Relationship between two sets of dates

Question

My wife has seizures a few times a month and takes medication for the seizures once a month. So I have two sets of dates - the dates of the seizures and the dates she got her medication.

I'm trying to determine whether there is any sort of pattern in the timing of the seizures vs. the medication. I hope to get some clues as to whether the medication is helping, whether she needs it more often, etc.

I'd appreciate any suggestions as to how to analyze these two sets of dates for any pattern/relationship. I'm not a data scientist, but I'm pretty good with Excel, so I'm hoping for something I can model in Excel.

Thanks!

Answer 1

A few simple ideas to observe whether there are any patterns:

• Plotting the days with seizure/without seizure together with the days when medication is taken across time. It's quite general but it might show some patterns.
• for each seizure, difference with the date of last seizure, i.e. number of days without a seizure. Plotting this value (simple plot of the number of days over time) might show whether there is any regularity and whether there is an evolution in the frequency.
• for each seizure, difference with the date of last medication. Plotting the distribution of this number (i.e. how often it's 1,2,3,...) might show if there is a direct relation between the two, e.g. if the number of days is generally more than 10 then it's likely that the effect of the medication wears down within 10 days. The medication might also have a delayed effect, which would be visible in the distribution.
• In case there is a long period of data available, you can calculate the average (for instance of days without seizure) over the last N days and plot the evolution of this average across time.

Probably you and your wife can think about other things which could be relevant. The general idea is to represent a potential parameter numerically and then observe how this value behaves.