Individualization of therapy in psychiatry: how personalized it may be?

This article provides a critical review of current approaches to personalized psychiatry, analyzing its potential, limitations, and ethical dilemmas. Despite progress in areas such as pharmacogenetics (e.g., the influence of CYP2D6 and CYP2C19 gene polymorphisms on psychotropic drug metabolism) and digital phenotyping (behavioral monitoring through wearable devices and machine learning algorithms), most methods remain within framework of research projects. Regulatory initiatives (FDA, CPIC) are gradually integrating genetic data into clinical recommendations; however, their implementation is limited by insufficient ethnic representation, low reproducibility, and small sample sizes. Digital tools, although capable of predicting relapses in affective or psychotic disorders, face challenges of data interpretation and risks of privacy breaches. Polygenic risk scores and biomarkers (e.g., neuroimaging patterns, cytokine levels) demonstrate limited predictive power at the individual level. Key risks of personalization include fragmented access to care, genetic-based stigmatization, and the replacement of clinical reasoning with algorithmic decision-making. Personalization is therefore justified only when supported by robust evidence (e.g., pharmacogenetics in treatment-resistant depression) and should complement rather than replace clinical practice. Successful integration requires validation of technologies, overcoming ethnic and socioeconomic barriers, and maintaining ethical standards and clinical expertise. Personalized psychiatry should be viewed not as a revolutionary paradigm, but as an evolutionary tool requiring cautious, stepwise implementation into clinical practice.