Konferensartikel

Major infectious disease threat to the public health are either naturally occurring or artificially induced bioterrorist attack. Whether it is pandemics or epidemics, both are the burden of any public health authority, which needs well preparedness, proper monitoring and early detection of an outbreak before it has gone so far. Nowadays, syndromic surveillance is the most practiced approach, which uses data prior to laboratory or physicians verification including absenteeism, over-the-counter and prescription pharmacy sales, internet search volumes and others. Moreover, various bio-sensors networks for disease surveillance purpose have been put in place. However, disease surveillance systems, which detects outbreak during incubation period (before the onset of the first symptoms) is an ongoing research topic. According to recent findings, the presence of illness and elevated blood glucose values are found to be highly correlated. Therefore, the purpose of this project is to use either the continuous blood glucose measurements or mobile phone based diabetic patient’s historical data, such as blood glucose, insulin, diet, and meal, to develop an algorithm that is cable of detecting a cluster of people with an elevated blood glucose levels within a specific areas. We have developed an interval prediction mechanisms that can predict a step ahead blood glucose values of the individual patient. Thus, the goal of the prediction interval model is to obtain a 100(1 a) % forecast interval single-step-ahead into the future for glucose response at a given input level of past blood glucose, insulin, physical exercise and carbohydrate. Therefore, the measured blood glucose value can be compared against the interval predicted by our model. At this point, we have carried out simulation using continuous blood glucose measurements (CGM) from some group of diabetes patients, which shows the effectiveness of our approach. The single step point prediction is found to be accurate with a root mean square error (RMSE) of 0.2121 mmol/l. Moreover, we are capable of following the individual blood glucose evolution accurately with a significance level of a =0.01. The system is currently under experiments and is not yet tested fully at a prototype stage.

diabetes mellitus, continuous blood glucose measurement (cgm), diabetes diary, blood glucose prediction, outbreak detection, electronic disease surveillance.