Furious Seasons

September 13, 2007

2005 Suicide Data Shows Decrease, SSRI Debate Might Follow

Remember that last week there was much news around 2004's suicide data, which showed an uptick, and how in the mind of many psych docs quoted in the press a black box warning added to anti-depressants in 2004 was to blame for the increase in suicides? Well, yesterday the National Center For Health Statistics (a branch of the CDC) released preliminary death data for 2005 (see Table B of report here).

The suicide rate dropped from 10.9 per 100,000 in 2004 to 10.6 per 100,000 in 2005, a decrease of about 3 percent--still unacceptably high. 2005 was the first full year of the black box warning of possible suicides and suicidal ideation, which technically only applied to ages 0 to 14 at the time (it's been expanded through 24 year olds more recently). The new data won't be broken out by age groups, genders and race until next summer, so for now it isn't possible to say how the 2004 and 2005 suicide data compared.

But something tells me that the less-use-of-SSRIs-causes-suicides argument heaved to by many in the psych field just took a beating. That may explain the press silence on this new data so far, since it would upset what many in the media, especially the Associated Press, gave voice to last week. In fact, in an AP article on the new data, reporter Mike Stobbe doesn't even mention the decrease in suicides while last week his colleague Greg Bluestein made it sound as if telling Americans SSRIs have problems lead to certain death. What amuses me greatly is that both reporters datelined their articles, which ran five days apart, Atlanta. That means they both likely work in the AP's Atlanta bureau, so for the AP not to mention the drop in suicides after the ink they expended last week strikes me as odd. Or really lazy.

I'm sure some other smart reporter will figure all of this out in the meantime and get a nice article out of it. Hopefully, he or she will understand that the evidence on SSRIs and suicide/suicidality is quite mixed and say so in print, regardless of how displeasing that argument might be for defenders of the SSRI faith and regardless of how much it annoys some advocates in the health care world who would prefer that complete information about the side effects and injuries sometimes caused by these drugs be kept from the public because the public cannot handle the truth.

Based on 2005's data, I'd say the public is doing just fine with the truth, however mixed it might be.

BTW, Psych Central has an interesting post on just how useful suicide hotlines are or aren't.

Posted by Philip Dawdy at September 13, 2007 12:05 AM

del.icio.us Digg it reddit

Comments

I used a suicide hotline once. It was very helpful. They had me on the phone for almost half hour and told me I could call back anytime.

I never have, but I have referred people to them as recently as a month ago.

Posted by: susan at September 13, 2007 02:26 AM

I called one and was told to go to a movie or call a friend. It was New Year's Eve I ended up having to explain no one I knew was home. Was when my daughter turned 18 and was sent to an adult psych unit, it was the worst night of my life and I think because I could tell the person had no idea how to talk to me, was the actual reason I was distracted. It was so absurd to me that they were clueless, it kept me on the phone 45 minutes.The person couldnt believe people were not around for me to call that I knew. Why the heck else would I have called a hotline? I was also never asked what I was going to do when I hung up the phone, and I think I should have been.

Posted by: Stephany at September 13, 2007 06:14 AM

Odd, the Atlanta Journal Constitution didn't report that story even though the CDC is here in Atlanta.

Posted by: Sally at September 13, 2007 06:36 AM

The problem is that 3% is within the margin of error that the preliminary statistics usually are off by. If you look closer to the end of that document (p 34), you'll see that the preliminary stats were off by about 2.5% in 2004, 2.7% in 2003 and 3.2% in 2002 for suicide.

However, if you look at age-specific groups -- where people expressed the most concern in the 15-24 age group -- then you see a larger discrepancy.

Specifically, the 2005 prelim. stats show a 9.8 rate compared with the 10.3 rate from 2004. Even if that rate is off by 3%, it's still going to show a decrease in this all-important age group.

At the end of the day for all age groups and overall numbers, you have a slight decrease from 2002 to 2003 (31,655 to 31,484), a fairly large increase from 2003 to 2004 (31,484 to 32,439), and a slight decrease from 2004 to 2005 (32,439 to 31,769). If you add the 3% error margin to the 2005 number, the numbers will be pretty similar from 2004 to 2005. Since the population continues to grow, however, the rate actually goes down.

And this is the problem from reading too much into statistics.

Posted by: John Grohol at September 13, 2007 06:47 AM

From my own blog:

As seen a Psychiatric Services News Report of the American Psychiatric Association

"In 2003, U.S. physicians wrote 15 million antidepressant prescriptions for patients under age 18, according to FDA data. In the first six months of 2004, antidepressant prescriptions for children increased by almost 8 percent, despite the new drug labeling."

The CDC reports an increase in youth suicide from 2003 to 2004 and some "experts" are blaming this increase on the decline in antidepressant use among youth. However, the Black Box warning was not approved until September 2004 and, even then, it took months before the use of antidepressants among youth began to decline.

Data Courtesy of the SSRI Stories Website

End result? The experts are blaming an increase in suicides on a decrease of antidepressants that did not take place.

Posted by: Sickmind Fraud at September 13, 2007 07:44 AM

There is some data that runs counter to the Psychiatric Services website linked above. Regardless, one should be cautious before reading much into a correlation, especially at such a large population level. Of course, some researchers and members of the media have been very non-cautious with their inferences from correlational data on the topic (i.e., based on correlational data over a limited time frame, they concluded decreased SSRI usage CAUSED more suicides), which I think is unfortunate.

I think Dr. Grohol's points are also well-taken. To quote him, "And this is the problem from reading too much into statistics" -- in this case, I agree.

Posted by: CL Psy at September 13, 2007 03:02 PM

I think everyone in the news media (as well as us) jumped the gun a bit by reporting the correlation as though we could read something important into the data.

For substantive policy decisions at the FDA level, I'd want at least 2-3 years worth of reliable data and trending to show something significant is going on. A one-year difference in the data is probably not enough to draw good conclusions from.

Posted by: John Grohol at September 13, 2007 03:59 PM