Consider this example:

A technologist made an omission on a QC form. In response to that error, the form was changed in a manner designed to reduce the likelihood of a repeat of that error. After the form was changed, multiple errors involving numerous technologists were discovered related to that form.

If aggregate error-related data had been reviewed, it would have been seen that the original omission on the form was a single event. No other technologist had ever made the same error. In changing a form that had never before caused problems, a new error-prone situation was created.