PREDICTION SYSTEM OF DISEASE ACTIVITY INFORMATION, AND LEARNING DEVICE AND METHOD OF SYSTEM

摘要

PROBLEM TO BE SOLVED: To allow a disease activity index to be predicted by using gait information, and thereby making a blood test unnecessary and allowing the disease activity index to be grasped at home.SOLUTION: In a server device SV, an accurate disease activity index containing blood data for each patient is calculated and the calculated accurate disease activity index is used as training data in carrying out multiple regression analysis (a stepwise method) having an mHAQ value inputted by the patient, a pain joint count and a swollen joint count, a peak frequency PF contained in gait information, and a coefficient of variation CV and an autocorrelation coefficient AC as explanatory variables, and thereby a prediction formula of the disease activity index is produced. Then, a physical condition information parameter subsequently inputted to a user terminal UT by the patient at home and a gait information parameter measured in the user terminal UT are received from the user terminal UT, and the received parameters are substituted into the prediction formula as variables in order to predict the disease activity index of the patient.