Monday, March 17, 2014

International Variations in Bipolar Disorder Rates

I was reading a very interesting book by Dr. John J. Ross, Shakespeare's Tremor and Orwell's Cough, and he mentioned a study of bipolar disorder rates across multiple countries found that the U.S. had the highest rates of bipolar disorder.  I dug around a little, and found that the study was this one, which used very large populations (61,392 people) in eleven different countries.  The U.S. had 4.4% of the population reporting bipolar disorder, the highest of any of the countries, with New Zealand just behind us at 3.9%.  The average for all eleven countries was 2.4%.

Unsurprisingly, "subthreshold bipolar" (the kind that makes for Wall Street CEOs, brilliant if difficult computer engineers, university faculty and other professionals) have a similar distribution: 2.4% of the U.S. population, compared to 1.4% for all countries.  Because the tendencies towards bipolar disorder and schizophrenia are genetic in origin, I sometimes wonder how much of American exceptionalism in the economic area over the last couple of centuries might have to do with this.

Another study that is quite interesting, and perhaps suggests caution in comparing the experience of other countries with decriminalization of intoxicants to the U.S. situation is this study in Sweden.  It found persons with bipolar disorder were 6.4x more likely to commit violent crimes than the general population, but this was almost entirely among those with bipolar disorder and substance abuse comorbidity.  Those without substance abuse comorbidity were only 1.3x more likely to commit violent crimes, and even this small difference disappeared when comparing to full siblings not suffering from bipolar disorder.  (My guess is that much of the non-substance abuse bipolar population was coming from homes where other siblings were also violent crimes.)


  1. Ethnic and socioeconomic factors are likely to be more diverse in the US. Our ethnic and socioeconomic composition, while flattening through diffusion, must still have substantial regional tendencies.

    What can justify a claim that a population drawn solely from within a 150 mile radius of Pittsburgh, PA from 1995-1999, says much about the entire US?

    The paper is not made less interesting by this labeling defect, but the entire US cannot be painted with the brush that they used.

  2. You raise an interesting point. I suspect that 150 mile radius of Pittsburgh is really pretty diverse, but not like a sample drawn from four or five different regions.