  (I encourage you to explore the measures that I didn't include.)

Ah, I hear you say, but this is only correlational.  When two variables, A and B, are correlated, there are usually several possible explanations.  A might cause B, B might cause A, or A and B might be caused by C.  So in this case, consuming antidepressants might make people feel happy and healthy; or, being happy and healthy might make people consume antidepressants; or, some other social factor might cause people to consume antidepressants and report that they feel happy and healthy.  I'll let you decide which of those sounds plausible to you.

Now, what does this prove?  Probably not very much; I'm not going to make any truth claims on the basis of some cute numbers.  After all, it's been "shown" that autism correlates better than .99 with sales of organic food.  But here's a thought experiment for you: Imagine what the positive psychology people would be telling us if the results had been the other way around --- that is, if Australia and Denmark and Canada had the lowest levels of antidepressant consumption.  Do you think it's just remotely possible that we might have heard something about that by now?

(*) The OECD has 34 member states, of which some, such as the USA and Switzerland, do not appear in the antidepressant consumption report.  All correlations reported in this post are based on comparisons in rank order among the 23 countries for which antidepressant consumption data are available.

Antidepressant consumption data here [XLS]
UN World Happiness Report 2013 here
OECD Better Life Index 2013 data here

[ Edit 2014-02-20: fixed the figures in the tables, following the discovery of a couple of minor typos in the data. Correlation of UN Happiness Report changed from .595 to .590. Correlation of OECD Better Life Index 2013 - Life Satisfaction changed from .617 to .621. Correlation of OECD Better Life Index 2013 - Self-reported Health changed from .734 to .730. ]