Insulin Pump Settings: Basal Rates, Carb Ratios, and Correction Factors Explained

What Is an Insulin Pump and How Does It Differ from Injections?

An insulin pump delivers rapid-acting insulin continuously through a subcutaneous cannula, replacing both long-acting basal insulin and mealtime injections. The pump uses only one insulin type, typically lispro (Humalog), aspart (NovoLog), or glulisine (Apidra), and mimics the pancreas by varying the delivery rate across the day rather than releasing a fixed depot. Adults with type 1 diabetes using CSII achieve a mean HbA1c reduction of approximately 0.3 to 0.6% compared with multiple daily injection (MDI) therapy, based on a Cochrane systematic review of 23 randomized trials (N=976) that also documented a lower rate of severe hypoglycemia in pump users [1]. The practical advantage is programmability: every metabolic variable gets its own dedicated setting, so errors can be isolated and corrected one at a time.

Unlike MDI, where a poorly timed long-acting dose affects the entire 24-hour period, pump basal rates can be split into up to 48 half-hour segments. That granularity matters most for the dawn phenomenon, exercise-related insulin resistance, and shift workers whose schedules invert their circadian rhythms [2].

Understanding Your Total Daily Dose: The Foundation of Every Setting

Before touching any individual pump parameter, calculate the total daily dose (TDD). TDD is the sum of all insulin delivered in a 24-hour period, including basal and bolus. It anchors every other calculation.

For adults already on MDI, TDD is straightforward: add the units of long-acting insulin plus the average units of rapid-acting insulin consumed per day across at least seven days. When switching from MDI to a pump, most clinicians reduce the MDI TDD by 20 to 25% to account for the improved absorption consistency of subcutaneous infusion compared with injection [3]. The ADA 2024 Standards of Care state: "When initiating CSII, the TDD may be reduced by 20% from the MDI TDD and redistributed as 50% basal and 50% bolus" [4].

A patient on 30 units of glargine plus an average of 22 units of aspart per day has a MDI TDD of 52 units. Reducing by 20% yields a pump starting TDD of approximately 41.6 units, with roughly 21 units allocated to basal and 21 units allocated to meal boluses across the day.

The FDA-cleared Omnipod 5 and Medtronic MiniMed 780G both use TDD as the primary input for their adaptive algorithms [5][6]. Getting TDD wrong by even 15% skews the automated corrections enough to require manual override multiple times daily.

Setting Basal Rates: The 24-Hour Insulin Floor

Basal rate is the continuous background delivery measured in units per hour. It should keep blood glucose stable when the patient is fasting, not exercising, and not stacking prior boluses.

How to Test Your Basal Rate

The basal rate test requires a 4, 6-hour fasting window. Blood glucose (or CGM reading) at the start and end of the window should vary by no more than 30 mg/dL in either direction if the basal rate is correct [7]. Testing each segment separately is practical: overnight from 10 pm to 6 am, morning from 6 am to noon, afternoon, and evening. A rise of more than 30 mg/dL during the test window suggests the basal rate is too low for that period. A drop of more than 30 mg/dL suggests it is too high.

Most adults with type 1 diabetes require higher basal rates between 3 am and 8 am due to the dawn phenomenon, a cortisol and growth hormone-driven rise in hepatic glucose output. A clinical study published in Diabetes Care (N=115) showed that the pre-dawn basal rate needed was on average 34% higher than the midnight rate in adults using CSII [8]. Programming the pump to ramp up delivery starting at 2:30, 3:00 am typically blunts morning fasting hyperglycemia without adding nocturnal hypoglycemia risk.

Pediatric patients warrant special attention. Children under 6 years old are more sensitive to insulin and often need basal rates below 0.1 units per hour during overnight hours. The JDRF T1D Exchange clinic registry (N=25,529 across 67 centers) reported that pump-using children aged 2, 6 had a mean HbA1c of 8.0% vs. 8.4% in MDI users of the same age cohort, suggesting that the programmability of pumps confers meaningful benefit even when doses are very small [9].

Insulin-to-Carb Ratio: Matching Mealtime Bolus to Carbohydrate Load

The insulin-to-carb ratio (ICR) defines how many grams of carbohydrate one unit of rapid-acting insulin will cover. An ICR of 1:10 means one unit covers 10 grams of carbohydrate.

Carb Counting Basics

Accurate ICR programming is useless without accurate carb counting. The ADA recommends that all patients using CSII receive structured carbohydrate counting education, typically a minimum of three to five dedicated sessions with a registered dietitian familiar with diabetes management [4]. Carbohydrate content is counted in grams from food labels or validated databases; fiber subtracts from total carbohydrate only when the fiber content exceeds 5 grams per serving. Fat and protein slow gastric emptying and can raise blood glucose 3 to 5 hours after eating, an effect relevant to large, high-fat meals like pizza, where a dual-wave or extended bolus may be more effective [10].

