Systematic Review of Available Models That Function as Solutions to Data Privacy in Deep Learning for Electronic Healthcare Systems
TL;DR
Overall, the findings underscore differential privacy’s growing role in healthcare data governance and emphasize SPDZ as a promising complementary framework for secure, privacy-preserving computation.
Abstract
Protecting patient data in healthcare systems requires robust privacy mechanisms that balance data security with utility. This systematic review examines the application of differential privacy (DP) and the SPDZ protocol in electronic healthcare systems. The review addressed three subquestions: the current research status of differential privacy in healthcare, approaches used to identify privacy challenges, and solutions proposed to mitigate risks across different subject domains. A comprehensive search yielded 4,970 studies, of which 316 were screened in full. After exclusions, 58 studies were analyzed. Results highlight increasing global interest in integrating SPDZ with differential privacy to strengthen data protection while enabling secure computation. Limitations include potential keyword bias, restricted temporal scope, and language constraints, which may affect coverage. Overall, the findings underscore differential privacy’s growing role in healthcare data governance and emphasize SPDZ as a promising complementary framework for secure, privacy-preserving computation.