Research and Diagnostic Algorithmic Rules (RADAR) for mood disorders, recurrence of illness, suicidal behaviours, and the patient's lifetime trajectory.

Abstract

The top-down Diagnostic and Statistical Manual/International Statistical Classification of Diseases categories of mood disorders are inaccurate, and their dogmatic nature precludes both deductive (as indisputable) and inductive (as top-down) remodelling of case definitions. In trials, psychiatric rating scale scores employed as outcome variables are invalid and rely on folk psychology-like narratives. Using machine learning techniques, we developed a new precision nomothetic model of mood disorders with a recurrence of illness (ROI) index, a new endophenotype class, namely Major Dysmood Disorder (MDMD), characterised by increased ROI, a more severe phenome, and more disabilities. Nonetheless, our previous studies did not compute Research and Diagnostic Algorithmic Rules (RADAR) to diagnose MDMD and score ROI, lifetime (LT), and current suicidal behaviours, as well as the phenome of mood disorders. Here, we provide rules to compute bottom-up RADAR scores for MDMD, ROI, LT and current suicidal ideation and attempts, the phenome of mood disorders, and the lifetime trajectory of mood disorder patients from a family history of mood disorders and substance abuse to adverse childhood experiences, ROI, and the phenome. We also demonstrate how to plot the 12 major scores in a single RADAR graph, which displays all features in a two-dimensional plot. These graphs allow the characteristics of a patient to be displayed as an idiomatic fingerprint, allowing one to estimate the key traits and severity of the illness at a glance. Consequently, biomarker research into mood disorders should use our RADAR scores to examine pan-omics data, which should be used to enlarge our precision models and RADAR graph.