From Symptoms to Function: The PAD-S Decision Matrix—A Transdiagnostic Psychotherapy Algorithm for Severe Mental Illness

Severe mental illness (SMI) is marked by diagnostic heterogeneity, fluctuating risk, and high functional impairment. Across diagnoses, clinicians must repeatedly decide how intensively to intervene, when to slow down, and how to safeguard the therapeutic process—often under conditions of elevated anxiety, rigid defenses, and shame-driven self-attack. This paper proposes the PAD-S decision matrix, a pragmatic psychotherapy decision algorithm that translates moment-to-moment clinical observations into four interoperable process nodes: Progression (PRO), Anxiety/Affect tolerance (ANX), Defense (DEF), and Superego/Shame attack (SUP). PAD-S has a dual meaning: (1) a clinical workflow—Perceive, Assess, Dose, Safeguard—and (2) the node set— Progression, Anxiety, Defense, Superego/shame. The algorithm is designed to be school-neutral, while being compatible with experiential dynamic therapies (EDT/ISTDP) and CBT-oriented skills and exposure work. To improve clinical utility and health-system alignment, PAD-S links process decisions to dimensional severity and functioning using ICD-11 language and ICF-based assessment (Mini-ICF-APP). We outline a documentation-ready set of node definitions, markers, dosing rules, and safety constraints, and illustrate application in two brief vignettes. Finally, we specify a research agenda for reliability, effectiveness, and implementation, and describe how PAD-S-coded psychotherapy transcripts could enable downstream computational analyses (e.g., SPICE) to accelerate psychotherapy process research.