Determining Your ICR

The 500 Rule provides the starting estimate for rapid-acting analogs: divide 500 by the TDD. A patient with a TDD of 40 units gets a starting ICR of 500/40 = 12.5, rounded to 1:12 or 1:13.

To verify that the ICR is accurate, the post-meal glucose two hours after a measured-carbohydrate meal should return within 30 to 50 mg/dL of the pre-meal glucose. A systematic overshoot (post-meal glucose more than 50 mg/dL above pre-meal) suggests the ICR is too conservative (ratio too high, not enough insulin per gram). A systematic undershoot (post-meal hypoglycemia) suggests the ratio is too aggressive [11].

ICR varies by time of day. Most people are more insulin-resistant at breakfast due to peak cortisol, so a breakfast ICR might be 1:8 while the lunch ICR is 1:12. The T1D Exchange registry found that approximately 62% of adult pump users program at least two distinct ICR settings across the day [9].

Insulin Sensitivity Factor (Correction Factor): Fixing Out-of-Range Glucose

The insulin sensitivity factor (ISF), also called the correction factor, predicts how many mg/dL one unit of rapid-acting insulin will drop blood glucose. An ISF of 50 means one unit is expected to lower glucose by 50 mg/dL.

The 1,700 Rule

For rapid-acting analogs, the standard estimation formula is: ISF = 1,700 divided by TDD. For regular human insulin the denominator constant is 1,500, but CSII pumps exclusively use rapid-acting analogs, so 1,700 applies in virtually all pump contexts [12]. A patient with TDD = 40 has a starting ISF of 1,700/40 = 42.5 mg/dL per unit, typically rounded to 40 or 45.

Validating the ISF

Validation requires a glucose above target (typically above 180 mg/dL) with no active meal bolus and no recent exercise. Administer a correction dose calculated from the current ISF, then check glucose at two hours. The result should land within 20 to 30 mg/dL of the programmed target. If glucose falls short of target by more than 30 mg/dL, the ISF may be too conservative (too high a number). If glucose overcorrects, the ISF is too aggressive (too low a number) [13].

ISF testing should be repeated after any illness, significant change in body weight, or change in physical activity level, because insulin sensitivity shifts substantially with each of these variables. A 2022 paper in Diabetes Technology and Therapeutics reported that ISF values changed by a mean of 18% following a 5 kg change in body weight in adults with type 1 diabetes (N=84) [14].

HealthRX Correction Factor Validation Framework (to use before every ISF adjustment):

1. Confirm pre-correction glucose is above target and steady on CGM for at least 20 minutes (no upward or downward arrow).
2. Confirm insulin on board (IOB) is zero or less than 0.2 units.
3. Confirm no meal consumed in the prior three hours.
4. Confirm no exercise planned within four hours.
5. Administer correction using current ISF, record pre-correction glucose, ISF used, and units delivered.
6. Check glucose at 90 minutes and again at 150 minutes.
7. Compute actual drop. Compare to predicted drop. Difference greater than 25% in either direction warrants ISF adjustment of 10 to 15%.
8. Repeat the test on two additional separate occasions before making a permanent setting change.

This eight-step sequence separates ISF error from confounders like exercise, meal-timing artifacts, and CGM lag. Your diabetes care team should review the log before any ISF change exceeding 15%.

Insulin on Board (IOB): Preventing Stacking Errors

Insulin on board (IOB) is the calculated estimate of active rapid-acting insulin still working in the body from prior boluses. It prevents the dangerous practice of stacking correction doses before the previous bolus has finished acting.

Most pump algorithms model IOB using a biexponential decay curve. Lispro and aspart have a duration of action of approximately 3 to 5 hours at typical subcutaneous doses, though individual pharmacokinetics vary [15]. Setting the IOB duration too short causes the pump to underestimate residual insulin, leading to hypoglycemia from stacked corrections. Setting it too long causes the pump to block necessary corrections, leaving glucose elevated.

A study in the Journal of Diabetes Science and Technology (N=60 adults, type 1 diabetes) found that using an IOB duration of 3.0 hours vs. 4.5 hours produced a statistically significant difference in nocturnal hypoglycemia rates: 8.3 events per patient-month vs. 3.1 events per patient-month, respectively (P<0.01) [16]. Clinicians should set IOB duration conservatively (4 to 5 hours) for patients who frequently experience nocturnal hypoglycemia and only shorten it when daytime correction response data clearly support doing so.

