What is a risk when reporting de-identified data in very small populations?

• A. De-identification eliminates all risk even in small populations.
• B. There is a residual risk of re-identification due to unique combinations of attributes.
• C. Small populations cannot be analyzed.
• D. De-identified data are always perfectly safe to publish.

Answer: (B)

In very small populations, removing direct identifiers doesn’t remove risk completely. The main risk is a residual chance of re-identification due to unique combinations of attributes that, when taken together, can distinguish a single person in the group. For example, a rare diagnosis coupled with a specific age and location might pinpoint an individual even after names and addresses are removed. External information can be used to link the de-identified record back to someone. So, de-identified data are not guaranteed to be risk-free in small populations, and additional safeguards like further aggregation, suppression of rare values, or controlled access may be needed. It’s not true that all risk is eliminated, that small populations can’t be analyzed, or that de-identified data are always perfectly safe to publish.