FormalPara Key Summary Points

Why carry out this study?

The clinical burden of type 1 diabetes in Sweden is growing, and novel treatments could help to keep individuals with the disease within glycemic targets and thereby lower the incidence of costly long-term diabetes-related complications.

However, the benefits of novel efficacious interventions must be judged against the costs of introducing such therapies as healthcare budgets come under increasing strain worldwide.

The present study therefore aimed to evaluate the long-term cost-effectiveness of a novel advanced hybrid closed-loop (AHCL) system (the MiniMed 780G system) versus intermittently scanned continuous glucose monitoring (isCGM) plus multiple daily insulin (MDI) or continuous subcutaneous insulin infusion (CSII) in people with type 1 diabetes in Sweden.

What was learned from the study?

Outcomes projected over patients’ lifetimes indicated that the MiniMed 780G system was associated with improved life expectancy and quality-adjusted life expectancy and increased costs versus is CGM plus MDI or CSII from a societal perspective in Sweden, resulting in an incremental cost-effectiveness ratio of SEK 373,700 per quality-adjusted life year gained.

Based on long-term projections, the MiniMed 780G system was considered a cost-effective treatment option in people with type 1 diabetes in Sweden.

Introduction

Type 1 diabetes is characterized by a loss of beta-cell function and consequent insulin deficiency, and represents one of the most common endocrine disorders in the world, particularly in pediatric populations [1]. In Sweden, the clinical burden of type 1 diabetes is growing, with the disease affecting more than 45,000 people in 2020, a marked increase since records began in 1996 [2]. Prolonged heightened blood glucose levels (measured via glycated hemoglobin [HbA1c]) can have a substantial impact on the incidence of diabetes-related complications over both the short and long term, as evidenced by the landmark Diabetes Control and Complications Trial (DCCT) and the Epidemiology of Diabetes Interventions and Complications (EDIC) study [3, 4]. Diabetes-related complications are associated with a significant economic burden in terms of both direct treatment costs and indirect costs relating to lost workplace productivity. Maintaining glycemic control within target ranges therefore represents the key goal for people with type 1 diabetes and for healthcare payers, as this reduces the incidence of long-term complications and avoids short-term complications such as hypoglycemia as well as the associated cost burdens. With healthcare systems worldwide coming under increasing budgetary strain, novel interventions for type 1 diabetes are often subjected to health economic analyses to evaluate whether they represent value for money for the healthcare payer and for society overall.

Type 1 diabetes requires lifelong treatment and management to avoid substantial morbidity and early mortality. Traditional treatments for type 1 diabetes have relied on self-injection of multiple daily insulin doses (MDI), which requires regular user-administered self-monitoring of blood glucose (SMBG) testing, and these therapies often represent the first-line treatment option for type 1 diabetes. However, innovations in diabetes technologies can help to relieve the burden of self-injection and SMBG testing on the user while offering improvements in glycemic control and subsequent reductions in long-term diabetes-related complications and hypoglycemic events [5,6,7,8,9,10,11,12,13]. Currently available diabetes technologies incorporate both a continuous glucose monitor (CGM, either intermittently scanned [isCGM] or real-time [rtCGM]), which can replace the need for SMBG testing, and insulin pumps for continuous subcutaneous insulin infusion (CSII), which can replace multiple daily self-injections. Combinations of CGM and CSII in sensor-augmented pumps (SAPs) can also embed predictive low-glucose management (PLGM) functions that, due to the communication between the two devices, allows partially automated insulin delivery based on blood glucose levels, improving glycemic control and reducing the risk of hypoglycemia [14, 15]. Mean HbA1c levels have been decreasing in the Swedish population with type 1 diabetes since 1996, and this trend correlates with an increased use of diabetes technologies over this time period [2]. However, the reported mean HbA1c level of people with type 1 diabetes in 2020 was 7.7%, which is above optimal levels [2, 16]. Novel diabetes technologies that can lower HbA1c levels while providing benefits for patients’ quality of life could therefore offer an attractive alternative to traditional MDI or older technological therapy.

