Mostly inspired by looking into medication activity. If, for example, there are medications given every two minutes and we generate activity spanning 3 minutes before and after, there will be significant overlap and "double counting." Instead, we would want to find where concurrent activity is at least 1 per time period (likely 1Min by default), and then end up spitting out an activity record with activity_start and activity_end spanning each of those.
Mostly inspired by looking into medication activity. If, for example, there are medications given every two minutes and we generate activity spanning 3 minutes before and after, there will be significant overlap and "double counting." Instead, we would want to find where concurrent activity is at least 1 per time period (likely 1Min by default), and then end up spitting out an activity record with activity_start and activity_end spanning each of those.