Transitional models in suicidal thoughts and behaviors - multivariate longitudinal designs.

Background: suicidal thoughts and behaviors (STB) are common in the society, with high mortality and a significant burden for society. Although prevalence is well-known, knowledge on the behavioral dynamics that drive the suicidal process and its timing are unknown.

Aims: to predict STB transitions among psychiatric patients with repeated visits to the ER(i.e. STB incidence among those without prior STB and STB persistence among those with prior STB, as well as the timing of these transitions);and to predict (the timing of) completed suicide among psychiatric patients who repeatedly use the ER of the University Hospital Gasthuisberg, Leuven and Groote Schuur Hospital (GSH) in Cape Town (South Africa)

Method: all patients referred to the emergency room will be recruited between 2020 and 2023 with regard to mental health characteristics, STB, service use and contextual factors (N~5,000). Prediction of transitions and temporal ordening will be done using multivariate and machine-learning techniques that will be used to develop predicted probabilities for a specific STB outcome on the individual patient-level.

Conclusion: this complex methodology will enable us to run valid concentration-of- risk models and yield essential knowledge on which patients are at high risk for the occurence and timing of (future) STB and death by suicide.