The prevalence of mental ill health is increasing, while demographic and productivity shifts in much of the world are likely to significantly diminish the pipeline of skilled practitioners. Further, traditional approaches to diagnosis and treatment selection are prone to under- and over-reporting and often divorced from underlying mechanisms of illness, limiting treatment discovery and implementation. Post-traumatic stress disorder (PTSD) in particular, is one of the more prevalent mental health challenges globally and is commonly associated with diagnostic delay. 

This presentation will discuss the results of an Australian pilot study, an AI driven multi-modal digital assessment paradigm for posttraumatic stress disorder that yielded an AUC of 0.75. Interestingly, the highest performing features in the paradigm were also the least intrusive from a clinical workflow perspective. These included narrative, acoustic and ecological momentary analysis data. 

Using the study as a springboard, opportunities to embed such paradigms across the care landscape, from discovery, screening and diagnostic decision support through to treatment tailoring will be discussed. Ultimately, we argue that the prospect of a more objective assessment approach applicable across settings holds great potential for advancing outcomes for our communities.