False Positives

False positives seem to be a big deal to me lately, like they really bother me. I get so many families requesting me to check a lab, run a test, do a scan. While I will do those when I feel appropriate, there's unfortunately little evidence that any tests really help with diagnosis of psychiatric disorders other than substance abuse. Most of the time everything comes back normal except for one teeny little thing and it gets you questioning "could that be a big deal?" Brain scans are the worst right now. There is no psychiatrist who can tell you what your kid has or what med they need based on a brain scan, and if they tell you they can, you are about to fork over several thousand dollars that your insurance company won't cover.

This got me thinking about false positives in typical diagnostic interviewing too. (Full disclosure: I am guilty of diagnosing kids incorrectly, I do it all the time I am sure). In psychiatry, we have a fairly standard interview that's designed to rule in or out several diagnostic categories: mood disorders, anxiety disorders, disorders that start in childhood, developmental disorders, psychotic disorders, etc... In the interview, we are asking all kinds of different questions to get a general idea about the diagnosis. It might take me 2-3 hours over a period of several appointments to come to a complete conclusion about this. But kids who get hospitalized get diagnosed with some pretty serious stuff all the time. The diagnosis du jour seems to be bipolar disorder (cue Jaws theme song). Lemme give a typical scenario- 8 yo gets really pissed off and threatens or harms someone, gets hospitalized, gets put on powerful antipsychotic medication, gets discharged with bipolar disorder. Pay particular attention to the order of those events.

False positives occur when the test comes back positive for a problem at a rate that is more often than the actual incidence of the disorder itself. Let's say there is a testing device called an anger-o-meter and it measures levels of anger (reference e-meter from Scientology) If you have a test that a lot of people show up positive for, then a lot of people are going to look sick. Anger level is often that test. Lots of clinicians have an internal amount of anger that they will consider normal or abnormal, and if a kid appears above that level, then they must have bipolar disorder. The problem is that anger is not specific to bipolar disorder. Using anger level as your measure of the presence of bipolar disorder is like using wetness on your driveway as a measure of whether it rained. That wetness could have come from a thousand different sources: sprinklers were on 5 minutes ago, pack of dogs peed on your drive, pipe burst in yard, someone emptied their cooler from tailgating, water balloon fight, squirt gun war.

Final note: I'm not saying an 8 year old can't have bipolar disorder, it does happen. But it's rare and the diagnosis should be made on the basis of the symptoms that are more precisely indicative of the disorder.