Understanding Missingness in Time-series Electronic Health Records for Individualized Representation

TL;DR

The insights in this work aim to bridge the gap between theoretical assumptions and practical observations in real-world EHRs and hope this work will open new doors for exploring directions for better representation in predictive modelling for true personalization.

Abstract

With the widespread of machine learning models for healthcare applications, there is increased interest in building applications for personalized medicine. Despite the plethora of proposed research for personalized medicine, very few focus on representing missingness and learning from the missingness patterns in time-series Electronic Health Records (EHR) data. The lack of focus on missingness representation in an individualized way limits the full utilization of machine learning applications towards true personalization. In this brief communication, we highlight new insights into patterns of missingness with real-world examples and implications of missingness in EHRs. The insights in this work aim to bridge the gap between theoretical assumptions and practical observations in real-world EHRs. We hope this work will open new doors for exploring directions for better representation in predictive modelling for true personalization.

Figures (1)

• Figure 1: An example showing the impact of different imputation approaches for time-series EHRs with feature-wise missingness. In (A), highly missing EHRs are shown with population-level statistics and reference tables. In (B), the impact of individualized imputation compared to population-based imputation is shown to affect patient representation and predictive modelling, respectively.