The Knowledge Management Lab
University of Toronto

Description

Mobile computing is making tremendous gains in popularity, thanks to advances in hardware, software and communications. Mobile computing promises to make data and services available to anyone, anywhere. To deliver on this promise, we need to develop technologies that will make it possible for non-technical users to manage their own data and to coordinate those data with data of others.

For example, we are developing a tool to allow physicians to prescribe drugs online. It is critical for a prescribing physician to have access to the details of a patient's medical history, but these data may be stored in a number of different locations. The tool will enable a doctor’s office to coordinate data with patient record databases belonging to other physicians, medical laboratories and pharmacies, and to get information from drug reference databases, medical reference databases and so on, so that it can generate advice on dose, adverse effects, interactions, cost, level of evidence and instructions for use. A doctor's office will be able to establish "acquaintances" with other databases containing patient data. Moreover, the set of acquaintances for any one office will be completely dynamic as the set of relevant databases changes over time. We are developing techniques for accessing information from a network of acquaintances and integrating it in response to a user formulating a single query with respect to her local database. For example, this will allow a family physician to find the results of any lab tests stored in the database of any acquainted specialist physician, pharmacy or medical laboratory.

A walk-in clinic may not wish to keep complete records on their irregular patients, in which case this sort of coordination at query time may be sufficient. On the other hand, for one of her regular patients, a family physician will want to keep as complete a medical history as possible in her own database. A key feature of this system is that it will be end-user programmable, meaning that the establishment of acquaintances and the coordination of data will be defined by office staff, rather than by database designers. For example, a staff person in a doctor’s office will be able to define co-ordination rules which specify conditions such as “An update of a patient’s prescription list should be propagated to the patient’s records owned by her cardiologist and her pharmacy”. Such a rule will be evaluated whenever there is an update to a patient’s prescription list, in order to ensure that the two patient databases evolve consistently.

The tool employs algorithms and software developed under the aegis of the Hyperion project, which conducts research into data management issues in the peer-to-peer computing paradigm.