1 Boring Old Man

Personalized Biological Testing In Psychiatry: Inevitable Reality or Impossible Dream?
Psychiatric Times
By Allen Frances, MD
July 18, 2012

^{At next year’s meeting of The International College of Neuropsychopharmacology [Collegium Internationale Neuro-Psychopharmacologicum, or CINP], the smartest people in psychiatric genetics will present their suggestions and report on progress in applying the techniques of personalized medicine as a whole new approach to psychiatric diagnosis. The need for out-of-the-box thinking is great. Our current diagnostic system is based on fallible and subjective clinical judgments. Progress in developing biological laboratory tests has been frustratingly slow, and except for Alzheimer disease is pretty much at a dead end. Dozens of candidate markers have been offered, but nothing ever seems to replicate.}

^{It has become increasingly clear that each of our psychiatric diagnoses is really a heterogeneous final common pathway with dozens, perhaps hundreds, of different pathogenetic pathways. There is not one way to develop symptoms of schizophrenia or bipolar disorder or autistim or OCD. Instead, there are probably hundreds of different genotypes that can converge on the same phenotype. Explanatory mechanisms do not replicate at least in part because patients may get their symptoms in their own very different ways. Here is how former CINP president and conference co-host Bob Belmaker describes the issue and the conference meant to address it:}

"The small effect size of antidepressants and even antipsychotics in large controlled trials has been the subject of much soul searching. These trials include very heterogeneous patient groups and the mean values may disguise huge individual variability. Personalized medicine is a new frontier of research to determine the best treatment for specific patients. CINP is organizing the first meeting on personalized medicine in psychiatry. Speakers include Tom Insel, Head of NIMH who has made personalized medicine in psychiatry one of his top goals; Jun Wang, head of the Beijing Genomic Institute where the government is pouring millions into drug development via the genome; and Shitij Kapur from the Maudsley Institute who will report on individual responses in schizophrenia."

^{Personalized psychiatric diagnosis has great promise and may be one of the few ways out of the current impasse — the constant flow of group mean studies giving non replicated or barely significant differences. But the path will not be easy and certainly will be slow. This is the painstakingly retail work of many professional lifetimes — akin to the steady, but very slow progress in finding the heterogeneous pathways of breast cancer. And there may be many blind alleys and false findings. If you do enough analyses you can always get positive, but meaningless, results based on chance. The statistical modeling of individual difference is still an art form in its early stages of development and first results will require extremely cautious and skeptical interpretation.}

^{The technique of personalized medicine will work best in situations with a favorable signal/noise ratio—conditions that have a well-demarcated and homogeneous clinical presentation with a preponderant biological contribution and patients who have a classic and severe presentation, an early onset, and many affected relatives. My guess is that the first fruits of the new approach will be in understanding severe, classic, early onset autism, OCD, and bipolar disorder. Even the best established contributions will probably account for just 2% to 3% of the variance — but each of these small steps will be a giant leap compared to our current stalemate in understanding the biological underpinnings of psychopathology. There will be no grand slams in this hard fought game [and no walks] — we must settle for an accumulation of singles.}

The mapping of the human genome is a miraculous advance, but brings new frontiers rather than a new destination. Somewhere down the line, it will make honest men of us when we use words like ‘genetic’ or ‘familial.’ As Dr. Frances says, there’s a hard fought game ahead of us, and the notion that genotypes will direct psychiatric drug choice is a leap unjustified at this moment. None of the attempts to define drug specificity so far have panned out. Algorithms, Sequencing, using two medications simultaneously – all the millions spent so far have produced nada. We don’t even know if these medications are symptomatic or disease-specific when they work, which is not as often as we’d like. We have no evidence that the conditions [MDD] in the form they’re widely diagnosed are hereditary, no evidence that the drug classes select unique patients – in short nothing solid that would help us pick a target population. The track records of the current investigators [iSPOT and EMBARC] is shaky by any criteria. Popularizers are people like Tom Insel and Charlie Nemeroff who seem to me to be ‘break-through freaks’ – enthusiastically jumping on any bandwagon that presents itself. And the whole enterprise is based on money – money for genetic testing, money for drugs, money for research, etc.