Access control refers to mechanisms for protecting access to confidential information, such as sensitive medical data. Management of access control policies, in applications that involve several collaborating parties, poses several challenges. One of these is in ensuring that each party in such a collaboration only obtains the minimal set of access permissions that they require for the collaboration. In a domain such as healthcare, it may be critical that access be minimized in this way, rather than allowing all parties equal access to the sensitive information. In practice, it is often difficult to manage access control policies to achieve this goal of minimality. This research will investigate new approaches to describing and reasoning about access control policies, to ensure that information that is shared in collaborative applications is only accessible to those parties in the application that require it. This work will develop both the theoretical foundations of, and prototype tools for, ensuring end-to-end security of data that is shared between organizations, with sharing of medical data as a particular source for use-case scenarios. On one dimension, the project will investigate the extension of workflow languages to describe protocols for sharing information in collaborate applications. On another dimension, relationship-based access control can describe access to information based on parties' roles in a collaboration. The project will investigate the use of spatial and temporal logics for relating these two, and user interfaces for presenting and reasoning about this information. Since ultimately software programs will perform the access and transfer of patient medical data, formal specifications will be extracted from workflow specifications, and used as a basis for checking software programs for their compliance with these access control policies. The broader impact of this research will be facilitated by the P.I.'s involvement with the NIH International Epidemiologic Databases on AIDS for Central Africa project (CA-IeDEA). The fruits of the research will be disseminated through interaction with open source communities involved in healthcare IT for low and middle income countries, as well as through interactions with lead healthcare IT decision-makers in the countries involved in CA-IeDEA.
TWC: Small: Workflows and Relationships for End-to-End Data Security in Collaborative Applications