The process of removing or modifying personal identifiers from data so that individuals cannot be directly identified from the dataset.
De-identification reduces privacy risk by removing or obscuring information that could identify a person. Techniques include removing direct identifiers (name, SIN, address), generalizing values (replacing exact age with an age range), suppressing rare values, and adding noise to numeric data.
De-identification is not the same as anonymization. Anonymization is a stronger standard — truly anonymized data cannot be re-identified even when combined with other datasets. De-identified data may still carry re-identification risk if combined with external information.
Example: A hospital de-identifies patient records before sharing them with researchers. Names, health card numbers, and exact dates of birth are removed. Postal codes are truncated to the first three characters. Ages over 90 are grouped as "90+". The resulting dataset supports research without exposing individual patients.
In Canada, de-identification practices for health data are guided by standards such as those published by the Canadian Institute for Health Information (CIHI). This is not legal advice.