Often studies only give you snapshots of a population – what one set of people looks like at hospitalization. What a different set looks like five years after hospitalization at a different hospital and in a different region. What fifty-year-olds look like right now, and what twenty-five-year-olds look like right now.
We often infer information about the course of a disorder based on who we can pick up at different points in life. But there’s no guarantee that the people we catch when they’re fifty were, at twenty-five, like the twenty-five-year-olds we’re catching now. We might have used recruiting techniques that caught (say) twenty-five-year-olds who are heavy drinkers, and fifty-year-olds who are drinkers now but were teetotalers at that age. There are a lot of ways these problems can come up, and researchers work diligently to do what they can with what is actually feasible to do, and we do our best to check on that knowledge in a variety of ways.
But it’s always good to have extensive longitudinal research to address questions about course of illness. Here’s an article (cite at bottom of entry) reporting on a large longitudinal study of people with bipolar I. I wanted to get ahold of the article to go into more depth, but my university doesn’t have it and it appears to be in either Portuguese or Italian, and since the closest thing I have is some Spanish my translation would be highly questionable.
This is from a decade-long project, the McLean-Harvard First Episode Project & International Consortium for Bipolar Disorder Research which followed people with bipolar disorder and psychotic disorders from their first hospitalization. This abstract only looked at data for bipolar I.
There are several findings I would prefer not to be true, but if we don’t consider the possibility, we can’t plan for them. People usually do not recover fully from their first episode, and they are very likely to have more episodes in the first two years (and to switch from depression to mania or vice versa),
Some conditionally good or bad stuff: Most people have the most problems early on with depression/dysphoria, and they tend to have a worse course. Initial mania or psychosis shows a better prognosis (interestingly enough). Very high rates of suicidal behavior accidents occurred early but not as much later on (this finding is pretty extensively reported). Early substance-use and anxiety go together. Prodromal symptoms (stuff indicating you’re about to have an episode) predicts bipolar disorder better than non-affective psychotic disorders (good for bipolar, bad for others).
Some good stuff: Most people didn’t cycle more and more over time (but if I’m reading the abstract right, they didn’t stick to a single steady cycle length, either). Also, how long people waited and how many episodes they’d had was unrelated to their response to mood stabilizers.
Salvatore, P., et al. (2007). Longitudinal research on bipolar disorders. Epidemiologia e psichiatria sociale, 16(2), 109-17.