Design and Evaluation of Privacy-Preserving Distributed Analytics for Collaborative Research

Abstract

Collaborative data analysis is essential in large-scale medical and social research. At the same time, privacy regulations and institutional policies restrict the direct exchange of sensitive records. This study describes the design and evaluation of a distributed analytics workflow that enables joint statistical analysis without transferring raw datasets. Secure aggregation and access auditing mechanisms were integrated into existing research infrastructures. Pilot deployments involving multiple research institutions demonstrate that core indicators can be computed with acceptable accuracy while maintaining compliance with data governance requirements. However, schema alignment and cohort definition remain major operational challenges. The findings indicate that privacy-preserving analytics requires sustained coordination between technical, legal, and organizational actors.