Clinical Outcome Prediction from Admission Notes using Self-Supervised Knowledge Integration

TL;DR

This paper introduces a novel approach for predicting clinical outcomes from admission notes by leveraging self-supervised knowledge integration and hierarchical ICD code information, improving predictive performance.

Contribution

It proposes a new admission-to-discharge prediction task, a clinical outcome pre-training method, and a simple way to incorporate ICD code hierarchy into language models.

Findings

Improved prediction accuracy over baselines

Enhanced transferability of models

Identified limitations in handling vital signs

Abstract

Outcome prediction from clinical text can prevent doctors from overlooking possible risks and help hospitals to plan capacities. We simulate patients at admission time, when decision support can be especially valuable, and contribute a novel admission to discharge task with four common outcome prediction targets: Diagnoses at discharge, procedures performed, in-hospital mortality and length-of-stay prediction. The ideal system should infer outcomes based on symptoms, pre-conditions and risk factors of a patient. We evaluate the effectiveness of language models to handle this scenario and propose clinical outcome pre-training to integrate knowledge about patient outcomes from multiple public sources. We further present a simple method to incorporate ICD code hierarchy into the models. We show that our approach improves performance on the outcome tasks against several baselines. A detailed analysis reveals further strengths of the model, including transferability, but also weaknesses such as handling of vital values and inconsistencies in the underlying data.