Studying Comorbidity Posted on May 2, 2006 Joseph R. Calabrese (bio) compares two approaches to the comorbidity question.

So when it comes to studying comorbid presentations of bipolar disorder, there’s a continuum of methodologic considerations. The oldest, the longest standing, and I would say the old fashioned methodologic consideration is to improve the clarity of the patient population you’re studying by having exclusion criteria that diminishes the variance in the data of the patient population under study. So this is an old idea that you can’t study folks who have alcohol and drug abuse, and what you’ve got to study is just the run of the mill uncomplicated bipolar patient. Well, the problem with that is most of our patients are complicated. So that limits generalizability, so that older methodologic consideration has a dramatic impact, and it decreases generalizability quite a bit. On the other hand, what we believe, and this is problematic as well, is that if you study the real world population of patients, the variance in the data increases substantially. So that, more often than not, translates into larger sample sizes being needed, but we don’t know enough to know what the variance in the real population of bipolar needs to be.

There is no other serious mental illness that has a higher lifetime prevalence of Axis I comorbidity, so this is really quite compelling. This means of all the illnesses, we should be most invested in studying the most common variant of illness, which in this instance is comorbid presentations of bipolar disorder.

« Back to Article