Patient Satisfaction and Clinical Experience With the Tandem t:slim X2 With Control-IQ Technology Advanced Hybrid Closed-Loop Insulin Delivery System

Abstract

Objective

The purpose of this study was to assess patient satisfaction and clinical experience with the t:slim X2 with Control-IQ Technology automated insulin delivery system.

Methods

This descriptive study used a retrospective electronic health record review of all individuals trained on the Control-IQ system between December 2019 and April 2022 in one adult endocrinology practice. A total of 99 patients using the Control-IQ system for at least 3 months completed the glucose monitoring satisfaction survey (GMSS). The primary outcome was overall satisfaction with the system as measured by the GMSS. Secondary outcomes included changes in A1C from baseline at 3, 6, and 12 months; the association between GMSS scores and A1C levels; and the use of the system’s sleep activity feature to achieve lower A1C levels.

Results

The overall satisfaction score for patients using the system was 4.0 ± 0.6 (possible score range 1–5). A1C decreased by 0.6% 3 months, 0.7% at 6 months, and 0.8% at 12 months. Participants with A1C levels ≤7% did not experience greater satisfaction compared to those with higher A1C levels.

Conclusion

Participants using the Control-IQ system reported a high rate of overall satisfaction and experienced significant reductions from baseline A1C at 3, 6, and 12 months.

In the realm of diabetes management, the evolution from multiple daily injection (MDI) insulin therapy to insulin pump therapy, and then, more recently, to automated insulin delivery (AID) reflects a span of advancements enabling insulin delivery with increasing levels of automation and precision to enhance glycemic control for people with diabetes. In the late 1970s, insulin pump therapy began to evolve. The MiniMed 502 insulin pump was approved by the U.S. Food and Drug Administration (FDA) in 1983, providing individuals with type 1 diabetes an alternative to an MDI regimen (1). Early insulin pumps provided continuous insulin delivery with flexibility in insulin dosing and helped eliminate the need for cumbersome calculations, although users still needed to manually input information such as their blood glucose readings and amounts of carbohydrates to be consumed. A meta-analysis of 10 studies in people with type 1 diabetes showed that insulin pump therapy provided better glycemic control than MDI regimens (2).

The FDA approved the Tandem t:slim X2 with Control-IQ Technology (Control-IQ) AID system (Tandem Diabetes Care, San Diego, CA) in December 2019. This device is an advanced hybrid closed-loop (HCL) system that uses continuous glucose monitoring (CGM) technology to maximize glycemic control. The Control-IQ technology can be used in people ≥6 years of age. It can automatically adjust insulin dosing based on a control algorithm that predicts individual glucose trends to improve glycemic time in range (TIR; 70–180 mg/dL) and reduce hypoglycemia (3). The goal of AID is to lessen the burden of diabetes.

The t:slim X2 pump pairs with a Dexcom G6 or G7 CGM system (Dexcom, San Diego, CA) and uses the Control-IQ algorithm to predict glucose levels 30 minutes in advance and automatically adjust insulin delivery accordingly. The system works to minimize glucose fluctuations and keep glucose levels within a defined range. The Control-IQ system features a sleep activity setting that is used to predict insulin needs and adjusts insulin delivery using a lower target glucose range (112.5–120 mg/dL) during sleep. Additionally, an exercise activity setting adjusts insulin delivery based on a higher target glucose range (140–160 mg/dL) for physical activity.

In a randomized controlled trial, Brown et al. (4) showed an increased TIR when using the Control-IQ system compared to sensor-augmented pump therapy over a 6-month period in people with type 1 diabetes. Another study looking at real-world use of the Control-IQ system demonstrated a 10% improvement in mean TIR after 1 year in participants with type 1 or type 2 diabetes (5). This research investigated user satisfaction and real-life clinical experience with the Control-IQ system in adults with diabetes.

Research Design and Methods

Evaluation and Survey

This descriptive study used a retrospective electronic health record review of all patients who trained on the Control-IQ system between December 2019 and April 2022 in an adult endocrinology practice located in the southeastern United States. This practice includes two endocrinologists and two nurse practitioners, as well as a nurse and a pharmacist who were both certified diabetes care and education specialists (CDCESs).

