DataTerminology.comNEW
The probability that anonymized data can be linked back to an individual’s identity.
DESCRIPTION
Re-identification risk quantifies the chance that de-identified records can be matched to individuals. It depends on dataset uniqueness, quasi-identifiers, and external data availability. Regulators require assessing and managing this risk before sharing or publishing data. Statistical disclosure control and expert determination are common approaches. Risk management balances privacy protection with analytical utility.
ALTERNATIVE DEFINITIONS
NISTIR 8053
Re-identification risk is the risk that information about an individual can be derived from de-identified data and associated with that individual.
Positions risk within de-identification lifecycle and data release.
U.S. Department of Health & Human Services
The risk of re-identification must be very small for data to be considered de-identified under the HIPAA Privacy Rule.
Links quantitative risk thresholds to HIPAA compliance criteria.
UK Information Commissioner’s Office
Organisations should assess the risk of re-identification, considering the data environment and adversaries’ likely resources and motivations.
Emphasizes contextual factors and adversary capabilities in risk assessment.
RELATED TERMS
K-anonymity l-Diversity t-Closeness Disclosure Risk Data Anonymization Data Protection Impact Assessment Data De-identification Quasi-Identifier
COMPETING TERMS
Differential Privacy Data Utility
KNOWLEDGE AREAS
Data Privacy & Compliance Data Governance Data Security
TARGET AUDIENCE
Data Privacy Officer Data Governance Manager IT Auditor Chief Information Security Officer Data Steward Data Scientist
EXTERNAL SOURCES
- Wikipedia - Data anonymization
- NISTIR 8053 - NIST guidance on de-identification
- HHS - HIPAA guidance on de-identification
- EDPB - EU anonymisation and risk assessment guidance
TOP QUESTIONS
Why is Re-identification Risk central to data sharing?
It determines whether data can be safely released without violating privacy. High risk implies stronger protections or reduced data granularity are necessary.
Category: Relevancy
How do k-anonymity and l-diversity affect Re-identification Risk?
They reduce risk by ensuring groups of records share quasi-identifiers and have diverse sensitive values, limiting identity inference.
Category: Techniques
What is the biggest misunderstanding about Re-identification Risk?
Assuming low risk in one context applies universally; risk is contextual and depends on auxiliary data, release environment, and adversary capabilities.
Category: Misunderstanding
How is Re-identification Risk quantified?
Through statistical models, uniqueness analysis of quasi-identifiers, attacker scenario modeling, and expert determinations under standards like HIPAA.
Category: Measurement
What trade-off exists between Re-identification Risk and Data Utility?
Stronger privacy protections often reduce data granularity and utility. Organizations must balance analytical needs with privacy constraints.
Category: Opinion
What are common mistakes when assessing Re-identification Risk?
Ignoring external datasets, underestimating adversaries, and failing to update risk assessments over time as data ecosystems evolve.
Category: Mistakes
Which controls reduce Re-identification Risk?
Differential privacy, generalization and suppression, controlled data access environments, and contractual restrictions limit linkage and inference.
Category: Controls
What is a hot topic around Re-identification Risk?
Evaluating risk in synthetic data and machine learning model outputs, including membership inference and attribute leakage.
Category: Topic
How should organizations manage Re-identification Risk over time?
Implement continuous privacy risk monitoring, re-assess before new releases, and adapt controls as auxiliary data grows.
Category: Management