Modelling Patient Trajectories Using Multimodal Information

TL;DR

This paper presents a multimodal approach to model patient trajectories from electronic health records, combining flexible feature integration and sequence modeling to predict clinical outcomes like readmission and disease progression.

Contribution

The study introduces a novel dual-architecture framework for patient trajectory modeling that handles diverse data types and temporal sequences, evaluated on the MIMIC-III dataset.

Findings

The first architecture effectively models readmission and diagnosis prediction from individual admissions.

Clinical text data did not significantly improve predictions, likely due to insufficient fine-tuning of clinicalBERT.

Sequence-based modeling with sliding windows achieved performance comparable to existing methods.

Abstract

Electronic Health Records (EHRs) aggregate diverse information at the patient level, holding a trajectory representative of the evolution of the patient health status throughout time. Although this information provides context and can be leveraged by physicians to monitor patient health and make more accurate prognoses/diagnoses, patient records can contain information from very long time spans, which combined with the rapid generation rate of medical data makes clinical decision making more complex. Patient trajectory modelling can assist by exploring existing information in a scalable manner, and can contribute in augmenting health care quality by fostering preventive medicine practices. We propose a solution to model patient trajectories that combines different types of information and considers the temporal aspect of clinical data. This solution leverages two different architectures: one supporting flexible sets of input features, to convert patient admissions into dense representations; and a second exploring extracted admission representations in a recurrent-based architecture, where patient trajectories are processed in sub-sequences using a sliding window mechanism. The developed solution was evaluated on two different clinical outcomes, unexpected patient readmission and disease progression, using the publicly available MIMIC-III clinical database. The results obtained demonstrate the potential of the first architecture to model readmission and diagnoses prediction using single patient admissions. While information from clinical text did not show the discriminative power observed in other existing works, this may be explained by the need to fine-tune the clinicalBERT model. Finally, we demonstrate the potential of the sequence-based architecture using a sliding window mechanism to represent the input data, attaining comparable performances to other existing solutions.