Current nutritional assessment, diagnosis and treatment regimes in the clinics are based on established rule-strategies developed from National nutrition procedures. However in the constantly growing patient database ofthe clinics there is a huge potential forthe re-use of data for research and development of a real-time CDSS based on cognitive computing that can be used to find new patterns that might be important for giving the right diagnosis and to predict treatment responses.
We aim at developing an intelligent and electronic real-time CDSS based on dynamic algorithms that are developing along with the constantly growing patient database. The underlying dynamic algorithmsofthe CDSS will enable the detection ofnew important patterns associated with certain treatments and thus be used to forsee a probable outcome for a patient. Because thealgorithms will be dynamic and shaped according to the growing data matrix, the CDSS will be ?trained? to give more and more precise predictions as the database is getting bigger. The "training" ofthe intelligent CDSS system requires both human training process, involving intensive cooperation between clinicians, statiticians, informaticians and software architects and in the next phases, machinelearning; which will be based on automated re-calculationofalgorithms.
The working process for real-time CDSS ranges from product development ofnew features using established technology to Information technology (IT)-innovation, using new technologies to solve new problems. A strong innovation perspective of this research work will be to enlarge the spectrum of possible applications to other fields than nutrition, including situations where patterns change rapidly such as for instance, in rapidly-evolving epidemics. In addition, there are large possibilities for "technology transfer" to foster similar initiatives in other European and non-European countries.