Integration of genomics data and electronic health records toward personalized medicine: A targeted review

Introduction: Diabetes disease is a group of metabolic diseases in which a person has high blood sugar, either because the pancreas does not produce enough insulin, or because cells do not respond to the insulin that is produced. Designing an automated system for regulating blood glucose in patients with diabetes is a solution that researchers have been paying close attention to in recent years. Therefore, safety is the minimum requirement for safety-critical systems such as the artificial pancreas. The present study introduces a safe, robust, performance-guaranteed optimal controller that can safely regulate blood glucose in the disturbance.Material and Methods: In this section, first, regulate blood glucose levels in simulation studies is evaluated. For this purpose, a dynamic model is used. The model includes a virtual patient, an insulin pump, and a continuous blood glucose level sensor. The virtual patient model represents the dynamics of insulin-glucose, carbohydrate-glucose, and exercise-glucose.Results: The need to not reset the controller parameters for patients in each category is one of the suggested controller's benefits. However, the PID controller needs to reset the parameters for each group of patients, the predictive control method requires the estimated model of the patient, and its performance is different on different days because the insulin-glucose dynamics for an individual changes day by day.Conclusion: Taking into account different sensitivities of body tissue to insulin, the results of evaluating the controller for two different groups of patients have shown that the controller is resistant to day-to-day changes in patients who may experience changes in insulin sensitivity, even with stress or medication and will not lose its optimal function. Based on the simulation results, the proposed controller can reduce the external disturbances' effect, whose amplitude is to a good extent within the body's physiological range.

Keywords:

Diabetes Mellitus , Type۱ Diabetes , Virtual Patient Model , Blood Glucose , Robust Optimal Control