Title
At the Implementation Cliff: Precision Psychiatry and the Limits of Clinical-Level Intervention (co-authored with Roshan Rane and Kerstin Ritter)
Abstract
Mental disorders account for approximately 16% of the global burden of disease. Projections based on WHO data indicate an escalation across the Global South by 2050. The high global burden of disease can be attributed, in part, to limitations in the current social-structural, clinical, and conceptual approaches to mental health. AI-based precision psychiatry promises to improve psychiatry through personalised treatment, diagnosis, and prognosis, although systematic reviews report an implementation failure rate exceeding 99%.
We argue that the late incorporation of implementation determinants in the AI-based precision psychiatry development pipeline results in design decisions that insufficiently align with real-world deployment contexts. We contend this is the primary driver of implementation failure.
In this article, we introduce a three-layered conceptual framework, comprising the conceptual, social-structural, and clinical layers, to describe the ecosystem within which AI-based precision psychiatry models are deployed. We then describe the prevailing development pathway for these models and show how implementation is currently treated as a final stage, rather than an integral part, of the development process. Drawing on the Global South as an illustrative case, we examine how this late positioning of implementation leads to critical contextual factors being overlooked during model development, increasing the risk that models will be poorly suited to the settings in which they are intended to be used.
We conclude that successful implementation requires a reconfiguration of the prevailing development pipeline to incorporate implementation determinants from the outset, through public engagement, co-design, and locally grounded validation practices.
About Stella
Stella Namuganza holds a bachelor’s degree in nursing and a master’s degree in clinical research. She is currently pursuing a PhD on the ethical, legal, and social aspects (ELSA) of digital health technologies. She works as a Research Associate at the University of Tübingen with the Research Hub Neuroethics (RHUNE) project at the Hertie Institute for AI in Brain Health. Her research employs meta-research approaches to investigate ELSA issues at the intersection of artificial intelligence and neuroscience, with a particular focus on the responsible development and implementation of emerging digital health technologies.