SIENA

Stimulation Initiative for European Neural Applications

Esprit Project 9811

Case Studies of Successful Applications

Quality Assurance and Increased Efficiency in Medical Projects

Company background

No information available

The problem

Type II diabetes is characterized by an abnormal insulin secretion and a decreased insulin effect on insulin sensitive tissues. During the first stage of type II diabetes diet alone or in combination with antidiabetic drugs can compensate for diabetes related metabolic changes. Several factors can finally result in an additional need for long-term insulin treatment to achieve stable metabolic control. This stage of type II diabetes is referred to as "second failure" (SF) of the oral diabetic drug treatment.

Neural network application

There is an obvious need for a reliable and real time differential diagnosis of SF in order to select and start an appropriate treatment assuring the best quality of life. Data of 573 diabetics available for the differential diagnosis has been given to a neural net.

Benefits

With a neural network the accuracy of SF diagnosis is increased. It helps to start a patient's treatment considerably earlier than could be achieved by customary diagnostic procedures and therefore contributes to an improvement of the patient's quality of life.

Generalization

This research shows the relevance of neural network technology in capturing a great amount of data in medical applications.

Contact Person

Diabetes Forschungsinstitut - G. Entenmann. Auf'm Hennekamp 65, D-40225 Düsseldorf. Tel.: +49 211 338 21

Reference for further reading

Frster, M. et. Al.: DIADOQ: Untersuchungen zum Aufbau von Wissensbasen in der Diabetologie mittels Neuronaler Netze. 39. Jahrestagung der GMDS. Dresden, 1994.