Multivariate prediction of temper outbursts in a sample of youth enriched for irritability using ecological momentary assessment data: A registered report.

Irritability and temper outbursts are among the most common reasons youth are referred for psychiatric assessment and care. Identifying in vivo clinical variables that precede the onset of temper outbursts would provide valuable clinical utility. Here, we provide the rationale for a study testing the performance of a classifier trained to predict temper outbursts in a group of clinically-referred youth presenting with symptoms of irritability and temper outbursts. Due to the large sample sizes needed for multivariate classification studies, here, we demonstrated the feasibility of our approach using a relatively large preliminary dataset. Our preliminary data included digital based event sampling from an existing Ecological Momentary Assessment dataset consisting of n = 54 participants with a total of 932 time points. We used this data to develop a logistic regression-based classifier for predicting the temper outburst prospectively. Our initial evaluation provided encouraging evidence for the possibility of predicting the presence of a temper outburst based on individual's momentary clinical responses (e.g., whether the participant is feeling grouchy, hungry, happy, sad, anxious, tired, etc.) prior to the outburst event, as well as external features (e.g., time of day, day of week). However, due to the risk of false positive discoveries and overfitting, these preliminary results are insufficient to conclusively establish the discovery of predictive rules for irritability in Ecological Momentary Assessment data. To more rigorously assess this classifier, we will collect a large confirmatory set, consisting of at least an additional 20 subjects with an expected total of 400 time points, in which will perform confirmatory analyses of the precision and recall of the classifier already fit using preliminary data. This work will potentially provide the foundation for the identification of features predictive of risk and future development of novel mobile-device-based interventions in youth affected with severe and impairing psychopathology.