Of course, the question remains, if we could predict schizophrenia, what would we do about it? Should every adolescent “at risk” be treated with antipsychotic medications? What is the evidence that medications forestall or preempt psychosis? If effective, how long would one administer the treatment? And even if predictors are 90% accurate (a high bar), what about the 10% of false-positive results that would lead to unnecessary treatment? All treated patients would be exposed to possible adverse effects of medication.

These questions and many others point out the need for re-thinking interventions in psychiatry—a process that has been under way for some time in other areas of medicine and is now receiving attention in psychiatry.⁸ Three current opportunities may be illustrative. The recent evidence of cognitive deficits in adolescents with schizophrenia years before the first psychotic episode suggests the importance of developing cognitive interventions to preempt the later phases of this illness.^9,10 Recent information technology approaches to cognitive remediation suggest the potential of using the brain’s inherent plasticity to alter circuitry and functioning without medication.¹¹

As a second example, posttraumatic stress disorder (PTSD) responds to cognitive-behavioral therapy (CBT). Early data suggest that CBT following a traumatic event can reduce the incidence of PTSD in those at greatest risk.¹² Third, data from several studies support the efficacy of lithium(Drug information on lithium) for reducing the number of suicides in those with mood disorders. In a meta-analysis of 32 trials, persons treated with lithium showed a 74% reduced risk of death from suicide.¹³

This shift in paradigm not only alters our approach to prevention, it also redefines mental disorders. Currently, mental disorders are diagnosed by late-stage symptoms, such as psychosis. If mental disorders are viewed as chronic illnesses with developmental onset, then we should be able to detect subtle, diagnostic changes in brain and behavioral functioning many years before the most disabling symptoms emerge. This suggests a shift in the definitions of schizophrenia and bipolar disorder that is similar to the change from myocardial infarction to atherosclerotic cardiovascular disease, which links diagnosis to pathophysiology and associated risk factors rather than to symptoms. Moreover, current psychiatric medications, when they are “effective,” are usually palliative and only reduce symptoms. A preemptive approach promises to reduce morbidity and mortality by intervening early, before the full syndrome develops, and re-aligning the trajectory of development so the individual identified as at risk has the greatest opportunity for the best outcome. In this sense, psychiatry can follow the path taken by oncology and cardiology.

What can we expect in the next decade? Is preemption feasible in this time frame? Recent breakthroughs in genomic and imaging technologies may well revolutionize psychiatric diagnosis by adding key elements and techniques for identifying patterns of risk that are amenable to early detection. Defining the genomic risk architecture of these disorders will require much larger studies than any published to date. Similarly, we will probably need more longitudinal, intrasubject studies to identify neural correlates of schizophrenia, mood disorders, and autism. However, there is every indication, from the use of these techniques in other areas of medicine, that we will be defining patterns of individual risk for mental disorders within the next few years in research settings, if not the community clinic.

In the meantime, there is much we can do with known predictors and proven intervention strategies that are in need of broader implementation in the community. We know already that a history of psychosis predicts a subsequent psychotic episode and that a suicide attempt is the highest risk factor fora subsequent completed suicide. Yet, more than 30% of those with a first psychotic episode will discontinue treatment in the first year, increasing their risk of relapse 5-fold, and fewer than 50% of adolescents presenting to emergency departments after a suicide attempt receive any follow-up.^14-16 Surely, we can do better. While research develops transformative information and tools for preemption, we need more effective implementation of currently available treatments to ensure the best outcomes. In the near-term, preventing relapse will have the greatest impact on public health.