Which statement best distinguishes de-identification from anonymization?

Which statement best distinguishes de-identification from anonymization?

• A. De-identification removes indirect identifiers but leaves direct identifiers intact.
• B. Anonymization removes or irreversibly obfuscates identifiers to prevent re-identification.
• C. De-identification means destroying the entire dataset.
• D. Anonymization is reversible.

Answer: (B)

The key idea is that de-identification and anonymization differ in how far they go to protect identity and whether re-identification is possible. De-identification reduces linkability by removing or masking identifiers, but it may still leave pathways (like indirect data points) that could allow someone to re-identify a person when combined with other information. Anonymization, on the other hand, takes the identifiers and removes or irreversibly obfuscates them so that re-identification is not feasible, even with additional data. That’s why the statement describing anonymization as removing or irreversibly obscuring identifiers to prevent re-identification best captures the distinction. It emphasizes irreversibility and the aim of making re-identification practically impossible. The other notions don’t fit: de-identification typically targets both direct and some indirect identifiers to reduce identifiability, not necessarily leaving direct identifiers intact. Destroying the entire dataset is not what de-identification means, and saying anonymization is reversible would undermine its purpose of preventing re-identification.

The key idea is that de-identification and anonymization differ in how far they go to protect identity and whether re-identification is possible.

De-identification reduces linkability by removing or masking identifiers, but it may still leave pathways (like indirect data points) that could allow someone to re-identify a person when combined with other information. Anonymization, on the other hand, takes the identifiers and removes or irreversibly obfuscates them so that re-identification is not feasible, even with additional data.

That’s why the statement describing anonymization as removing or irreversibly obscuring identifiers to prevent re-identification best captures the distinction. It emphasizes irreversibility and the aim of making re-identification practically impossible.

The other notions don’t fit: de-identification typically targets both direct and some indirect identifiers to reduce identifiability, not necessarily leaving direct identifiers intact. Destroying the entire dataset is not what de-identification means, and saying anonymization is reversible would undermine its purpose of preventing re-identification.