Building personalized machine learning models using real-time monitoring data to predict idiographic suicidal thoughts

Abstract

Suicide risk is highest immediately after psychiatric hospitalization, but the field lacks methods for identifying which patients are at greatest risk, and when. We built personalized models predicting suicidal thoughts after psychiatric hospital visits (N=89 patients), using ecological momentary assessment (EMA; average EMA responses per participant=311). We built several idiographic models, including baseline autoregressive and elastic net models (using single train/test split) and Gaussian Process (GP) models (using an iterative rolling-forward prediction method). Simple GP models provided the best prediction of suicidal urges (R2average=0.17), outperforming baseline autoregressive (R2average=0.10) and elastic net (R2average=0.06) models. Similarly, simple GP models provided the best prediction of suicidal intent (R2average=0.12) compared to autoregressive (R2average=0.08) and elastic net (R2average=0.04). Here we show that idiographic prediction of suicidal thoughts is possible, though accuracy currently is modest. Building GP models that iteratively update and learn symptom dynamics over time could provide important information to inform development of just-in-time adaptive interventions.