pellicle
Professional Dingbat, Guru and Merkintologist
Well due to a combination of "silly season" - traveling around - not being sure what day it is (due to the public holidays on mid-weeks) I managed to take my dose of warfarin twice (yes yes ... I know ...)
So I took 14mg on Thursday evening instead of 7mg.
I decided to take this opportunity to see how my model performed on "double" dose rather than missed dose.
I took my INR within an hour of taking the 14mg (so I had a working baseline) and observed what would happen.
As it was panning out my INR had been creeping up to the 3.1 area (tested previously on Saturday) and I'd been slowly (in 0.5mg adjustments) been (attempting) winding it back down to lower numbers only to find that it was 3.4 on Thursday (just after taking the pills).
I had been anticipating a swing back down (as it normally peaks about 3 ish and returns back to 2.3ish in a sort of 3 week sinewave.
I decided that I'd drop the dose for Friday (meaning I took 0mg) and then took 4mg on Saturday and will take 6 tonight.
Below is the graph of what my model produces (orange line) against the actual measured data (yes, I've been measuring daily).
To my eye te differences are that my model predicts a slightly higher INR (and if I'd measured at different times I may indeed have observed that) but its not "clinically significant" (as the jargon goes). It does predict a slightly lower dump than I've measured (2.5 measured, 1.8 predicted) but I put that down to the fact that my model does not have the capacity to deal with the inevitable "buffering" of all the stuff involved.
None the less I feel that I'm getting close to making a publishable software (and my friends are encouraging me to publish my current work in a Journal).
Happy New Year to everyone
So I took 14mg on Thursday evening instead of 7mg.
I decided to take this opportunity to see how my model performed on "double" dose rather than missed dose.
I took my INR within an hour of taking the 14mg (so I had a working baseline) and observed what would happen.
As it was panning out my INR had been creeping up to the 3.1 area (tested previously on Saturday) and I'd been slowly (in 0.5mg adjustments) been (attempting) winding it back down to lower numbers only to find that it was 3.4 on Thursday (just after taking the pills).
I had been anticipating a swing back down (as it normally peaks about 3 ish and returns back to 2.3ish in a sort of 3 week sinewave.
I decided that I'd drop the dose for Friday (meaning I took 0mg) and then took 4mg on Saturday and will take 6 tonight.
Below is the graph of what my model produces (orange line) against the actual measured data (yes, I've been measuring daily).
To my eye te differences are that my model predicts a slightly higher INR (and if I'd measured at different times I may indeed have observed that) but its not "clinically significant" (as the jargon goes). It does predict a slightly lower dump than I've measured (2.5 measured, 1.8 predicted) but I put that down to the fact that my model does not have the capacity to deal with the inevitable "buffering" of all the stuff involved.
None the less I feel that I'm getting close to making a publishable software (and my friends are encouraging me to publish my current work in a Journal).
Happy New Year to everyone
