Warn users when PHI is potentially entered from non-published evidence. The user should be able to identify why the text was flagged and elect to edit the text or ignore the warning.
Alongside this: consider updating the interfaces with additional guidelines on how to best avoid PHI (for example: use age groups instead of specific ages, remove locations and labs when possible).
Warn users when PHI is potentially entered from non-published evidence. The user should be able to identify why the text was flagged and elect to edit the text or ignore the warning.
Alongside this: consider updating the interfaces with additional guidelines on how to best avoid PHI (for example: use age groups instead of specific ages, remove locations and labs when possible).
One option is to integrate with and leverage this AWS product: https://docs.aws.amazon.com/comprehend/latest/dg/how-medical-phi.html
First discussed on the VCI working group 2/25/21.