Recently, advanced hybrid closed-loop (AHCL) systems have been developed that allow automated insulin administration via CSII in response to fluctuations in blood glucose levels measured by the CGM. The MiniMed 780G represents an AHCL system that automatically adjusts basal insulin delivery, as well as providing safe correction bolus doses as required, via an algorithm that updates insulin delivery every 5 min. This system was evaluated versus SAP therapy with PLGM in a recent study [17]. That randomized, open-label, crossover study, performed at two centers and conducted in automated-insulin-delivery-naïve participants with type 1 diabetes (aged between 7–80 years; n = 60), found that time in range (TIR) was improved with AHCL versus SAP plus PLGM, with higher use of the automated mode and no increase in hypoglycemia [17]. The automated features of AHCL systems could also relieve the burden of insulin administration from people with type 1 diabetes, thereby offering quality-of-life benefits and reduced fear of hypoglycemia [18].

Given the scope for improvement in type 1 diabetes care in Sweden, and the potential benefits offered by AHCL systems, the present study aimed to evaluate the long-term cost-effectiveness of the AHCL MiniMed 780G system compared with isCGM plus either MDI or CSII in the Swedish setting, and thereby estimate whether this system can offer both clinical benefits and value for money.

Methods

Model Overview

The IQVIA CORE Diabetes Model (version 9.0) was used to project clinical and cost outcomes over modeled patients’ lifetimes, in line with guidance on the cost-effectiveness of interventions for diabetes [19]. This model is a validated, web-based, non-product-specific diabetes analysis tool designed to project the long-term health outcomes of novel interventions for the treatment of type 1 and type 2 diabetes [20,21,22]. The model incorporates numerous submodels, each with a semi-Markov structure, that use time-, state-, time-in-state-, and diabetes-type-dependent probabilities derived from published sources. Relevant model outputs include life expectancy, quality-adjusted life expectancy (expressed in quality-adjusted life years [QALYs]), cumulative incidence and time to onset of diabetes-related complications, direct costs arising from the treatment of diabetes-related complications, indirect costs arising from lost workplace productivity, incremental cost-effectiveness ratios (ICERs) based on both direct costs and combined (direct and indirect) costs, as well as cost-effectiveness scatterplots and acceptability curves.

Clinical Data

Baseline cohort characteristics, including age, duration of diabetes, proportion male, HbA1c, systolic and diastolic blood pressure, total cholesterol, high-density lipoprotein (HDL) cholesterol, and body mass index (BMI) were sourced from the FUTURE study, a 12-month, prospective, observational, multicenter, real-world study investigating the impact of isCGM on glycemic control (Table 1) [23]. A total of 1913 people with type 1 diabetes for > 3 months were included, with 78% receiving MDI therapy. The mean age of the cohort was 45.8 years, with a duration of diabetes of 23 years, a proportion male of 53.9%, and a mean HbA1c of 7.8%.

Table 1 Baseline characteristics, treatment effects, and adverse event rates applied in the base case analysis
Full size table

Treatment effects for the MiniMed 780G system and isCGM plus MDI or CSII were sourced from a recent randomized crossover trial and the FUTURE study, respectively (Table 1) [17, 23]. In the MiniMed 780G arm, HbA1c was reduced by 0.5%, and the rate of severe hypoglycemic events (requiring the assistance of a third party) was set to zero, in line with the results from the randomized trial [17]. In the isCGM plus MDI or CSII arm, no changes from baseline in HbA1c were applied, and the rate of severe hypoglycemia (requiring the assistance of a third party) was set to 63.9 events per 100 patient-years, in line with results from the FUTURE study [23]. Following the application of treatment effects in the first year of the analysis, HbA1c was assumed to remain constant for the remainder of each patient’s lifetime.

