This is a randomized crossover trial with 1:1 randomization to the admission sequence of using the Control AP system (rMPC - Naïve Model Predictive Control) vs. Experimental AP system (EnMPC - Ensemble Model Predictive Control) over approximately 4 months. Eligible participants will proceed to the Data Collection Phase for approximately 28 days, during which they will participate in regimented exercise activities. If the participant collected adequate data during the Data Collection Phase, they will be randomized and undergo the study admissions in the assigned sequence. Each admission is approximately 36 hours in length and will consist of one afternoon of exercise and one without.
Exercise remains a challenge to AP systems; more specifically, by the time exercise is detected it is often too late to avoid hypoglycemia without the ingestion of rapid carbohydrates or the use of rescue injections, such as glucagon. To this avail, the investigators propose to add a novel Model Predictive Control module to the proven USS system. This module is designed to compute insulin doses every 5 minutes that are designed to "optimally" maintain glycaemia around a target of 120mg/dL. The optimality is defined mathematically as minimizing deviations from basal rate injections and the distance between current and future (up to 2h) glycaemia from a physiologically feasible trajectory back down (or up) to a pre-specified target. Furthermore, the novel control system, labelled Multi Stage MPC or Ensemble MPC, accounts for a preset number of exercise scenarios during the prediction horizon, these scenarios being derived from the user historical record; this setup allows the control system to anticipate expected exercise bouts up to 2h in advance while maintaining the condition for optimal glycemic control. By adding such module to a well validated system, the investigators expect an improvement in protection against hypoglycemia during and immediately after physical activity without increase in hyperglycemia. To demonstrate the feasibility of this approach, the novel anticipatory system will be compared to a naïve AP system during highly supervised hotel admissions with afternoon exercise. Participants will be asked to exercise regularly in the late afternoon during a month of data collection to generate the patterns to be anticipated.
Study Type
INTERVENTIONAL
Allocation
RANDOMIZED
Purpose
TREATMENT
Masking
NONE
Enrollment
15
This AP controller has the ability to anticipate exercise activity by use of trends seen during the Data Collection Period.
This AP controller does not have the ability to anticipate exercise activity.
University of Virginia
Charlottesville, Virginia, United States
Number of Hypoglycemic Occurrences in Relation to Exercise Activity
Number of hypoglycemic occurrences immediately before, during, and immediately after exercise (\~5-7pm) as defined by more than one consecutive CGM values below 70 mg/dL or hypoglycemic treatment per glycemic guidelines.
Time frame: 2 Hours
Average CGM
Average CGM value
Time frame: 36 Hours
Percent Time CGM Below 54 mg/dL
Percentage of time CGM was below 54 mg/dL
Time frame: 36 Hours
Percent Time CGM Below 70 mg/dL
Percentage of time CGM was below 70 mg/dL
Time frame: 36 Hours
Percent Time CGM Between 70 and 180 mg/dL
Percentage of time CGM was between 70 and 180 mg/dL
Time frame: 36 Hours
Percent Time CGM Between 70 and 140 mg/dL
Percentage of time CGM was between 70 and 140 mg/dL
Time frame: 36 Hours
Percent Time CGM Above 180 mg/dL
Percentage of time CGM was above 180 mg/dL
Time frame: 36 Hours
Percent Time CGM Above 250 mg/dL
Percentage of time CGM was above 250 mg/dL
Time frame: 36 Hours
CGM Coefficient of Variation
Coefficient of Variation of the CGM Values
Time frame: 36 Hours
CGM-based Low Blood Glucose Index
CGM-based Low Blood Glucose Index (LBGI) which is a metric used to quantify the risk of hypoglycemia (low blood sugar) based on self-monitored blood glucose (SMBG) data. LBGI is based on a nonlinear transformation of blood glucose values that corrects for the asymmetry of the glucose scale. This transformation maps glucose values into a risk space (minimum risk = 0), where higher values correspond to higher risk. Values \<1 suggest low risk of hypoglycemia.
Time frame: 36 Hours
CGM Standard Deviation
Standard Deviation of the CGM Values
Time frame: 36 Hours
This platform is for informational purposes only and does not constitute medical advice. Always consult a qualified healthcare professional.