Prediction of Sepsis in the Emergency Room with Pulse Wave applied Machine Learning

Unidentified deterioration in patients admitted to the emergency department
(ED) results in an increased mortality rate. Sepsis, septic shock and
cardiovascular instability are frequent problems in these patients. (Abnormal)
changes in clinical manifestation of these patients often precede the
deterioration itself. Failure to recognize these changes may result in
inappropriate treatment and increased mortality. Therefore, early
identification of deterioration of the patient may contribute to improve
outcome of deteriorating patients.
The hemodynamic profile of these patients provide information of the state of
the patient and potentially holds predictive information for the prediction of
deterioration. Therefore, in this research, we will conduct a prospective data
collection study at the ED. We will collect the continuous noninvasive arterial
pressure waveform signals with the ClearSight (CS) finger cuff. With these
data, a machine learning framework will be developed to predict sepsis, septic
shock and cardiovascular instability in patients admitted to the ED.

The primary aim of this study is early identification of sepsis (and septic
shock) and cardiovascular instability, based on the hemodynamic profile of ED
patients. Continuous noninvasive arterial pressure waveform signals will be
collected with the ClearSight (CS) finger cuff. In combination with clinical
data from patients* electronic medical record (EMR), a machine learning
framework will be developed for the prediction tasks.

This is a non-randomized prospective observational study. Electronic data
collection of continuous noninvasive arterial pressure waveform signals takes
places with the CS/EV1000/HemoSphere system in patients admitted to the ED.
This study is divided in two phases.
Phase 1:
A pilot phase comprising 200 ED patients. Standard of care is performed and
timing and dosing is left to the judgement of the attending emergency
physician.

Phase 2:
A non-randomized prospective observational data collection study in at least
1000 and maximally 1500 (dependent on the interim analysis) ED patients of with
at least 20% sepsis and septic shock patients. An amendment will be submitted
if there are changes to the protocol.

There are no additional risks or benefits associated with participation. There
are no investigational devices used in this study. There are no additional
risks associated with the use of the CS/EV1000/HemoSphere monitor other than
described in the Instructions for Use. There are also no risks associated with
the study procedures.