Individual pump training sessions are provided by the CDCESs. Patients are followed by the CDCESs for 2 weeks after starting the system and have a scheduled follow-up appointment in 4–6 weeks after the initial training. During the initial 2 weeks, settings are adjusted and concepts covered in the initial training are reinforced.

For this study, surveys were offered to patients as they came for regular follow-up appointments if they had used the Control-IQ system for at least 3 months. Surveys were completed at the beginning of the appointments, while clinic staff downloaded devices and printed reports. Surveys were optional, and rates of declining the survey were not tracked. Patients provided consent when completing the paper survey.

Measures

The primary outcome was overall satisfaction with the system as measured by the type 1 diabetes version of the Glucose Monitoring System Satisfaction (GMSS) survey (6). The GMSS is a 15-item scale that measures key factors associated with device satisfaction, with different versions for type 1 and type 2 diabetes. In addition to measuring total satisfaction, the GMSS contains four subscales. These measure “Openness” (e.g., “Helps me be more open to new experiences in life”), “Emotional Burden” (e.g., “Makes me feel more frustrated with my diabetes”), “Behavioral Burden” (e.g., “Is too much of a hassle to use”), and “Trust” (e.g., “Gives me numbers I don’t entirely trust”). The survey scoring requires reverse coding for negatively worded items. Scores of 1 or 2 indicate unfavorable satisfaction scores (“strongly disagree” or “disagree”), a score of 3 indicates a neutral satisfaction level, and scores of 4 or 5 indicate favorable satisfaction scores (“agree” or “strongly agree”). Higher subscale and total scores indicate greater satisfaction (6).

Secondary outcomes included changes in A1C from baseline to 3, 6, and 12 months; the association between GMSS scores and A1C levels; and the use of the system’s sleep activity feature to achieve lower A1C levels. Baseline A1C levels were assessed before starting the Control-IQ system. Data collection also included percentages of TIR, time above range (TAR), and time below range (TBR); type of insurance; and diabetes technology devices used before starting the Control-IQ system. Finally, at the end of the survey there were two additional questions related to use of the system touch screen and the use of infusion sets. These two questions were crafted because of recurrent patient questions and concerns. The idea was to collect information to inform improved training practices in the future.

Statistical Analysis

Baseline parameters of age, sex, type of diabetes, diabetes duration, and baseline A1C were assessed. Descriptive statistics included mean, SD, range, and percentage. Changes in A1C were examined using repeated ANOVA with pairwise comparisons and Bonferroni corrections. Student t tests were used to assess 1) A1C threshold and GMSS scores and 2) the use of sleep activity and TIR. χ² analysis was used to assess whether A1C levels correlated with the use of the sleep activity feature. Finally, Student t tests with Bonferroni corrections were used to compare current study (Control-IQ) satisfaction scores to previously collected satisfaction scores for the MiniMed 670G HCL system (Medtronic, San Juan, Puerto Rico) (7). This comparison was made because the MiniMed 670G was the most recent technology evaluated using the GMSS tool. Statistical significance was determined at P ≤0.05. All analyses were performed using SYSTAT 13 statistical software (Systat Software, San Jose, CA).

Results

All patients at the endocrinology clinic using the Control-IQ system for at least 3 months were eligible for the study, and 99 patients completed the GMSS. When the last survey was completed, ∼134 patients had been started on the Control-IQ system, so 74% of those eligible completed the survey. Table 1 depicts the characteristics of the study population. The mean age was 53.3 ± 15.3 years and the majority of participants were female (61.6%). The average duration of diabetes was 27.4 years (range 2–72 years). The mean baseline A1C was 7.89 ± 1.0%, and most patients had type 1 diabetes (n = 82). Most participants (58%) had commercial insurance, whereas 33% were covered by Medicare. Before completing the survey, the mean duration of use for the Control-IQ system was 12 months (range 3–26 months). Participants spent 94% of the time with the system in automated mode. Interestingly, two patients had used the system for only 3 months at the time of the survey, and both had baseline A1C levels <7%.

Table 1.

