We have presented a health-care decision support
system, which has been used experimentally on a large
database of clinical records, in order to show how data
warehousing and data mining can effectively support
evidence-based medicine. The proposed system enables
medical guide lines identification by exploiting evidence
based clinical history of a patient, standard protocols, and
other patients histories. A user study has been presented
to provide an early system validation focusing on usability
aspects.
A broader validation focusing on the quality of the
system provided information and the results of inferences
is currently being accomplished. In particular, other than
the medical data sources we had available, we also needed
the active involvement of medical doctors to assist us in
writing CD forms and scripts for specific medical domains,
and to help us evaluate the quality of inference results.
To this end, with the help of medical doctors from our
medical school we are currently integrating data sources
concerning a specific disease (Sepsis), for which we are
also writing CD forms and scripts. This will also allow us to
perform more sophisticated interpretations of textual
information contained within medical records, for which
we are also investigating the use of sketch recognition
strategies, due the handwritten text abounding in legacy
clinical records