Everyone is going crazy... do the numbers prove it?

November 20, 2007 6:23 AM

Are there any statistics that suggest mental disorders (from depression to schizophrenia and anywhere in-between) have been growing (in terms of people being diagnosed) over the years?
posted by JaySunSee to Science & Nature (12 answers total)

Not so much statistics as a comprehensive review on recent literature to come out on the subject. Might point you in the right direction.
posted by subajestad at 7:14 AM on November 20, 2007

There is a vast body of literature (and statistics) on how childhood autism diagnosis is increasing. Per usual, however, people bicker about if it is just increased awareness (i.e. ADHD style) causing more childhood diagnoses rather than an actual statistically significant change.
Example of the statistics.
posted by hybridvigor at 8:02 AM on November 20, 2007

This page may be of interest. Cites Cross-National Collaborative Group (JAMA 1992 Dec 2;268(21):3098-105) for an increase in the cumulative lifetime rates of major depression with each successively younger birth cohort at all sites, with the exception of the Hispanic samples. Hmm, that's kind of old -- maybe find more recent material that cites to that study.
posted by ClaudiaCenter at 8:58 AM on November 20, 2007

Yes, absolutely. The National Comorbidity Survey and its followups, the NCS-R, NCS-2, and NCS-A have demonstrated a rising incidence of nearly every major category of psychiatric disorder in the United States over the past two or three decades. The studies were epidemiologically rigorous and things like ascertainment bias have pretty well been ruled out owing to the consistency of methods.
posted by ikkyu2 at 12:01 PM on November 20, 2007

ikkyu's link seems off. It's here. Be sure to check the publications tab. If you don't have access to electronic databases, you can have Kessler's office send you any of the publications (just don't be too greedy) right off the publications page.
posted by toddst at 12:52 PM on November 20, 2007

Yes, there has been an increase in the number of diagnosed mental disorders, but do not make the mistake of confusing the rise of diagnosis with the rise of actual disorders. Ascertainment bias or not, these studies do not take into account the overwhelming evidence of misdiagnosis, overdiagnosis, and a general overemphasis of the act of diagnosis itself in mental health.

It is completely possible that the world isn't any crazier than usual, we're just being told that we are.
posted by tjvis at 1:03 PM on November 20, 2007 [2 favorites]

If you are interested in the question of how a society's understanding of what mental diseases are possible affects the incidence of those mental diseases, you want to look at Ian Hacking's work from the late 1990s.
posted by LobsterMitten at 1:43 PM on November 20, 2007

Response by poster: Hoo boy. I gots me some reading to do.

Thank you everybody for your input (so far).
posted by JaySunSee at 5:37 PM on November 20, 2007

toddst: that website was, in fact, what I was trying (and failing) to link to.
posted by ikkyu2 at 12:58 PM on November 21, 2007

these studies do not take into account the overwhelming evidence of misdiagnosis, overdiagnosis, and a general overemphasis of the act of diagnosis itself in mental health.

Yes, they do. (Incidentally, what you're talking about, that's what ascertainment bias is.) The original NCS was performed in a representative sample population, then replicated a decade later in a larger population using the exact same methodology. It showed a rise in the incidence and prevalence of the disorders under study. People have tried to assail its conclusions on statistical as well as methodological grounds and failed, occasionally publically humiliating themselves in the literature in the process.

So, in fact, you're completely wrong. You might want to learn a little bit about what you're talking about before you start refuting things that (I'm guessing) you've never even heard of before you opened this thread.
posted by ikkyu2 at 1:02 PM on November 21, 2007

Jeez, someone's panties got twisted in a bunch. I've never encountered someone so thoroughly convinced of their own righteousness.
Actually, I do know what I'm talking about, as I had to research this very topic as a psych major. One of the biggest critiques of the NCS was that a substantial portion of respondents classified as cases were statistically insignificant, based on the rather general diagnostic requirements of the DSM-III, and the even more general DSM-IV for the NCS-R. The fact that the NCS-R used the same methodology but with a diagnostic tool that has been significantly changed to broaden the number and scope of disorders should call into question any attempt to find an increase in incidence and prevalence of the population as a whole.
The NCS and its counterparts tend to be lumped together as some sort of definitive guide to the nation's mental health. However, the purpose of the study was to find links between instances of primary and secondary mental health disorders and the probability of using one to predict the other. Comorbidity (the C in NCS) in psychology is a virtual admission that diagnosis is inexact because specific disorders are too difficult to distinguish from one another to be usefully valid. This seems to be quite a major obstacle for many people in taking the study's findings at face value.
That might not be good enough for you, ikkyu2, but I'm incredibly surprised that someone who purports to be an expert on the subject (at least one who has heard of it before opening this thread) has not encountered any of these very valid criticisms before. Either way, please don't personally attack people when you know absolutely nothing about them. It's not very Buddhist of you.
posted by tjvis at 12:17 PM on November 29, 2007

If you cared to spend a minute reading my profile, you'd learn that I'm not Buddhist.

If you cared to spend a couple hours reading reports of the NCS-R, you'll see that the diagnostic data collected was analyzed several ways, including a strict 1-to-1 correspondence with the NCS data (for purposes of the comparison of interest.) The NCS-R tool collected more data than the NCS data, but the R means "replication" and that means that the original NCS tool was used first and completed before the additional data was collected.

I'm not sure how long it would take you to figure out the difference between "clinically insignificant" and "statistically insignificant," but it would certainly help people trying to parse your arguments. I think what you're trying to say is that the original NCS tool overascertained cases that were clinically insignificant and didn't belong in the statistics as cases, which is certainly a defensible criticism. I do not see how that fault, even if present, however, could influence the interpretations of trends in data collected over time with the same tool.
posted by ikkyu2 at 2:07 PM on November 29, 2007