The Foundation of Closed Loop Insulin Delivery

Closed loop systems—often called artificial pancreas systems—automate insulin delivery by integrating a continuous glucose monitor (CGM), an insulin pump, and a control algorithm. The algorithm processes real-time glucose readings and adjusts the pump’s infusion rate to maintain blood glucose within a predefined target range. While these systems manage many day-to-day adjustments on their own, they rely heavily on properly configured user settings. Commercial systems vary: hybrid closed loops require manual meal announcements while fully automated loops are still in development. To truly personalize your therapy, you need a deep understanding of how your specific device handles basal delivery, bolus calculations, and correction doses. This article provides a rigorous, data-driven framework for customizing those settings.

Core Parameters That Define Algorithm Behavior

Every closed loop algorithm leans on a set of adjustable parameters that dictate how aggressively or conservatively the system responds to glucose trends. Incorrect values can lead to frequent hypoglycemia or prolonged hyperglycemia, so careful, evidence-based tuning is essential.

Insulin Sensitivity Factor (ISF)

ISF represents how much your blood glucose drops when you take one unit of insulin. For example, an ISF of 1:50 mg/dL per unit means each unit lowers glucose by 50 mg/dL. ISF is not a fixed number; it shifts throughout the day due to circadian rhythms, activity levels, stress, illness, and hormonal cycles. Many pumps let you program multiple ISF values across different time blocks, which is critical for fine-tuning. A common starting point is the “1700 rule”: 1700 divided by your total daily insulin dose (TDD) gives you a rough ISF in mg/dL. From there, adjust by 5–10% based on three to five days of CGM pattern analysis. For instance, if you consistently see hypoglycemia at mid-morning after bolus correction, your ISF may be too high (too aggressive). Lower it by 5 mg/dL per unit and reassess.

Carbohydrate Ratios

The carbohydrate ratio (e.g., 1 unit per 10 grams) determines how much insulin you get with meals. Ratios often vary by time of day: many people need a tighter ratio at breakfast due to the dawn phenomenon, while lunch and dinner may require looser ratios, especially after exercise. Some systems allow separate ratios for meals, snacks, and corrections. To optimize, collect post-meal glucose data for five days. If two hours after a meal you spike above 180 mg/dL, consider decreasing the ratio (more insulin per gram) by 5–10%. If you hit a low within three hours, increase the ratio (less insulin). A practical strategy: start with a baseline ratio from your total daily insulin and meal carb intake (TDD / total daily carbs gives a rough ratio), then adjust time blocks based on logged excursions.

Basal Rates

Basal insulin covers the liver’s background glucose production and cellular glucose utilization. The closed loop algorithm modulates basal delivery in small increments every five minutes, but it performs best when the programmed basal rate closely matches your underlying physiological needs. Common patterns include the dawn phenomenon (glucose rising between 3–8 AM) and the “foot on the floor” effect (a sharp rise upon waking). To identify trends, examine the night-time CGM profiles for three consecutive nights. If you see a steady rise from 2–5 AM, increase basal by 0.05–0.1 units/hour during that period. If you see a drop, decrease basal. Wait three days before making further adjustments, and always change one block at a time. Some algorithms allow a separate “override” basal percentage for exercise or illness; use that feature rather than altering the underlying profile frequently.

Target Glucose Range

The target glucose defines what the algorithm aims for—typically 100–120 mg/dL (5.6–6.7 mmol/L). Lower targets (e.g., 100 mg/dL) tighten control but increase hypoglycemia risk, especially if you have hypoglycemia unawareness. Higher targets (e.g., 130 mg/dL) improve safety margins but may reduce time-in-range. Most modern systems offer time-specific targets: a higher target during exercise, sleep, or driving. Modifying the target is one of the most direct ways to change system aggressiveness. Start conservatively: if you’re currently at a target of 120 mg/dL and have more than 4% time below 70 mg/dL, raise the target to 130 mg/dL and see if lows improve. After two weeks of stable control, you can gradually lower by 5 mg/dL increments every one to two weeks, monitoring hypoglycemia rates.

Active Insulin Time (Duration of Insulin Action)

This parameter tells the algorithm how long insulin remains active in your body. Typical settings range from 3 to 5 hours for rapid-acting analogs (insulin lispro, aspart, glulisine). Setting active insulin time too short (e.g., 2 hours) risks insulin stacking and delayed hypoglycemia, because the algorithm may deliver another correction too soon. Setting it too long (e.g., 5.5 hours) can delay corrections and cause prolonged post-meal hyperglycemia. To find your optimal value, skip a meal and observe how long after a bolus your glucose continues to drop before stabilizing. If you still see effects 4 hours after a bolus, set active insulin time to 4 hours. Adjust by 30-minute increments and evaluate over a week. Some users with higher insulin resistance may need longer active times.

A Structured Framework for Customization

Personalizing closed loop settings is an iterative, data-driven process. Adjusting multiple parameters simultaneously can destabilize control and mask cause-and-effect relationships. Follow this systematic approach.

