Exploratory Hidden Markov Factor Models for Longitudinal Mobile Health Data: Application to Adverse Posttraumatic Neuropsychiatric Sequelae

• Activity Number: 281 - New Methods with Applications in Mental Health Statistics
• Type: Contributed
• Date/Time: Tuesday, August 9, 2022 : 10:30 AM to 12:20 PM
• Sponsor: Mental Health Statistics Section
• Author(s): Lin Ge* and Xinming An and Donglin Zeng and Samuel McLean and Ronald Kessler and Rui Song
• Companies: North Carolina State University and The University of North Carolina at Chapel Hill and University of North Carolina and The University of North Carolina at Chapel Hill and Harvard Medical School and North Carolina State University
• Keywords: Continuous-Time Hidden Markov Model; Multinomial Logistic Model; Log-Linear Model; Multivariate Longitudinal Data; Mental Health; AURORA

Abstract

Adverse Posttraumatic Neuropsychiatric Sequelae (APNS) are common after traumatic events and cause burdens for society. Many studies have investigated the challenges in diagnosing APNS symptoms. However, progress has been limited by the subjective nature of traditional measures. This study is motivated by the objective mobile device data collected from the Advancing Understanding of RecOvery afteR traumA (AURORA) study. We develop both discrete-time and continuous-time exploratory hidden Markov factor models to model the dynamic psychological conditions of individuals with either regular or irregular measurements. The proposed models extend the conventional hidden Markov models to allow high-dimensional data and feature-based nonhomogeneous transition probability. To find the maximum likelihood estimates, we develop a stabilized expectation-maximization algorithm with initialization strategies. Simulation studies are carried out to assess the performance of parameter estimation and model selection. Finally, an application to the AURORA data is conducted, which captures the relationships between heart rate variability, activity, and APNS consistent with existing literature.