Using a Personalized Decision Support Tool to Help People With Type 1 Diabetes Manage Exercise

People with type 1 diabetes frequently experience variable glucose responses before, during, and after physical activity, including exercise-associated hypoglycemia and the need for unplanned carbohydrate intake. Although consensus-based recommendations exist to guide exercise preparation, many individuals continue to rely on trial-and-error strategies that can be burdensome or difficult to generalize across activities.

This study evaluates the netIOB \& Exercise Toolkit (NEXT), a software-based, clinician-facing decision-support system intended to contextualize exercise preparation using recent continuous glucose monitoring (CGM) data and insulin delivery history. The Toolkit estimates net insulin-on-board relative to an individual's programmed basal insulin profile and presents structured, non-automated suggestions related to exercise timing, insulin target adjustments, and carbohydrate intake. All decisions remain under clinician and participant control.

The primary study population consists of adults with type 1 diabetes who complete three supervised, moderate-intensity exercise sessions in a randomized, three-period crossover design. Each session is conducted under a different pre-exercise strategy: consensus standard-of-care guidance, usual personal care routines, or guidance informed by the NEXT Toolkit. Sessions are separated by washout periods to minimize physiologic carryover, and glycemic, insulin delivery, and behavioral outcomes are assessed during exercise and for up to 48 hours post-exercise.

In addition, a healthy adult control group completes a single supervised exercise session without pre-exercise glucose or insulin management interventions. This group is included to provide comparative physiologic context for exploratory analyses.

Exploratory analyses will examine relationships between pre-exercise net insulin-on-board (netIOB) strata and glycemic, insulin, and treatment outcomes across exercise sessions. These analyses are intended to characterize patterns of response rather than to test formal efficacy hypotheses.

Participants in both the type 1 diabetes and healthy control groups will wear a commercially available continuous ketone monitoring (CKM) sensor (SiBio KS1) for exploratory data collection during exercise and defined post-exercise monitoring periods. CKM data are collected solely for research purposes, are not used to guide clinical or study-related decisions, and are not intended to evaluate device performance.

The NEXT Toolkit is classified as a non-significant risk, IDE-exempt investigational tool. Results from this pilot study are intended to inform future research design and the development of transparent, open-access methods for studying exercise physiology and glucose management in diabetes.

• Standard-of-Care Guidance (SoC) (Behavioral): Study staff provide insulin and carbohydrate adjustment advice based on consensus exercise management guidelines (Moer et al., 2024), tailored to the planned activity.
• Usual Care (UC) (Behavioral): Participants independently manage their insulin and carbohydrate decisions around exercise, using their routine practices without study-provided guidance.
• netIOB & Exercise Toolkit (NEXT) (Device): An investigational software tool generates individualized insulin and carbohydrate adjustment recommendations before and after exercise using CGM data and recent insulin delivery. Study physicians review and relay these recommendations to participants.
• Continuous Glucose Monitoring (Device): Continuous glucose monitoring using an FDA-cleared system (Dexcom G7) for research data collection during and after exercise. CGM data are used to assess glycemic responses and are not used to evaluate device performance or accuracy.
• Continuous Ketone Monitoring (Device): A commercially available continuous ketone monitoring (CKM) device will be used for exploratory data collection during and after exercise. Ketone data are collected for research purposes only and are not used for clinical decision-making or to evaluate device performance.

Primary outcomes

• Composite Score of Glycemic Safety and Carbohydrate Burden [Baseline (1 hour pre-exercise), Exercise (0 to 90 minutes), Early Post-Exercise (0 to 6 hours after exercise), Mid Post-Exercise (6 to 24 hours after exercise), Late Post-Exercise (24 to 48 hours after exercise); total of 50.5 hours per visit (3 visits).]: This composite outcome reflects glycemic risk during exercise and the subsequent post-exercise monitoring period. The score is based on a Magni risk-derived metric computed from continuous glucose monitoring (CGM) data, with the glycemic effects of rescue carbohydrate treatments removed to isolate underlying glycemic risk. Each evaluated timepoint is given a score ranging from 0 to 100, where lower scores indicate lower glycemic risk and higher scores indicate higher glycemic risk. Scores across the defined observation windows are summarized across intervention conditions.