Study Population Baseline Characteristics

Characteristic	Study Population (n = 99)
Age, years	53.3 ± 15.3
Female sex	61 (61.6)
Caucasian race	93 (94)
Diabetes duration, years	27.4 ± 14.7
Baseline A1C, %	7.89 ± 1.0
Diabetes type
Type 1	82 (82.8)
Type 2	17 (17.2)
Duration of Control-IQ use, months	11.7 ± 4.8
Time spent with system in closed-loop automated mode, %	94.1 ± 5.9
Insurance
Commercial	57 (58)
Medicare	33 (33)
Other (Medicare + Medicaid or Medicaid only)	9 (9)

Data are mean ± SD or n (%).

The mean overall satisfaction score for patients using the Control-IQ system was 4.0 ± 0.6 (possible score range 1–5). Scores for the four subscales, along with explanations for each, are shown in Table 2. For A1C, there was a 0.6 percentage point decrease (n = 99) at 3 months (95% CI 0.471–0.786), a 0.7 percentage point decrease (n = 88) at 6 months (95% CI 0.552–0.932), and a 0.8 percentage point decrease (n = 63) at 12 months (95% CI 0.578–0.991), all of which were statistically significant (P <0.001). Participants with A1C levels ≤7% at the time of the survey did not experience greater satisfaction (i.e., higher GMSS scores) when compared to patients with A1C levels ≥7.1% (95% CI −0.225 to 0.239, P = 0.950). In addition, participants using the scheduled sleep activity feature (n = 44) did not have a higher incidence of A1C <7% (P = 0.685) or greater TIR (P = 0.240) compared to those not using the sleep activity feature.

Table 2.

GMSS Survey Scores

Scale/Subscale	Score*
Total score (higher scores indicate greater satisfaction)	4.00 ± 0.6
Openness subscale (higher scores indicate greater openness)	3.94 ± 0.6
Emotional Burden subscale (higher scores indicate greater burden)	2.04 ± 0.7
Behavioral Burden subscale (higher scores indicate greater burden)	1.86 ± 0.6
Trust subscale (higher scores indicate greater trust)	3.96 ± 0.8

Scores are mean ± SD. *Based on a 1–5 Likert scale.

For adults with type 1 or type 2 diabetes, the American Diabetes Association’s (ADA’s) Standards of Care in Diabetes—2024 recommends a TIR (70–180 mg/dL) >70%, with <4% TBR (<70 mg/dL) and <25% TAR (>180 mg/dL). There are less strict guidelines for older adults, including TIR >50% with <1% TBR (<70 mg/dL) and <50% TAR (>180 mg/dL) (8). In this study, the mean TIR was 67.2% for adults (23–64 years of age) and 70.1% for older adults (65–83 years of age). Forty-six of the 99 participants achieved a TIR >70%. Additionally, the mean TAR was 32.2% for adults and 28.6% for older adults. Older adults experienced a mean 1.3% TBR, which is slightly higher than the recommendation, and younger adults had a mean TBR less than the recommendation, at 1%. Figure 1 shows CGM metrics for participating adults (23–64 years of age) and older adults (65–83 years of age) compared to ADA glycemic target recommendations for each age-group (8).

Figure 1.

CGM metrics for participating adults (23–64 years of age) and older adults (65–83 years of age) compared to ADA glycemic target recommendations for each age-group (8).

Prior use of technology included MDI with CGM (n = 22), MDI without CGM (n = 5), insulin pump only (n = 18), pump with CGM but not AID (n = 28), t:slim X2 pump with low-glucose suspend capability (Basal-IQ) (n = 14), and other AID system (n = 12). Table 3 displays baseline A1C, total satisfaction scores, and secondary outcomes based on the prior use of technology. Interestingly, the highest satisfaction scores were in patients previously using a pump with CGM but not AID, another AID system (i.e., the MiniMed 670G), or MDI with CGM. All groups experienced a reduction in A1C, but the largest A1C reduction was in the group previously using MDI with CGM. Patients previously using an AID system still experienced a 0.48 percentage point reduction in A1C after starting the Control-IQ system.

Table 3.

