Improving Forecasts of Suicide Attempts for Patients with Little Data

TL;DR

This paper introduces Latent Similarity Gaussian Processes (LSGPs), a novel method that improves suicide attempt forecasting for patients with limited data by leveraging similarities among patients, addressing heterogeneity and data scarcity.

Contribution

The paper presents LSGPs, a new modeling approach that captures patient heterogeneity and improves prediction accuracy for patients with little data.

Findings

Outperforms most baseline models in preliminary tests

Leverages patient similarity to enhance predictions

Provides new insights into patient heterogeneity

Abstract

Ecological Momentary Assessment provides real-time data on suicidal thoughts and behaviors, but predicting suicide attempts remains challenging due to their rarity and patient heterogeneity. We show that single models fit to all patients perform poorly, while individualized models improve performance but still overfit to patients with limited data. To address this, we introduce Latent Similarity Gaussian Processes (LSGPs) to capture patient heterogeneity, enabling those with little data to leverage similar patients' trends. Preliminary results show promise: even without kernel-design, we outperform all but one baseline while offering a new understanding of patient similarity.