Automated cardiac arrest detection using a smartwatch: A simulation study to DETECT sudden falls

While survival from out-of-hospital cardiac arrest has markedly improved over
the past decade, for victims of unwitnessed cardiac arrest medical assistance
often comes too late. Automated cardiac arrest detection and alarming would be
an ideal solution to provide early help for this large subset. In the DETECT
project, we aim to develop a smartwatch with the functionality of automated
cardiac arrest detection and alarming. The primary sensor technology used to
detect cardiac arrest is photoplethysmography (PPG). This is an
easy-to-understand technology based on reflection of light to detect absence of
pulsatile flow at the wrist. In the DETECT-1 study, a PPG-based algorithm for
cardiac arrest detection is being developed using data from patients with
short-lasting induced circulatory arrests. However, to come to a reliable
cardiac arrest detection algorithm with low false positives, we need to take
into account additional sensor data to confirm or exclude the presence of
circulatory arrest. Accelerometer sensors measure acceleration and provide
information on human movement. Since a first manifestation of cardiac arrest is
sudden physical collapse without subsequent movement, these sensor data may
provide valuable information to exclude or confirm cardiac arrest. For example,
in case a patient continues to walk, this cannot be a cardiac arrest. In case,
a patient collapses and shows no movement, a cardiac arrest is more likely.

In the DETECT-2 study, we aim to construct an algorithm for detection of
cardiac arrest related falls using wrist-derived accelerometer signals from
simulated sudden falls and non-fall movements. The sensitivity and false
positive rate of the developed algorithm for sudden fall detection will be
studied. Additionally, collected data of the PPG sensor will be used to study
false positive rates of the recently developed first PPG-based cardiac arrest
detection algorithm (DETECT-1).

The DETECT-2 is a Dutch prospective simulation study performed in a controlled
setting.

In this study, the participants will be asked to simulate falls. Risks
associated with these falls could be bruises, abrasions and in the worst case
more serious injuries, such as fractures. To minimize the risk of injuries, the
study will be performed in a controlled setting on a soft surface and under
supervision of trainer. Furthermore, the included subjects will be healthy to
minimize the risk of injury as well.