Utilizing machine learning to improve clinical trial design for acute respiratory distress syndrome

E Schwager,G Boverman,S Schiffer,B Gross,P Boehme,H Truebel,J J Frassica,M Xu-Wilson,K Jansson,Y Chang,A Rahman

doi:10.1038/s41746-021-00505-5

Abstract

Heterogeneous patient populations, complex pharmacology and low recruitment rates in the Intensive Care Unit (ICU) have led to the failure of many clinical trials. Recently, machine learning (ML) emerged as a new technology to process and identify big data relationships, enabling a new era in clinical trial design. In this study, we designed a ML model for predictively stratifying acute respiratory distress syndrome (ARDS) patients, ultimately reducing the required number of patients by increasing statistical power through cohort homogeneity. From the Philips eICU Research Institute (eRI) database, no less than 51,555 ARDS patients were extracted. We defined three subpopulations by outcome: (1) rapid death, (2) spontaneous recovery, and (3) long-stay patients. A retrospective univariate analysis identified highly predictive variables for each outcome. All 220 variables were used to determine the most accurate and generalizable model to predict long-stay patients. Multiclass gradient boosting was identified as the best-performing ML model. Whereas alterations in pH, bicarbonate or lactate proved to be strong predictors for rapid death in the univariate analysis, only the multivariate ML model was able to reliably differentiate the disease course of the long-stay outcome population (AUC of 0.77). We demonstrate the feasibility of prospective patient stratification using ML algorithms in the by far largest ARDS cohort reported to date. Our algorithm can identify patients with sufficiently long ARDS episodes to allow time for patients to respond to therapy, increasing statistical power. Further, early enrollment alerts may increase recruitment rate.

Highlights

Great progress in pharmaceutical therapies has been achieved over the last decades, some diseases, especially in the intensive care unit (ICU), are still characterized by high mortality, and efficient therapies would be well-received[1]
From 3,180,903 ICU stays with respiratory charting information contained in the Philips’ eICU Research Institute database, 51,555 ICU patients were identified as having acute respiratory distress syndrome (ARDS) defined by low oxygenation, respiratory failure with mechanical ventilation in the absence of congestive heart failure (Fig. 2; see “Methods” section)
Hospital mortality among ICU patients diagnosed with severe ARDS was 39%

Summary

Introduction

Great progress in pharmaceutical therapies has been achieved over the last decades, some diseases, especially in the intensive care unit (ICU), are still characterized by high mortality, and efficient therapies would be well-received[1]. Clinical trials in the ICU are highly complex and prone to failure. A prominent example is acute respiratory distress syndrome (ARDS). It is associated with a mortality of up to 40% and significantly reduced quality of life among survivors[2]. Three factors stand out as reasons why so many trials in ARDS have failed, despite promising preclinical results, observational data, and clinical expert knowledge4: 1. Different aetiologies for ARDS, combined with diverse patient comorbidities, lead to different disease phenotypes that may require different treatment[4]. ICUs vary in their standard of care, making it difficult to disentangle the effects of ICU practice and proposed therapies[1]

Objectives

Results

Conclusion