2
$\begingroup$

Is there any available statistical method for assessing whether anonymisation performed on patient data in a medical dataset is sufficient?

$\endgroup$
2
  • 1
    $\begingroup$ Sufficient for what purpose? $\endgroup$
    – whuber
    Jun 12 '20 at 21:00
  • 1
    $\begingroup$ For the anonymization. Is there a statistical method that tells you when you have anonymized the data enough so that the patients cannot be identified. $\endgroup$
    – Cesare
    Jun 12 '20 at 21:01
2
$\begingroup$

There is no one-size-fits-all method. Proper anonymization, or de-identification, invariably requires both training and thought. You’ll find essentially this answer if you read regulations such as the US HIPAA (Health Insurance Portability and Accountability Act) that govern confidentiality of patient data.

$\endgroup$
1
  • $\begingroup$ I was thinking of using information theory. Like if the information conveyed from the data about the patient identity is less or equal to that after one scrambles the data, then we should be able to conclude that the patient cannot be identified. But it is just an idea right now. $\endgroup$
    – Cesare
    Nov 23 '20 at 6:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.