Step 1: Establish Baseline Data

Before making any changes, review the last 14–30 days of CGM data with a healthcare provider who specializes in closed loop systems. Generate standard reports: average glucose, time-in-range (70–180 mg/dL; 3.9–10 mmol/L), time below 70 mg/dL, time above 180 mg/dL, standard deviation, and coefficient of variation. The ambulatory glucose profile (AGP) is the gold standard for visualization. Identify primary concerns: is it fasting hyperglycemia? Post-meal spikes? Nocturnal hypoglycemia? Your provider can validate safe starting values for ISF, carbohydrate ratios, basal rates, and target range. Never adjust core settings without professional approval, especially in the first three months of pump therapy.

Step 2: Analyze Patterns with Time-Blocked Windows

Focus on three-hour windows around meals and overnight. Use the CGM software to overlay multiple days and spot consistent patterns. For example, if you see a steady rise from 2–5 AM and another from 11 AM–2 PM after lunch, you need separate solutions for each. Some systems provide “auto-mode” or “closed loop” reports that show how often the algorithm delivers automatic micro-boluses. If the system frequently issues auto-corrections, your ISF or carbohydrate ratios likely need revision. Document these patterns in a log or within the software annotations.

Step 3: Single-Variable Adjustments

Change only one parameter at a time. For example, if you detect a persistent post-breakfast spike, adjust the breakfast carbohydrate ratio by 10% (more insulin per gram) and wait three to five days. Keep the same ISF, basal rates, and target. Evaluate the change objectively: did the 2-hour post-meal glucose drop? Did it increase hypoglycemia? If the adjustment improves time-in-range without causing lows, keep it. If it leads to more lows, revert to the previous value and try a 5% adjustment instead. Use a simple spreadsheet or your device’s logbook to track changes and outcomes. Avoid changing settings more frequently than once every three days, as the algorithm needs time to settle.

Step 4: Leverage System Feedback Metrics

Modern systems provide valuable metrics: “time in closed loop,” “auto-corrections delivered,” “percentage of basal delivered automatically,” and “number of hypoglycemic events triggered by auto-suspend.” For example, on the Tandem t:slim X2 with Control-IQ, you can see how many auto-corrections were delivered per day. If the system is delivering more than 2–3 auto-corrections per meal, it likely indicates that your manual bolus or meal ratio is off. On the Medtronic 780G, the SmartGuard function adjusts basal on the fly, but if you see frequent high alerts, your active insulin time or ISF may need tuning. These metrics are powerful feedback; use them to guide adjustments rather than relying solely on fingerstick data.

Step 5: Adapt to Dynamic Life Factors

Exercise, illness, menstrual cycles, stress, alcohol, travel, and changes in weight all influence insulin sensitivity. Most systems have dedicated modes or temporary targets. For example, “exercise mode” typically raises the target glucose to 140–160 mg/dL or reduces basal by up to 50%. Create separate profiles for days with physical activity. Test these settings in a controlled environment: walk for 30 minutes at moderate pace with your exercise mode active, and monitor glucose every 15 minutes. Document the effect and refine the profile incrementally. Similarly, for illness, use a separate sick-day profile with increased basal and tighter targets if you tend to run high. The key is to preempt changes rather than react after extremes have already occurred.

Advanced Techniques for Demanding Scenarios

Once your baseline settings are stable, you can explore more nuanced strategies for specific situations that standard settings can’t handle well.

Time-Blocked Basal and ISF Profiles

Users with significant circadian variability benefit from multiple time blocks—most pumps allow up to 24. For example, you might have a slightly higher basal from 3–7 AM (0.1 units/hour more) to counteract the dawn phenomenon, then lower it from 1–4 PM (0.05 units/hour less) on days you typically take a walk after lunch. ISF can also be time-blocked: many people are more insulin sensitive in the afternoon due to activity. Adjust ISF by 10–15% during that window. Make changes in small increments (0.025–0.05 units for basal, 5 mg/dL per unit for ISF) and review after 1–2 weeks. Some algorithms let you set “ISF profile” separate from basal profile; take advantage of this granularity.

Extended and Combo Boluses for Complex Meals

High-fat, high-protein, or high-fiber meals cause delayed, sustained glucose rises that exceed the duration of a standard rapid-acting bolus. Extended boluses deliver a portion upfront and the remainder over 1–3 hours. Combo boluses (a mix of immediate and extended) are available on many pumps. For example, for a pizza meal, start with 60% immediately and 40% over 2 hours. The exact split depends on your personal response. Track the post-prandial curve: if you spike at 3 hours and then come down slowly, increase the extended portion or the duration. If you go low at 1.5 hours, reduce the immediate portion. This technique can dramatically reduce post-meal excursions without increasing hypoglycemia risk, but it requires careful testing. Work with your dietitian to identify which meals benefit from extended delivery.

