B0740
Title: State space model multiple imputation for missing data in non-stationary multivariate time series
Authors: Xiaoxuan Cai - Columbia University (United States) [presenting]
Abstract: Mobile technology provides scalable methods for collecting physiological and behavioural biomarkers in patients' naturalistic settings, as well as opportunities for therapeutic advancements and scientific discoveries regarding the aetiology of psychiatric illness. Continuous data collection yields a new type of data: entangled multivariate time series of outcome, exposure, and covariates. Missing data is a pervasive problem, and ecological momentary assessment (EMA) in psychiatric research via mobile devices is no exception. However, complex data structures of multivariate time series and non-stationarity make missing data a major challenge for proper inference. Most available imputation methods are designed for longitudinal data with limited follow-up times or for stationary time series. A novel data imputation solution is proposed based on the state space model and multiple imputations to properly address missing data in non-stationary multivariate time series. Its advantages are systematically demonstrated over other widely used missing data imputation strategies by evaluating its theoretical properties and empirical performance in simulations of stationary and non-stationary time series subject to various missing mechanisms. The proposed method is applied to investigate the association between digital social interaction and negative mood in a multi-year smartphone observational study of bipolar and schizophrenia patients.