Baseline A1C and Control-IQ Outcomes (Changes in A1C, TIR, TBR, TAR, and Total Satisfaction Scores) Based on Prior Use of Technology

Previously Used Technology	Baseline A1C, %	Change in A1C, %*	TIR, %	TBR, %	TAR, %	Total Satisfaction Score
MDI with CGM (n = 22)	8.25 ± 1.2	−1.21	67.5 ± 17.4	1.1 ± 1.8	31.4 ± 17.8	4.0 ± 0.5
MDI without CGM (n = 5)	7.58 ± 0.94	−0.86	67.0 ± 12.0	0.4 ± 0.9	32.6 ± 11.7	3.9 ± 0.5
Pump only (n = 18)	7.94 ± 1.16	−0.72	67.3 ± 13.1	1.1 ± 1.2	31.7 ± 13.3	3.7 ± 0.7
Pump with CGM, but not AID (n = 28)	7.86 ± 0.95	−0.76	68.6 ± 13.0	1.0 ± 1.1	31.6 ± 14.6	4.2 ± 0.5
Basal-IQ pump (n = 14)	7.86 ± 1.03	−0.66	66.0 ± 11.9	0.9 ± 1.2	32.8 ± 14.3	3.9 ± 0.5
Other AID system (n = 12)	7.38 ± 0.54	−0.48	71.8 ± 7.1	2.0 ± 1.7	26.2 ± 7.6	4.1 ± 0.5

Data are mean ± SD or mean. *A1C was evaluated retrospectively at 3, 6, and 12 months after completion of the survey.

When scores from the GMSS were compared to scores from a previous study in patients using the MiniMed 670G, participants using the Control-IQ system reported less behavioral burden (e.g., “Is too much of a hassle to use”) and trusted the numbers (i.e., the glucose sensor readings obtained with the system) more than patients using the MiniMed 670G (P = 0.039 and P = 0.057, respectively) (7). The mean total satisfaction score for the Control-IQ system (4.0) was higher than for the MiniMed 670G system (3.8), but the difference was not statistically significant (P = 0.073).

At the end of the survey, there were two additional questions. The first question was regarding problems using the system touch screen (“Do you have problems using the touch screen? Yes or No?”). Despite the CDCESs receiving many complaints from patients struggling with the touch screen during the initial training, only 5% of participants reported a problem using the touch screen. The second question focused on the pump infusion sets, and participants could choose more than one answer (“Do you have problems with the infusion set? Yes or No?; Choose any and all that apply: a) inserting in body, b) disconnecting, or c) bent or kinked cannulas [small piece under the skin that delivers insulin]”). Eleven participants reported problems inserting the sets, 18 reported problems disconnecting the set from their body, and 31 reported bent or kinked cannulas after insertion.

Discussion

The validated GMSS is a reliable measurement of quality of life and treatment satisfaction in patients using glucose monitoring systems (6). The current study revealed favorable subscale mean scores for openness and trust, as well as for emotional and behavioral burden for the Control-IQ system. Overall device satisfaction scores revealed that most patients were satisfied with the system. It is important to note that all mean satisfaction scores were favorable. Table 3 shows that patients previously using an insulin pump with CGM that was not an AID system and those previously using another AID system reported the highest satisfaction scores after transitioning to the Control-IQ system. Polonsky et al. (6) previously reported that greater device satisfaction, as reflected by the GMSS, is linked to greater well-being, lower diabetes distress, and fewer negative attitudes toward blood glucose monitoring.

Several recent studies have evaluated satisfaction with newer diabetes technology, although none have used the GMSS tool to review satisfaction. Matejko et al. (9) demonstrated that people with type 1 diabetes who were naive to insulin pump therapy and CGM, who later transitioned to the MiniMed 780G HCL system, experienced improved glycemic control and significant improvements in quality of life. The control group (who used MDI and self-monitoring blood glucose) remained unchanged. Additionally, Polonsky et al. (10) evaluated psychosocial outcomes for people with type 1 diabetes who used the Omnipod 5 AID system (Insulet, Acton, MA). Study participants completed various psychosocial outcomes questionnaires before and 3 months after initiating the Omnipod 5 system. Patients using the Omnipod 5 system demonstrated significant improvements in psychosocial measures such as diabetes distress, hypoglycemia confidence, insulin delivery satisfaction, diabetes treatment satisfaction, and system usability. Finally, Pinsker et al. (11) reviewed real-world patient-reported outcomes and glycemic results after patients initiated the Control-IQ system. Participants completed multiple psychosocial questionnaires, including tools that evaluated satisfaction. Improvements in psychosocial outcomes and persistent achievement of glycemic control were noted. Patients reported high device-related satisfaction, improved quality of life, and reduced diabetes impact. These findings are generally consistent with the current study and help to confirm that people living with diabetes who use advanced diabetes technology are more satisfied and experience significantly improved quality of life.

