Doctor AI: Predicting Clinical Events via Recurrent Neural Networks

TL;DR

Doctor AI employs recurrent neural networks to analyze electronic health records, accurately predicting future diagnoses and medications, demonstrating high recall and adaptability across institutions.

Contribution

This work introduces a novel RNN-based model for predicting clinical events from EHR data, achieving high accuracy and generalizability across different healthcare settings.

Findings

Achieves up to 79% recall@30 in diagnosis prediction

Outperforms baseline models in predictive accuracy

Successfully adapts across different healthcare institutions

Abstract

Leveraging large historical data in electronic health record (EHR), we developed Doctor AI, a generic predictive model that covers observed medical conditions and medication uses. Doctor AI is a temporal model using recurrent neural networks (RNN) and was developed and applied to longitudinal time stamped EHR data from 260K patients over 8 years. Encounter records (e.g. diagnosis codes, medication codes or procedure codes) were input to RNN to predict (all) the diagnosis and medication categories for a subsequent visit. Doctor AI assesses the history of patients to make multilabel predictions (one label for each diagnosis or medication category). Based on separate blind test set evaluation, Doctor AI can perform differential diagnosis with up to 79% recall@30, significantly higher than several baselines. Moreover, we demonstrate great generalizability of Doctor AI by adapting the resulting models from one institution to another without losing substantial accuracy.