SAFE-AP: New methods for an efficient and safe domiciliary Artificial Pancreas in type 1 diabetes


The artificial pancreas, or automatic glucose control, is the technology big companies in the diabetes area are betting on since it is expected it will revolutionize the management of type 1 diabetes, freeing the patient from the current burden of decision making and improving metabolic control. Nowadays several artificial pancreas prototypes have been validated in controlled inpatient studies. However, the domiciliary use of the artificial pancreas requires, besides an efficient controller able to face the daily life conditions, the safety mechanisms that allow its use without additional risk for the patient:
(a) Fault tolerance with regard to the instrumentation and communication between devices.
(b) Patient supervision allowing for risk assessment, beyond impending hypoglycemia, such as altered patient states or controller malfunctioning.
(c) Controller robustness and efficiency in face of disturbances like typical meals in the patient’s diet, exercise of different nature, diverse situations of stress and concomitant diseases.
(d) Optimization of the controller tuning and risk minimization in the clinical practice facing groups of patients with particular metabolic characteristics, transient alterations and metabolic changes due to the progression of the disease.
These challenges constitute the main objective of this project: the development of new methods and tools for efficiency and long-term safety of the artificial pancreas at home. This is reflected in the following specific objectives:
Objective 1. Development and validation of methods for the detection and diagnosis of faults in the instrumentation. Obstructions in insulin infusion, loss of signal by the continuous glucose monitor, failed sensors, calibration errors, etc. will be considered, among others.
Objective 2. Development and validation of methods for improving the accuracy of continuous glucose monitors and mitigating errors in the glucose readings or trends due to metabolic states related to hemodynamic changes associated with exercise and stress, present in the patient's daily life.
Objective 3. Development of methods for risk assessment. Prediction of hypoglycemia and severe hyperglycemia will be addressed, as well as detection of abnormal or altered states of the patient outside the operating range of the controller and controller malfunction or detuning due to metabolic changes.
Objective 4. Development and validation of control algorithms for fault tolerance and robustness to disturbances.
Objective 5. Design and implementation of an artificial pancreas for home use based on "smartphone", integrating previous methods for efficiency and safety during the life of the system.
Objective 6. Design and implementation of a remote supervision system to optimize performance and safety. The system will be a tool to generate knowledge using massive data mining techniques ("big data") to support decision making of medical personnel.

Start: 01/01/2014


End: 31/12/2016


Funder: MINECO


Reference: DPI2013-46892-C2-2-R


Grant: € 200, 860


Project coordinator: Jorge Bondía (UPV)


IIiA coordinator: Josep Vehí




Universitat Politècnica de València