Target Glucose Range: Individualized by Risk Profile

The pump's target glucose sets the central value the correction bolus calculator aims for. The ADA 2024 Standards of Care recommend a pre-meal target of 80 to 130 mg/dL and a post-meal peak below 180 mg/dL for most non-pregnant adults with type 1 or type 2 diabetes [4]. However, these targets are adjusted substantially for:

• Pregnant individuals with pre-existing diabetes (fasting target <95 mg/dL, 1-hour post-meal <140 mg/dL per ACOG guidelines) [17]
• Older adults with hypoglycemia unawareness (target often relaxed to 120 to 150 mg/dL)
• Children under 6 years, where the ADA recommends a range of 100 to 180 mg/dL to reduce hypoglycemia risk [4]

Advanced hybrid closed-loop systems like the Medtronic MiniMed 780G allow setting a variable target; the default automated target is 100 mg/dL, though FDA labeling permits user adjustment between 100 and 120 mg/dL [6]. The Omnipod 5 system uses a fixed 110 mg/dL target in automated mode, which cannot currently be user-modified [5].

Basal-Bolus Regimens on Pumps: How the Settings Work Together

A basal-bolus regimen on a pump replicates physiologic insulin secretion more closely than any injection schedule. The basal program covers overnight fasting needs and inter-meal periods. The meal bolus uses the ICR and carb count. The correction bolus uses the ISF and target glucose. IOB subtracts from the suggested correction to avoid stacking.

The closed-loop or hybrid closed-loop systems (Omnipod 5, MiniMed 780G, Tandem Control-IQ) layer a real-time CGM signal on top of these four parameters and adjust basal delivery automatically every 5 minutes. The FLAIR trial (N=113 adolescents, crossover design) compared standard pump therapy plus CGM against the Medtronic 670G hybrid closed-loop system and showed a mean HbA1c improvement of 0.5% and 15% more time in range (70 to 180 mg/dL) with the closed-loop device (P<0.001) [18]. Even in automated systems, the manually entered settings form the baseline from which the algorithm makes adjustments; a grossly incorrect ICR or ISF forces the algorithm to work against a flawed foundation.

The ATTD 2023 consensus defined time in range (TIR, 70 to 180 mg/dL) as the primary CSII outcome metric, recommending a target of greater than 70% TIR for most adults with type 1 diabetes and greater than 50% TIR for older adults or those with significant hypoglycemia risk [19].

Adjusting Settings Over Time: When and How to Make Changes

Pump settings are not set-and-forget parameters. Body weight, activity level, illness, medications (especially corticosteroids), hormonal changes, and seasonal shifts in physical activity all alter insulin requirements. A general principle endorsed by the American Association of Clinical Endocrinology (AACE) is that any single pump setting should be changed by no more than 10 to 20% at a time, with a minimum of two to three days of observation before making a subsequent change [20].

Changes should happen one parameter at a time. Changing basal rate, ICR, and ISF simultaneously makes it impossible to identify which adjustment produced a given glucose response. Systematic pump optimization typically proceeds in this order: first verify basal rate accuracy, then validate ISF, then refine ICR, and last adjust the target glucose and IOB duration [7].

CGM data across 14 days provides the most reliable substrate for settings reviews. A 14-day ambulatory glucose profile (AGP) shows overnight trends (guiding basal adjustment), post-meal patterns (guiding ICR), and correction response curves (guiding ISF). Your diabetes care team should review AGP data at least every 3 months per ADA recommendations, and more frequently after any significant life change [4].

Illness protocols typically require a temporary basal rate increase of 20 to 50% to compensate for stress-hormone-driven insulin resistance. Most pumps allow a temporary basal rate that automatically returns to baseline after a defined duration, typically 30 minutes to 72 hours, without permanently altering the programmed settings [3].

Special Populations: Pediatric, Pregnancy, and Athletes

Pediatric Patients

Children grow rapidly, and TDD increases roughly in parallel with body weight and pubertal progression. Basal rates in prepubertal children often total 0.2, 0.4 units per hour across the full day, while adolescents in the peak of puberty may require two to three times the adult weight-based dose due to growth hormone-driven insulin resistance [9]. Settings should be reviewed at every clinic visit, typically every 8 to 12 weeks for children under 12.

Pregnancy

Insulin requirements increase steadily through the second and third trimesters, rising by 50 to 100% above pre-pregnancy TDD in many patients. The CONCEPTT trial (N=215 pregnant women with type 1 diabetes) found that CGM use during pregnancy significantly improved neonatal outcomes, with a reduction in large-for-gestational-age births from 41.8% to 31.8% (P=0.0210), underscoring the need for tight glucose control and frequent settings review throughout gestation [21]. ACOG and ADA both recommend weekly pump settings review during the second and third trimesters [4][17].

Athletes and Active Individuals

Exercise acutely increases insulin sensitivity and can lower glucose precipitously during and for up to 24 hours after aerobic activity. A temporary basal rate reduction of 50 to 80% beginning 60 to 90 minutes before moderate-intensity aerobic exercise reduces exercise-induced hypoglycemia without impairing performance. The consensus guidelines published in Lancet Diabetes and Endocrinology recommend that carbohydrate intake during aerobic exercise of more than 60 minutes should be 30, 60 grams per hour regardless of pump adjustments, because basal suspension alone may be insufficient to prevent late hypoglycemia [22].