Predicting and interpreting healthcare trajectories from irregularly collected sequential patient data using AMITA

Abstract

In personalized predictive medicine, accurately modeling a patient's illness and care processes is essential, given their inherent long-term temporal dependencies. However, Electronic Health Records (EHRs) often contain episodic and irregularly timed data, resulting from patients' sporadic hospital admissions, leading to unique patterns for each hospital stay. Consequently, constructing a personalized predictive model requires careful consideration of these factors to accurately capture the patient's health journey and assist in clinical decision-making.

Long Short-Term Memory (LSTM) is an effective model for handling sequential data, such as EHRs, but it encounters two major limitations when applied to EHRs: it is unable to interpret the prediction results and it ignores the irregular time intervals between consecutive events. To tackle these limitations, we present a novel deep dynamic memory neural network called Adaptive Multi-Way Interpretable Time-Aware LSTM for irregularly collected sequential data “AMITA”. The primary objective of AMITA is to leverage medical records, memorize illness trajectories and care processes, estimate current illness states, and predict future risks, thereby providing a high level of precision and predictive power.

To enhance its capabilities, AMITA extends the standard LSTM model in two key ways. Firstly, it incorporates frequency measurement and the most recent observation to enhance personalized predictive modeling of patient illnesses, enabling a more accurate understanding of the patient's condition. Secondly, it parameterizes the cell state to handle irregular timing effectively, utilizing both elapsed times and a frequency-based decay factor, which considers both measurement frequency and contextual information. Furthermore, the model capitalizes on both to comprehend the impact of interventions on the course of illness on the cell state, facilitating the memorization of illness courses and improving its ability to capture the temporal dynamics of healthcare data, accommodating variations and irregularities in event and observation timing.

The effectiveness of our proposed model is validated through empirical experiments conducted on two real-world clinical datasets. The results demonstrate the superiority of AMITA over current state-of-the-art models and other robust baselines, showcasing its potential in advancing personalized predictive medicine by offering a more accurate and comprehensive approach to modeling patient health trajectories.