The history, current status, and possible future of precision mental health

In evidence-based mental health practice, decisions must often be made for which there is little or no empirical basis. A common example of this is when there are multiple empirically supported interventions for a person with a given diagnosis, where the aim is to recommend the treatment most likely to be effective for that person. Data obtained from randomized clinical trials allow for the identification of patient characteristics that could be used to match patients to treatments. Historically, researchers have focused on individual moderators, single variables that interact statistically with treatment type, but these have rarely proved powerful enough to inform treatment decisions. Recently, researchers have begun to explore ways in which the use of multivariable algorithms might improve clinical decision-making. Common pitfalls have been identified, including the use of methods that provide overoptimistic estimates of the gains that can be expected from the applications of an algorithm in a clinical setting. It is too early to tell if these efforts will pay off and, if so, how much their use can increase the efficiency and effectiveness of mental health systems. It behooves the field to continue to learn and develop the most powerful methods that can produce generalizable knowledge that will advance the aims of precision mental health.