Secondary outcomes

• Percent-Based Continuous Glucose Monitor Metrics [Baseline (1 hour pre-exercise), Exercise (0 to 90 minutes), Early Post-Exercise (0 to 6 hours after exercise), Mid Post-Exercise (6 to 24 hours after exercise), Late Post-Exercise (24 to 48 hours after exercise); total of 50.5 hours per visit (3 visits).]: Assess standard continuous glucose monitoring (CGM) metrics during defined time windows relative to exercise (0-90 minutes, 0-6 hours, 6-24 hours, and 24-48 hours post-exercise). These include: * Percent Time in Range (TIR): 70-180 mg/dL * Percent Time Below Range (TBR): <70 mg/dL and <54 mg/dL * Percent Time Above Range (TAR): >180 mg/dL and >250 mg/dL
• Glucose Concentration and Variability Metrics from Continuous Glucose Monitor [Baseline (1 hour pre-exercise), Exercise (0 to 90 minutes), Early Post-Exercise (0 to 6 hours after exercise), Mid Post-Exercise (6 to 24 hours after exercise), Late Post-Exercise (24 to 48 hours after exercise); total of 50.5 hours per visit (3 visits).]: Mean glucose level, glucose standard deviation (SD), and coefficient of variation (CV) will be calculated using CGM data collected over defined time periods relative to exercise. These metrics describe overall glycemic variability and will be used to assess the stability of glucose control following each intervention. Units of Measure: Mean glucose: mg/dL SD: mg/dL CV: %
• Frequency of Hypoglycemic Events [Exercise (0 to 90 minutes), Early Post-Exercise (0 to 6 hours after exercise); total of 7.5 hours per visit (3 visits).]: Number of hypoglycemic episodes per participant at thresholds of <70 mg/dL and <54 mg/dL, stratified by treatment arm. Events will be classified as treated or untreated.
• Glucose Nadir Value [Exercise (0 to 90 minutes), Early Post-Exercise (0 to 6 hours after exercise); total of 7.5 hours per visit (3 visits).]: Lowest glucose value (nadir mg/dl) recorded per participant during exercise period and 0-6 hours post-exercise, as measured by CGM.
• Time to First Hypoglycemic Event [Exercise (0 to 90 minutes), Early Post-Exercise (0 to 6 hours after exercise); total of 7.5 hours per visit (3 visits).]: Time (in minutes) from exercise start to the first hypoglycemic episode (<70 mg/dL) for each participant.
• Carbohydrate Consumption for Hypoglycemia Prevention/Treatment [Exercise (0 to 90 minutes); total of 1.5 hours per visit (3 visits).]: Quantify carbohydrate use related to exercise using: * Total grams of carbohydrate consumed * Number of carbohydrate treatment episodes * Categorization as prophylactic or reactive
• Total and Programmed Insulin Delivery [Baseline (1 hour pre-exercise), Exercise (0 to 90 minutes), Early Post-Exercise (0 to 6 hours after exercise), Mid Post-Exercise (6 to 24 hours after exercise), Late Post-Exercise (24 to 48 hours after exercise); total of 50.5 hours per visit (3 visits).]: Compare the total insulin delivered (actual) to the amount that would have been delivered based on the participant's pre-programmed basal rates. This includes both basal and bolus insulin.
• Frequency of User-Initiated Pump Adjustments [Baseline (1 hour pre-exercise), Exercise (0 to 90 minutes), Early Post-Exercise (0 to 6 hours after exercise), Mid Post-Exercise (6 to 24 hours after exercise), Late Post-Exercise (24 to 48 hours after exercise); total of 50.5 hours per visit (3 visits).]: Number of times participants manually modified insulin delivery during the exercise period, including suspensions, temporary targets, exercise mode, or basal rate changes.
• Duration of User-Initiated Pump Adjustments [Baseline (1 hour pre-exercise), Exercise (0 to 90 minutes), Early Post-Exercise (0 to 6 hours after exercise), Mid Post-Exercise (6 to 24 hours after exercise), Late Post-Exercise (24 to 48 hours after exercise); total of 50.5 hours per visit (3 visits).]: Cumulative duration of user-initiated pump modifications, including time spent in temporary targets, pump suspensions, and exercise modes.
• Baseline and Peak Ketone Levels [Baseline (immediately prior to exercise) and Exercise (0 to 90 minutes from exercise).]: Baseline and peak ketone concentrations (mmol/L) measured using continuous ketone monitoring during the exercise period. Outcomes will be explored across pre-exercise strategies in participants with type 1 diabetes and relative to healthy control participants.
• Ketone Exposure (AUC) During and After Exercise [Exercise (0 to 90 minutes), Early Post-Exercise (0 to 6 hours after exercise) per visit.]: Ketone exposure quantified as area under the curve (AUC) from continuous ketone monitoring during exercise (0-90 minutes from exercise start) and early post-exercise (0-6 hours after exercise). Outcomes will be explored across pre-exercise strategies in participants with type 1 diabetes and relative to healthy control participants.