Exercise Modes and Activity-Specific Profiles

Exercise lowers glucose during and often for many hours afterward, but the effect varies by intensity and duration. Most closed loop systems offer a dedicated exercise mode that raises the target (e.g., to 140–160 mg/dL) or reduces basal by a percentage (e.g., 50% for 2 hours). However, not all exercise is the same: light walking may require a modest increase in target, while high-intensity interval training (HIIT) can initially spike glucose due to stress hormones, requiring a different approach. Create at least two exercise profiles: one for aerobic (e.g., >30 minutes of steady activity) and one for anaerobic (e.g., sprints, weight lifting). For aerobic exercise, start a temporary target 30 minutes before and keep it active until glucose stabilizes after exercise. For anaerobic, you may need a higher target and possibly a small bolus to prevent the stress-induced spike. Monitor your CGM during activity and adjust the profile parameters over several sessions. Don’t forget the post-exercise period: delayed hypoglycemia can occur 6–12 hours later. Consider increasing your overnight target or setting a temporary basal reduction for the next 8 hours.

Safety Considerations and Common Pitfalls

  • Always involve your healthcare team. Never adjust core settings by more than 20% without professional guidance. Share data via cloud platforms like Dexcom Clarity, Tidepool, or Glooko so your provider can review changes remotely. If you cannot reach your provider, err on the side of conservative settings.
  • Make incremental changes. Adjust only one parameter at a time, and by no more than 15%. Wait at least three full days between adjustments. Frequent, large changes destabilize the algorithm and make it impossible to assess true efficacy. Document each change and its outcome.
  • Avoid overtly aggressive targets. Chasing a perfect time-in-range (e.g., above 80%) can paradoxically increase hypoglycemia as the algorithm applies frequent auto-corrections. A realistic goal is 70–80% time-in-range with <4% time below 70 mg/dL. If you are achieving 70% but with 5% lows, raising your target by 5 mg/dL often reduces lows without significantly harming time-in-range.
  • Verify your active insulin time. If you experience stacking lows after corrections, your active insulin time may be set too short. If you have prolonged post-meal highs despite correct boluses, it may be set too long. Validate by observing the glucose curve after a skipped meal.
  • Be aware of algorithm limitations. Some systems use adaptive learning that can shift settings based on your behavior. However, these algorithms may not rapidly adapt to abrupt changes like illness, travel across time zones, or new exercise routines. When in doubt, revert to a known-good profile and manually manage the situation.
  • Watch for “insulin stacking” loops. If the system delivers repeated small auto-corrections and your glucose continues to drop, cancel those corrections in manually if your device allows, or temporarily suspend delivery. Not all systems have a “cancel all pending corrections” option, so be prepared to overrule the pump if necessary.

Partnering with Your Healthcare Team

Successful personalization is a team effort. Beyond initial setup, your endocrinologist, certified diabetes educator, and dietitian can help interpret advanced metrics like coefficient of variation, standard deviation, and time-in-tight-range (70–180 mg/dL). Schedule quarterly reviews with extra visits during major tuning phases. Use reporting platforms like Tidepool, Glooko, or Dexcom Clarity to share standardized Ambulatory Glucose Profiles. These platforms can also highlight subtle patterns, such as rising glucose during the final hour of active insulin time. In addition, many pump manufacturers offer patient support programs with certified trainers who can answer specific questions about your device’s algorithms. For example, Tandem’s support team can help interpret Control-IQ reports, and Medtronic offers online coaching for SmartGuard. Don’t hesitate to use these resources.

The next generation of closed loop systems will incorporate additional biometric sensors—heart rate, skin temperature, accelerometers—to automatically adjust settings without user input. Some devices already learn from historical data and propose new profiles. Machine learning algorithms in development can personalize ISF and carbohydrate ratios in real time based on hundreds of data points, including meal composition, activity level, stress markers, and sleep quality. Automated meal detection using CGM rate-of-change patterns is being tested in clinical trials and may reduce the need for manual carbohydrate counting. Even as automation increases, understanding the foundational parameters remains valuable for troubleshooting and for interacting effectively with the system. Stay current with firmware updates from your pump manufacturer—many introduce new features like adjustable auto-correction aggressiveness, automated profile switching based on heart rate, or improved exercise detection. The FDA’s Artificial Pancreas Device System page provides regulatory updates and safety information.

Conclusion

Customizing your closed loop system transforms diabetes management from reactive glucose monitoring into proactive, personalized control. By methodically adjusting insulin sensitivity factors, carbohydrate ratios, basal rates, target ranges, and active insulin time—and by applying time-specific profiles and exercise modes—you can achieve better time-in-range, fewer hypoglycemic events, and a significantly improved quality of life. Always prioritize safety: collaborate with your healthcare team, make incremental changes, and rely on data to guide decisions. With patience and a structured approach, your closed loop system becomes a reliable partner in your daily diabetes care.

For additional resources, consult the American Diabetes Association, JDRF, Diabetes UK, and T1D Exchange. These organizations offer patient-centered educational materials and clinical updates that support your journey toward optimal closed loop use.