The Risks to Patient Privacy from Publishing Data from Clinical Anesthesia Studies

Area of Science:

• Medical Informatics
• Computer Science
• Anesthesiology

Background:

• Anonymization of patient data is standard practice before publishing research.
• De-identification methods are assumed to protect patient privacy in supplemental digital content.
• Recent research indicates that de-identified health data may still pose privacy risks.

Purpose of the Study:

• To examine the privacy risks associated with publishing de-identified data from small anesthesia studies.
• To evaluate the effectiveness of current de-identification methods against potential privacy attacks.
• To propose improved editorial policies for safeguarding patient privacy in published research.

Main Methods:

• Analysis of privacy implications for small-scale anesthesia studies (randomized trials, observational studies, case series).
• Review of computer science methods for privacy attacks on health information.
• Calculation of 'population uniqueness' using Texas state data for surgical patients.
• Assessment of re-identification risk for a randomly selected patient.

Main Results:

• De-identified data from small anesthesia studies can be vulnerable to privacy breaches.
• A 'population uniqueness' calculation for Texas surgical patients showed a 42.8% (SE < 0.1%) probability of re-identification.
• This risk level is considered unacceptably high and may underestimate risks in smaller populations.

Conclusions:

• Current de-identification practices may not adequately protect patient privacy in published anesthesia research.
• A significant risk of patient re-identification exists even with anonymized data.
• New editorial policies are needed to balance research transparency with robust patient privacy protection.