Sex: All, ages: 18 Years to 60 Years, accepts healthy volunteers

Inclusion Criteria Inclusion Criteria: All Participants (Type 1 Diabetes and Healthy Control Groups)

* Adults between the age of 18-60 years
* Able to perform moderate intensity walking for 60 minutes (target 40-60% age-predicted maximal heart rate).
* Willing and able to comply with study procedures, including supervised exercise visits and device wear
* Able to provide written informed consent

Inclusion Criteria: Type 1 Diabetes Group Only

* Clinical diagnosis of type 1 diabetes for >1 year, based on the investigator's clinical judgement
* Current use of continuous subcutaneous insulin infusion with Tandem Control-IQ and a compatible continuous glucose monitor (CGM) for >1 month prior to enrollment
* Stable insulin delivery regimen, with no planned changes to insulin pump settings or insulin dosing strategy during the study period
* Consistent CGM use during the month prior to enrollment (>80% data availability)

Inclusion Criteria: Health Control Group Only

* No diagnosis of diabetes or other disorders of glucose metabolism
* Not using insulin or glucose-lowering medications

Exclusion Criteria Exclusion Criteria: All Participants

* Intercurrent illness or medical condition that precludes safe participation in moderate-intensity exercise (e.g., unstable cardiopulmonary disease, uncontrolled arrhythmia, or uncontrolled hypertension), as previously assessed by the participant's primary care physician
* Known coronary artery disease with symptoms limiting moderate physical activity, or history of myocardial infarction, percutaneous coronary intervention, or coronary artery bypass grafting within the past 12 months
* Pregnancy, lactation, or plans to become pregnant during the study period
* Renal insufficiency with estimated GFR <45 mL/min/1.73 m², dialysis dependence, or adrenal insufficiency
* Concurrent participation in another interventional drug or device study within 30 days prior to enrollment
* Inability to comply with study procedures or safety requirements (e.g., inability to achieve target heart-rate zone, attend scheduled visits, or enable required device data access), or otherwise deemed unsuitable by the investigator

Exclusion Criteria: Type 1 Diabetes Group Only

* Use of non-CSII insulin delivery, including long-acting injectable or inhaled insulin, during the study period
* Use of medications with potential to substantially affect glycemia (e.g., SGLT-2 inhibitors, GLP-1 receptor agonists, or GIP agonists), unless on a stable regimen with no planned changes during the study
* Use of systemic corticosteroids within 4 weeks prior to participation
* History of severe hypoglycemia (requiring third-party assistance) or diabetic ketoacidosis within the prior 6 months

• Rayhan Lal, MD, Principal investigator, Stanford University
• Dessi dessi@stanford.edu, PhD, Study director, Stanford University