Sunday, March 14, 2010

Cognitive Dissonance - When Your Brain Won't Let You See the Facts

Just a quick follow-up to a blog piece I published last week. The piece was a response to a totally irresponsible Newsweek cover story, entitled, Why Antidepressants Don’t Work.

The headline may as well have said, “Flush All Your Meds Down the Toilet Right Now.” No doubt, a lot of people did, and no doubt some of them are paying in full measure.

The hook for the Newsweek piece was a meta-analysis published earlier this year in JAMA. According to Newsweek:

In an analysis of six large experiments in which, as usual, depressed patients received either a placebo or an active drug, the true drug effect—that is, in addition to the placebo effect—was "nonexistent to negligible" in patients with mild, moderate, and even severe depression.

Earlier, Benedict Carey in the NY Times reported:

Some widely prescribed drugs for depression provide relief in extreme cases but are no more effective than placebo pills for most patients, according to a new analysis released Tuesday.

Clearly, Newsweek was riffing off the NY Times. Based on these accounts, I naturally assumed that JAMA had published a study highly critical of antidepressants. But, then again, why would JAMA publish this kind of article at all? This is hardly the MO of medical journals, which can best be described as friends of Pharma.

So I went to the JAMA piece and read it. Or, to be more precise, I misread it. Yes, the JAMA study was highly flawed for all the reasons I stated in my blog piece, but this observation I made was completely wrong:

Medical journals are notorious for publishing industry propaganda disguised as research. This one went the other way. 

No, it didn’t. The JAMA piece was actually strongly supportive of antidepressant treatment. It was only days later that a light went off in my head. Here’s my best explanation for what happened:

The NY Times and Newsweek accounts led me to assume that I would be reading a critical study. Although I was highly skeptical of the NY Times and Newsweek accounts (as there were obvious rebuttals they didn’t report), I nevertheless went to the JAMA piece with certain filters already in place in my brain. In other words, my brain was likely to screen out any information that contradicted my operating assumption. The experts refer to this as “cognitive dissonance,” and we see this happening all the time in daily life.

Politics is a prime example. Republican brains filter information in a way that make them think their shit don't stink, even when the facts blatantly contradict their views. The brains of Democrats are wired the same way.

Of all things, Benedict Carey authored a 2005 NY Times piece related to this, based on a study that suggests our genes may influence our gut-level responses to political and social issues. Ironically, his article may explain his own egregious serial misreporting on mental illness (see, for instance, my blog piece on HealthCentral, Antipsychotics - the NY Times Gets it Right - and Wrong).

Newsweek led off with two meta-analyses conducted by Irving Kirsch of the University of Connecticut. Using the FDA database that included unpublished drug trials, Kirsch found that antidepressants performed only minimally better at best against placebos. There are various ways of disputing Kirsch’s conclusions, but no one has done it by coming up with a better meta-analysis.

The JAMA meta-analysis was a direct response to Kirsch (and another supporting study). According to the JAMA piece:

One limitation to these meta-analyses is the restricted range of baseline severity scores included in their constituent studies.

In other words, the authors (Jay Fournier et al from the University of Pennsylvania) thought that the inclusion in these trials of patients with less than severe depression may have polluted the study sample. So, they set out to do a more refined meta-analysis, and judging from the article, seemed very happy with the result:

For patients with very severe depression, the benefit of medications over placebo is substantial.

See? goes the thinking. Antidepressants DO work. The catch is they don’t work for everyone. To tease out the efficacy data the authors needed for severe depression, they were left with no choice but to make this major concession concerning the benefit of antidepressants for depressions less extreme:

True drug effects ... were nonexistent to negligible among depressed patients with mild, moderate, and even severe baseline symptoms ...

The authors thought they had successfully rebutted Kirsch (though conceding various points of agreement), and in the process had defended the efficacy of antidepressants. The NY Times and Newsweek saw it another way.

This, in turn, affected how I would interpret the article. My eyes saw one thing (an article defending antidepressants), but my brain told me I was seeing something else (an article critical of antidepressants). My factual error fortunately did not effect my analysis. But it serves as a sober reminder of how careful we all need to be. The phenomenon of cognitive dissonance, as much as any mental illness, can take over our brains and rob us of our ability to reason.

No comments: