Validation of Respiration Rate Algorithms

Continuous accurate unobtrusive respiratory rate monitoring may lead to improved patient outcomes, as respiratory rate is thought to be a sensitive marker of patient deterioration. Currently systems are not suitable for long term monitoring, particularly in ambulant patients as they are too restrictive. To ensure that our algorithms are suitable for use in a clinical context we need to demonstrate their performance not only in the optimal situation, healthy volunteers at rest, but also in more challenging situations such as where the person being monitored is moving and also in patients who have conditions which may affect their physiology in such a way that the accuracy of the respiration rate estimation may be affected.

No previous study has systematically tested algorithms deriving respiratory rate from either the ECG or the photoplethysmography (PPG) waveforms in a real -world setting.

The algorithms work by looking for changes in intervals between heartbeats and also changes in the sizes of the ECG and PPG waveforms, both of which may be caused by respiration. These changes tend to diminish with increasing age and also conditions which alter the chest movement and cardiac reflexes. Thus it is important to test our algorithms' accuracy in participants exhibiting these conditions. It is also important to ensure that the calculations of respiratory rate are accurate across a range of heart rates and respiratory rates. Our testing covers all these variables.