Changes in other physiological parameters, such as blood pressure and serum lipid levels, were not applied in lieu of treatment-specific data, and baseline values were assumed to follow the model’s default progression equations based on published sources. No changes in BMI were applied, and this parameter was assumed to remain constant over the duration of the analyses. The rate of diabetic ketoacidosis was set to zero in both treatment arms. This approach allowed the impacts of different levels of glycemic control and hypoglycemic events (the two parameters evaluated in the clinical data sources) to be assessed.

Cost Data and Utilities

Costs were accounted from a Swedish societal perspective and expressed in Swedish krona (SEK). Selected cost outcomes were also expressed in euros (EUR), applying an exchange rate of SEK 1.0 = EUR 0.1. All analyses were performed using a first-order Monte Carlo approach, with future clinical and cost benefits discounted at 3.0% per annum, in line with pharmacoeconomic guidance for the Swedish setting [24]. Direct costs captured the costs of treating diabetes-related complications and the costs of patient management, which were taken from published sources (Table 2) [25,26,27,28,29,30,31,32,33,34]. Indirect costs arising from lost workplace productivity were based on the days off work estimates published by Sørensen and Ploug and the 2019 average salaries for men and women in Sweden [35, 36]. Indirect costs were only accrued while simulated individuals were below the set retirement age (64 years) [37].

Table 2 Direct costs associated with the treatment of diabetes-related complications applied in the analyses
Full size table

Pharmacy costs captured the costs of the MiniMed 780G device, the isCGM, basal and bolus insulin, the CSII pump, cannula, and reservoir, as well as the SMBG testing apparatus and training to use the insulin pump. Resource use, including the doses of insulin applied in the MDI arm and the cost of isCGM plus MDI or CSII, was based on a weighted average of patients in the FUTURE study (77.8% receiving MDI and 22.2% receiving CSII) [23]. Annual costs for each treatment arm were calculated based on resource use and pharmacy costs, and totaled SEK 75,644.75 for treatment with MiniMed 780G and SEK 29,971.05 for treatment with isCGM plus MDI or CSII.

Health-state utilities and event-based disutilities associated with diabetes-related complications were taken from published sources [38,39,40,41,42,43,44,45,46]. In the MiniMed 780G arm, a treatment-specific utility increase of 0.0552 was applied to capture the improvement in quality of life associated with reduced fear of hypoglycemic events in patients using SAPs, based on the results of the INTERPRET study (which noted a decrease of 6.9 in the hypoglycemic fear survey) and a quality-of-life translation instrument developed by Currie et al. [47,48,49]. For the isCGM plus MDI or CSII arm, no significant improvement in the hypoglycemic fear survey score was reported in the FUTURE study, and therefore no utility improvement was assumed [23].

Sensitivity Analyses

Modeling the long-term outcomes of diabetes from short-term data is associated with uncertainty. A series of sensitivity analyses were therefore performed to evaluate the robustness of the base case findings. These included evaluating the effect of over- or underestimating the HbA1c benefit with MiniMed 780G by increasing and decreasing the benefit in the 780G treatment arm in several steps between − 0.8% to − 0.4%, and testing the effect of a better-controlled cohort by lowering the baseline HbA1c from 7.8% to 7.5% in the MiniMed 780G arm (based on the findings of the US Pivotal trial) [17].

Compliance with Ethics Guidelines

This article is based on previously conducted studies and does not contain any studies with human participants or animals performed by any of the authors.

Results

Base Case Analysis

Long-term projections indicated that the MiniMed 780G system was associated with an improvement in life expectancy of 0.16 years and an improvement in quality-adjusted life expectancy of 1.95 QALYs versus isCGM plus MDI or CSII (Table 3). The improved clinical outcomes were a result of a reduced incidence and an increased time to onset of diabetes-related complications with MiniMed 780G, as well as quality-of-life benefits arising from reduced fear of hypoglycemia (Fig. 2).

