Generally in clinics diseases are diagnosed with their severity assessed by history taking, physical examination and investigations. This requires abundance of clinical knowledge, skills and experience. Difference in individual doctors' ability and in individual hospitals' condition makes it difficult to correctly diagnose and assess diseases and severity. For example many problems often arise in choosing indicators of symptoms, signs and investigatory findings, relationship between those indicators, their contribution rate to diagnosis and severity assessment, constant upgrade and correction of patients according to certain diseases, quantitative diagnosis and severity assessment ensuring objectivity and promptitude. Therefore in the age of knowledge driven economy, the usage of information technology is widely studied to rapidly and correctly diagnose and assess diseases and severity by using less indicators without secondary burden to patients.
Recently developed hypertension severity assessment system by artificial neural network enables not only doctors but patients to assess their hypertension severity by inputting their symptoms, signs and investigatory findings on computer, which leads to continual scientific management of hypertension.
Having chosen optic indicators useful for hypertension severity assessment through statistical test, Department of medical informatics , Pyongyang Medical college,
This computer aided assessment system consists of modules for inputting patient's data, diagnosis, process and displaying treatment.
Data input module has indicators useful for hypertension severity assessment as well as general indicators. Here patients' basic data is inputted. Inputted according to indicators, qualitative indicators are automatically quantitated and quantitative indicators undergo operation.
The indicators useful for hypertension severity assessment were scientifically taken by t-test and x2-test.
As a result of x2-test among qualitative indicators genetic disposition, psychentonia, stiffness of backneck), tinnitus, numbness in extremities, hypomnesia, a sense of oppression of heart, increased first sound, TV1>TV6, RV1/TV5>10mm, arteriole constriction, papillary edema, facial edema, decreased hypouresis, anorexia, hyposexuality were proven to be significantly useful. Among quantitative indicators suffered period, cardiac anxiety, cardiac pain, dyspnea, enlarged left area of cardiac dullness, IIa>IIp, RV5>26mm, SV1+RV5>36mm, enlarged aortic arc, enlarged left 4th arc, hemorrhage of fundus oculi, proteinuria were proven to be significantly useful.
As a result fo t-test among quantitative body weight, diastolic blood pressure, pulse rate, diameter of aorta, left ventricular posterior wall thickness, left ventricular diastolic diameter, left ventricular systolic diameter, especially age, systolic blood pressure and pulse pressure were proven to be significantly useful.
The diagnosis and process module undergoes BP artificial neural network process to assess the hypertension severity.
BP artificial neural network process was conducted in MATLAB Neural Network Toolbox 4.0.
The treatment display module makes it possible to constantly and scientifically manage the hypertension referring to the strategy according to the severity.
93.7% of hypertension patients(443) were correctly diagnosed by computer aided hypertension severity assessment system.
We are going to positively introduce this system into clinics to assess its efficacy and to actively conduct the academic interchange and collaboration of bionics and cardiology.
