Automatic adaptation of basal therapy for Type 1 diabetic patients: A Run-to-Run approach. (January 2017)

• Record Type: Journal Article
• Authors: Toffanin, Chiara; Messori, Mirko; Cobelli, Claudio; Magni, Lalo
• Is Part Of: Biomedical signal processing and control. Volume 31(2017)
• Page Start: 539
• Page End: 549
• Publication Date: 2017-01
• Subjects: Run-to-run -- Automatic adaptation -- Stability -- Piecewise affine system; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2016.09.002
• Languages: English
• ISSNs: 1746-8094

Abstract

Abstract : Highlights: New Run-to-Run algorithm based on clinical performance indices for basal adaptation. Algorithm convergence verified by the Lyapunov theory for piecewise affine systems. Robustness in silico evaluation on 100 adult patients of the UVA/Padova simulator. Time in range significantly increased in a week with a time below range reduction. Abstract: Type 1 diabetic patients require adjustments of their basal therapy due to insulin requirement changes. A very promising automatic approach is based on the so-called Run-to-Run (R2R) strategy, which adjusts the insulin therapy based on the performance measured during the previous run, usually of 1 day. Previous R2R approaches were based on a few blood glucose measurements. In this paper, the use of a subcutaneous continuous glucose monitoring is exploited in order to obtain more relevant clinical performance indices, such as the percentage of time spent below 70 mg/dl, above 180 mg/dl, and the average glucose. Different priority is given to hypo- and hyperglycemia control and, in particular, the primary goal is to reduce the time below 70 mg/dl, whereas the secondary is to reduce the time above 180 mg/dl and to reach a desirable glucose concentration. In doing so, a switching control law is achieved. A procedure for the convergence analysis of the proposed R2R approach is introduced by resorting to the Lyapunov theory for piecewise affine systems. Performance is studied for the algorithm by means of an extensive in silico trial performed on 100 adult patients of the UVA/Padova simulator. Five scenarios characterized by over- and underestimation of the initial basal therapy, random variations of the patient insulin sensitivity, meal variations, and intra-day circadian variability are considered. After a week the time in range [70–180] mg/dl is significantly ( p -value < 0.001) increased with a reduction of the time spent below 70 mg/dl by using the proposed algorithm.