Page 355 - Glucose Monitoring Devices

362 Index

Measurement error (Continued) dataset description, 189, 190f
model validation, 93e95, 94f performance assessment, 192
parameters and second-order statistical performance metrics, 194e196, 194f, 195t
description, 93t posterior distributions, 196, 197f
scatter plots, 91f prior derivation, 190e191
YSI and SMBG-YSI preprocessing, 89e90 interstitial concentration, 178
random error, 79 Kalman filter-based approaches, 180e181
in silico clinical trials (ISCTs), 80e81 linear matrix inequalities techniques, 182
state-of-the-art modeling method linear time-invariant system, 178
constant-SD zones identification, 85 local dynamics models, 181
maximum-likelihood fitting, 85e87 low-pass filtering nature, 178e179
model validation, 87e88, 95t multistep calibration algorithm, 182
nonparametric approaches, 84 nonlinear and time-dependent relationship, 173
parametric approach, 84 parametrization, 177e178
training and test sets, 85 patients’ discomfort, 173
systematic error, 79 problem statement
type 1 diabetes (T1D), 80 calibrated glucose profile, 175
Medtronic Guardian Connect continuous glucose calibration function, 175
monitoring (CGM) systems, 128 electrical current profile, 176f
Medtronic’s SmartGuard self-adjusting insulin electrical current signal, 174
delivery system, 149 estimate quality, 176
Medtrum’s A6 Touchcare system, 149 glucose concentration profile, 174
Menarini Glucoday system, 213, 215f linear regression, 175
Metabolic models, 58 mathematical model, 174
Metabolic variability, 58 measurement noise, 175
Microneedle-array sensors, 116e117 numerical determination, 175
Minimally invasive continuous glucose moni- output, 177f
toring (CGM) sensor calibration sensor sensitivity, 175
autoregressive (AR) models, 181 two-point calibration, 175
Bayesian calibration algorithm recursive approaches, 183
BG-to-IG kinetics, 187e188 time variability, 179, 179f
current signal calibration, 189 MiniMed 670G system
dynamic system, 185f adjunctive agents, 339
fixed time-domain functions, 185 auto mode, 337e338
glucose concentration, 184 bihormonal systems, 339
low-pass filtering nature, 184 challenges, 339e340
model parameters, 185e189 forced system, 338
parameter initialization, 186 Guardian sensor 3, 336e337
parameters vector, 185 hotel-based study, 335e336
parameter update, 189 hybrid closed-loop systems, 339e340
physical domains, 184 insulin pump, 337
posteriori estimate, 186 3-month single-arm study, 336
SMBG sample, 188 Nightscout Project, 340e346
temporal domain of validity, 185 outpatient studies, 335e336
blood glucose-to-interstitial glucose (BG-to-IG) patient considerations, 346
kinetics, 176e177, 178f, 180 pediatric population, 336
challenges, 183 research unit-based study, 335
computational complexity, 182 retrospective analysis, 338
deconvolution-based Bayesian approach, 182 MiniMed predictive low glucose management
Dexcom G4 Platinum (DG4P) CGM sensor (PLGM) system, 278
calibrated CGM profiles, 192e194, 193f Model predictive control (MPC) algorithms,
calibration scenarios, 191 294e295, 297e299, 333