With regard to glycemic control, Brown et al. (4) observed a 0.33 percentage point reduction in A1C after 6 months of using closed-loop technology (mean baseline A1C 7.4%). Participants in the current study had a higher mean baseline A1C of 7.9% and experienced a more substantial mean decrease in A1C at 6 months (−0.7 percentage point, P <0.001). The change in A1C for the current study occurred quickly, with the majority of the A1C reduction taking place within 3 months of initiating use of the system. Regardless of previously used technology, all groups in this study experienced A1C lowering. Participants previously using MDI with CGM experienced the most substantial A1C reduction after starting the HCL system. Of note, Polonsky et al. (6) previously observed a correlation between lower A1C levels and higher GMSS total satisfaction scores. However, the current study did not consistently demonstrate this correlation despite significant improvements in A1C and consistent reports of device satisfaction.

The current study also demonstrated an average TIR of 67.2%, which was slightly below the target of >70% for younger adults not at risk for hypoglycemia, indicating room for improvement. The lower TIR could be attributed to less than half of participants (44%) using the sleep activity feature. In a 12-month real-world study, Breton and Kovatchev (5) concluded that the use of the sleep activity feature increased TIR at night, especially in people with type 1 diabetes. However, the current study did not demonstrate a correlation between use of the sleep activity feature and improved TIR. Additionally, postprandial excursions resulting from late or missed boluses and over-correcting for hypoglycemia also may contribute to lower TIR (12). Of note, there was minimal risk of hypoglycemia in the younger adults, with a mean TBR of 1%, which was far below the ADA goal of <4%. Moreover, the ADA recommends a TIR >50% for adults >65 years of age, and the current study surpassed that goal with a mean TIR for older adults of 70.1% with minimal additional hypoglycemia (TBR 1.3%, which is slightly greater than the recommendation of <1%) (8). This study further demonstrated that advancements in diabetes technology help patients attain better glycemic control.

Although patient satisfaction scores and clinical outcomes are favorable with the Control-IQ system, it is important for medical providers and patients to have realistic expectations. The Control-IQ system is not 100% automated; patients must count and input carbohydrate amounts for meals and snacks. The additional questions on the survey used in this study showed that most patients did not have a problem using the touch screen but did report a variety of challenges with infusion sets. The most common problem was kinked or bent cannulas. Kanapka et al. (13) reported the straight insertion sets used with the Tandem pump have a higher failure rate than the angled insertion or steel infusion sets. This finding highlights the need to ensure that patients are able to troubleshoot and resolve infusion site failures.

Cost continues to be a barrier to diabetes technology. For people with no insurance and those who are underinsured, CGM and insulin pump therapy are cost prohibitive. In the current study, over half (58%) of the participants had commercial insurance, and 42% were covered by Medicare and/or Medicaid. The Centers for Medicare & Medicaid Services and some commercial plans have strict eligibility requirements and cumbersome paperwork, which also limits access.

Limitations

One limitation of this study is its small sample size, although it did include three times as many patients as the previous Medtronic satisfaction study (7). Impressively, the current study recruited a diverse patient population, including people with type 1 or type 2 diabetes with varying ages, duration of diabetes, and previous diabetes technology use. Moreover, patients could be included in this single-response study if they had used the Control-IQ system for at least 3 months, allowing for a diverse patient population with regard to duration of Control-IQ use. However, this diversity could also be viewed as a limitation because there was no consistency related to duration of use at the time of the survey and because there was no baseline GMSS evaluation before starting the Control-IQ system to compare changes in satisfaction. The lack of a control group could also be viewed as a limitation.

Conclusion

This study evaluated real-life experience with the Control-IQ HCL system. The system offers people with diabetes an additional option and may help them better manage their diabetes with less burden. Participants using the Control-IQ system reported a high rate of overall satisfaction and experienced significant reductions from baseline A1C at 3, 6, and 12 months. Although higher total satisfaction scores did not significantly correlate with lower A1C levels, both results were favorable independently. Future research should evaluate patient satisfaction with the more recently approved systems from Medtronic and Insulet compared to the Control-IQ system.