The concept of privacy is essential, as it simultaneously considers other core values such as respect, individuality, dignity, and personal autonomy. Safeguarding data in health research is also crucial since the process generally involves gathering, storing, and utilizing a vast amount of Personally Identifiable Information (PII) or Protected Health Information (PHI), which can be sensitive. This study aims to examine privacy measures in healthcare, providing a comparative analysis of privacy problems and technical solutions. To extract articles, we searched the PubMed database. Manual searches were also conducted in Google Scholar and the reference lists of relevant papers to ensure comprehensiveness. The results show, among others, that the pseudonymization of healthcare data is a widely valued approach as it addresses the challenges posed by other methods. Users and patients can also be actively involved in this process. In some cases, AI models are not only vulnerable to breaches but can also be exploited to amplify privacy attacks. Additionally, while standards like FHIR provide guidance on privacy and security, full compliance with regulations such as HIPAA or GDPR may not always be feasible. Therefore, implementers must carefully consider relevant legal requirements to mitigate privacy risks.
Invisible Threats: Rethinking Privacy in Digital Healthcare
Tabari P.;De Rosa M.;Costagliola G.;Fuccella V.
2026
Abstract
The concept of privacy is essential, as it simultaneously considers other core values such as respect, individuality, dignity, and personal autonomy. Safeguarding data in health research is also crucial since the process generally involves gathering, storing, and utilizing a vast amount of Personally Identifiable Information (PII) or Protected Health Information (PHI), which can be sensitive. This study aims to examine privacy measures in healthcare, providing a comparative analysis of privacy problems and technical solutions. To extract articles, we searched the PubMed database. Manual searches were also conducted in Google Scholar and the reference lists of relevant papers to ensure comprehensiveness. The results show, among others, that the pseudonymization of healthcare data is a widely valued approach as it addresses the challenges posed by other methods. Users and patients can also be actively involved in this process. In some cases, AI models are not only vulnerable to breaches but can also be exploited to amplify privacy attacks. Additionally, while standards like FHIR provide guidance on privacy and security, full compliance with regulations such as HIPAA or GDPR may not always be feasible. Therefore, implementers must carefully consider relevant legal requirements to mitigate privacy risks.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.


