The reason for this can be found in the small print (drop down in the corner of each plot). We decided the best way of harmonising the various specifications coming in from different directions was to make it a visualisation of the treatment records. A representation of the data we have been sent rather a clinical picture of patients. The more processing we do the more sources of error there are and the more things that need to be rigorously defined. So since patients can have multiple treatment records overlapping with a specific window the number of treatment records can be larger than the number of patients.
@George-D-S Are you happy with this explanation? @James-Medcalf and I are working through a new spec for the krt stats. Part of this will involve a bit of a reshuffle. The cohorts we will have will be more like: new, incident, prevalent and the stats will be calcualated with the latest treatment.
The reason for this can be found in the small print (drop down in the corner of each plot). We decided the best way of harmonising the various specifications coming in from different directions was to make it a visualisation of the treatment records. A representation of the data we have been sent rather a clinical picture of patients. The more processing we do the more sources of error there are and the more things that need to be rigorously defined. So since patients can have multiple treatment records overlapping with a specific window the number of treatment records can be larger than the number of patients.
@George-D-S Are you happy with this explanation? @James-Medcalf and I are working through a new spec for the krt stats. Part of this will involve a bit of a reshuffle. The cohorts we will have will be more like: new, incident, prevalent and the stats will be calcualated with the latest treatment.