Recently I have been engaged in an interdisciplinary research in cooperation with researchers of medicine and applied mathematics at other universities, investigating the control of blood glucose. More specifically, this study, using data from patients in a postoperative intensive care unit (ICU), is aimed at developing a blood glucose control algorithm that is useful in medical practice. This study is different from the fields of study presented above, but I would like to introduce it here as a study related to IMI toward the application of mathematics and mathematical sciences to industrial society.

Immediately after surgery, the blood glucose levels of ICU patients rise rapidly because of stress caused by surgical invasion, cardiotonics, and other factors. Maintaining the blood glucose level in an adequate range by insulin administration is considered important because high blood glucose concentrations can cause multiple organ failure, coma, poorer outcomes, and other complications. Nevertheless, blood glucose control by insulin is difficult because of various factors such as delayed action of insulin and variation in insulin sensitivity. Another difficulty is that seriously poor outcomes might result from hypoglycemia triggered by insulin administration. Currently, blood glucose control in ICUs is conducted by nurses under predetermined conditions. Since no standard control algorithm has been established, blood glucose control is based largely on the experience of nurses under present circumstances. Such methods of blood glucose control are a burden on them. Some standardized method must be found.

Through cooperative study, we intend to develop a standardized method to assist blood glucose control in a way that is useful in actual clinical situations, and strive to estimate an algorithm used by experienced nurses to judge insulin administration based on a combination of mathematical approaches and observations of medical practice.

Regarding mathematical approaches to such problems, mathematical models of blood glucose and insulin related to blood glucose control have been designed to date. Existing models have different capacities and objectives, with difficulties such that mathematical models which express kinetics in the body include many unmeasurable variables. Parameters that are specific to each patient must be estimated. By contrast, various new data-driven study methods have been developed in recent years in the field of data science. Our cooperative study has adopted conventional methods and novel methods in an endeavor to contribute to the development of algorithms that are better suited for the study objectives.

In cooperative studies such as those of blood glucose control, for which medical observations are important, conducting studies cooperatively with surgeons and Certified Nurse Specialists in ICUs is crucially important, not only from the viewpoint of mathematics and mathematical sciences. We will proceed steadily with cooperative studies, carefully communicating with medical specialists, and thereby improving the application of mathematics and mathematical sciences to the field of medicine and to industrial society.