Possible supervisor: Annette ten Teije
We have extensively modelled a diabetes treatment protocol in ASBRU. This protocol describes how patients should be treated. Via CBO, we also have available data that describes how a large set of patients have actually been treated. Using automatic analysis of these patient-data, and comparing them with machine-generated predictions of the prescribed treatment for these patients, it should be possible to analyse how well the doctors have complied with the protocol. Clearly, any large-scale deviations from the protocol are a cause for adjusting either the protocol, or the behavior of the doctors.