Table 3 Base case analysis results
Full size table

Total direct costs were estimated to be SEK 838,285 (EUR 83,829) higher with the MiniMed 780G system compared with isCGM plus MDI or CSII over a patient’s lifetime, due to the higher acquisition cost of the device and the increased survival and further treatment of individuals over the long term (Fig. 1). However, higher treatment costs were partially offset by cost savings arising from the avoidance of diabetes-related complications, most notably renal complications (mean cost savings of SEK 36,704 [EUR 3,670] per person). Moreover, indirect costs were estimated to be SEK 110,877 (EUR 11,088) lower with the MiniMed 780G system compared with isCGM plus MDI or CSII, leading to combined (direct and indirect) cost increases of SEK 727,408 (EUR 72,741) with the MiniMed 780G system over a patient’s lifetime.

Fig. 1
figure 1

Total costs accumulated in the base case analysis over a patient’s lifetime. CSII continuous subcutaneous insulin infusion, isCGM intermittently scanned continuous glucose monitor, MDI self-injection of multiple daily insulin doses, SEK Swedish krona

Full size image

Estimation of long-term clinical outcomes indicated that both life expectancy and quality-adjusted life expectancy were improved with the MiniMed 780G system compared with isCGM plus MDI or CSII, at an increased cost from a Swedish societal perspective (Table 3). The MiniMed 780G system was therefore associated with an ICER of SEK 373,700 per QALY gained, based on combined costs, versus isCGM plus MDI or CSII. Based on a willingness-to-pay threshold of SEK 500,000 per QALY gained in Sweden (as recommended by the Swedish Agency for Health Technology Assessment [SBU] for high-cost interventions), the MiniMed 780G system was considered a cost-effective treatment option compared with isCGM plus MDI or CSII for the treatment of type 1 diabetes [50].

Sensitivity Analyses

Sensitivity analyses showed that the results of the base case findings were robust to changes in baseline HbA1c and treatment effects (Table 4). Reducing the HbA1c benefit with the MiniMed 780G system to − 0.4% resulted in reduced clinical benefits and cost savings, with the ICER increasing to SEK 387,755 per QALY gained. Conversely, increasing the HbA1c benefit to − 0.6%, − 0.7%, and − 0.8% resulted in increased clinical benefits and cost savings with 780G, yielding ICERs of SEK 358,016, 346,607, and 332,476 per QALY gained, respectively. Reducing baseline HbA1c to 7.5% in the 780G arm (as opposed to 7.8% in the base case analysis) led to increased quality-adjusted life expectancy benefits with the AHCL system, but also increased costs due to the greater survival and further treatment of patients over the long term. The MiniMed 780G system was therefore associated with an ICER of SEK 250,547 per QALY gained versus isCGM plus MDI or CSII.

Table 4 Sensitivity analyses results
Full size table

Discussion

Demonstrating the cost-effectiveness of novel diabetes technologies, particularly AHCL systems that are associated with a high initial outlay, is crucial for increasing the uptake of efficacious treatment options for type 1 diabetes. From a societal perspective in Sweden, the MiniMed 780G system was found to be a cost-effective treatment option compared with isCGM plus MDI or CSII in people with type 1 diabetes. Greater reductions in HbA1c and severe hypoglycemic events yielded improved life expectancy and quality-adjusted life expectancy over the long term, a result of the reduced incidence and increased time to onset of diabetes-related complications with the MiniMed 780G system (Fig. 2). Increased treatment costs were partially offset by cost savings from the avoidance of diabetes-related complications and associated lost workplace productivity. The MiniMed 780G system was therefore projected to offer clinical benefits for people with type 1 diabetes while providing value for money for healthcare payers in Sweden.

