Read on Twitter
New paper!
We introduce #GAMs as a tool for analysing medical monitoring data. We also present the idea that GAMs can be used to decompose a medical waveform into individual, physiologically relevant, components.
w/ @ucfagls and @SimonTVistisen https://link.springer.com/article/10.1007/s10877-022-00873-7
We introduce #GAMs as a tool for analysing medical monitoring data. We also present the idea that GAMs can be used to decompose a medical waveform into individual, physiologically relevant, components.
w/ @ucfagls and @SimonTVistisen https://link.springer.com/article/10.1007/s10877-022-00873-7
Example 1: Pulse pressure ("the strength of a heart beat") varies with ventilation (mechanical). We can use a cyclic spline to represent the repeating effect of ventilation on pulse pressure.
This is especially useful when we only have a few beats per ventilation (high respiratory rate, low heart rate), where the pulse pressure variation in any individual ventilation does not represent the "true" pulse pressure variation.
Example 2: A steady central venous pressure (CVP) curve can be decomposed into separate cyclic effects: heart cycle, respiratory cycle, and the interaction of the two.
We can present the sum of these three effects as a plane (mainly for fun).
Sample #Rstats code (tutorial) and sample data is available here: https://github.com/JohannesNE/gam-medical-signals-supplementary
