Mobile Healthcare (mHealth) systems use mobile smartphones and portable sensor kits to provide improved and affordable healthcare solutions to underserved communities or to individuals with reduced mobility who need regular monitoring. The architectural constraints of such systems provide a variety of computing challenges: the distributed nature of the system; mobility of the persons and devices involved; asynchrony in communication; security, integrity and authenticity of the data collected; and a plethora of administrative domains and the legacy of installed electronic health/medical systems.
The volume of data collected can be very large; together with the data, there is a large amount of metadata as well. We argue that certain metadata are essential for interpreting the data and assessing their quality. There is great variety in the kinds of medical data and metadata, the methods by which they are collected and administrative constraints on where they may be stored, which suggest the need for flexible distributed data repositories. There also are concerns about the veracity of the data, as well as interesting questions about who owns the data and who may access them.
We argue that traditional notions of relational databases, and security techniques such as access control and encryption of communications are inadequate. Instead, end-to-end systematic (from sensor to cloud) information flow techniques need to be applied for integrity and secrecy. These need to be adapted to work with the volume and diversity of data collected, and in a federated collection of administrative domains where data from different domains are subject to different information flow policies.