we'll extract these in a separate study definition because we only need to extract them once if we use "number_of_matches_in_period" for the entire study period

This isn't used anymore as we're defining the code variables using "number_of_matches_in_period"

This extracts variables once. In doing so, name this study definition more generally

See /issues/149

As part of this handle 0 total events in top 5

As part of this handle descriptions without match top 5

As part of this order columns in the two versions of top 5 consistently

This gives consistent width to the plots. The legend settings here work for the possible set of legends we have but may need to be adapted if these change in the future.