Fig. 2
figure 2

Cumulative incidence of diabetes-related complications in the base case analysis. CSII continuous subcutaneous insulin infusion, isCGM intermittently scanned continuous glucose monitor, MDI self-injection of multiple daily insulin doses

Full size image

The selection of isCGM plus either MDI or CSII as a comparator was based on the uptake of CGM and CSII in Sweden. While CGM was in use in 84% of the population with type 1 diabetes in Sweden in 2020, similar levels of uptake have not been observed for insulin pumps, with only 26% of the population covered [2]. Therefore, many patients still rely on MDI when responding to prompts from the CGM, and a comparator combining both approaches was considered appropriate. Baseline cohort characteristics and insulin dose resource use from the FUTURE study, which investigated the impact of isCGM on glycemic control, were also chosen to reflect the current state of type 1 diabetes therapy in Sweden, with a high proportion utilizing CGM but fewer people using insulin pumps [23]. Moreover, these data from the National Diabetes Register indicate that there is scope for improving the treatment of type 1 diabetes in Sweden, with insulin pumps currently underutilized. Given the improvements in HbA1c associated with improved diabetes care and greater use of effective technologies, the support and reimbursement of novel cost-effective technologies should continue. The present study has demonstrated that a next-generation AHCL device is a cost-effective treatment option for people with type 1 diabetes in Sweden. Prescribing such a device could therefore help to further improve diabetes care, considering the high mean levels of HbA1c and low insulin pump uptake still observed in people with type 1 diabetes in Sweden [2, 16].

Fear of hypoglycemia was a key factor in the present analysis. Fear of hypoglycemia is common in populations with type 1 diabetes, and can affect quality of life both directly and indirectly through an increased burden of self-management [51, 52]. A history of severe hypoglycemic events, as well as certain situations (such as driving or being in the workplace, particularly an industrial setting), can also drive an increased fear of future events, and can act as a barrier to physical activity or other activities, further limiting patients’ quality of life [51, 53, 54]. The psychological burden of fear of hypoglycemia should also be considered, as this can affect both adult patients and parents or caregivers of younger individuals with type 1 diabetes [55, 56]. Given the reductions in hypoglycemia observed with the AHCL system in the recent trial and applied in the present analysis, the application of an additional utility for avoidance of fear of hypoglycemia was considered appropriate to capture not only the physical effects of hypoglycemia but also the associated psychological distress [17].

A limitation of the analysis, inherent to all long-term cost-effectiveness analyses, is the projection of long-term outcomes from short-term clinical data. That acknowledged, this is one of the essential tenets of long-term health economic modeling, and it remains one of the best available options for informing decision making in the absence of long-term clinical trial data, with guidelines for computer modeling of diabetes interventions recommending the projection of outcomes over patients’ lifetimes [19]. Moreover, every effort was made to limit uncertainty by using a published and validated health economic model and conducting sensitivity analyses around model inputs [20,21,22]. The use of data from a crossover trial could also be viewed as a weakness of the analysis. However, no randomized head-to-head study comparing an AHCL system with CGM in combination with either MDI or previous-generation devices has been published, and the crossover trial therefore represented the most robust evidence source currently available. Future clinical studies should focus on elucidating potential benefits for glycemic control and hypoglycemia in head-to-head comparisons of diabetes technologies, with subsequent health economic analyses utilizing these data to provide further measurements of cost-effectiveness for novel devices.

The inability to include TIR data could also be seen as a limitation. Published evidence has indicated that the MiniMed 780G system is associated with improved TIR and reduced hyperglycemia versus SAP plus PLGM and a previous HCL system, respectively [17, 57]. In that sense, the exclusion of these data was conservative from the perspective of the 780G system. Nonetheless, as TIR becomes more central to country-specific recommendations for treatments for type 1 diabetes, novel modeling methods should be developed to utilize these data and to help further inform decision makers.

Conclusions

Long-term projections indicate that the AHCL MiniMed 780G system is likely to represent a cost-effective treatment option versus isCGM plus MDI or CSII in people with type 1 diabetes in Sweden.