INTRODUCTION
Diabetes imposes a substantial global disease burden,1, 2 affecting an estimated 589 million adults aged 20–79 years (11.1% of this age group) in 2024.3 Projections indicate this number will rise to 852.5 million (13.0%) by 2050, paralleled by healthcare expenditures that reached $1.015 trillion in 2024 and are expected to exceed $1.043 trillion by 2050.3 China has the world's largest diabetes population,4 with 148.0 million adults living with diabetes in 2024,3 96% of whom have type 2 diabetes (T2D).1 However, only 16.5% of T2D patients achieved adequate glycaemic control.2 In 2021, T2D accounted for 174 515 deaths in China and a disability‐adjusted life year (DALY) rate of 569.8 per 100 000 population.5 While numerous glucose‐lowering therapies have emerged in recent years, many adults with T2D eventually require and benefit from insulin therapy.6 However, the complexity of treatment regimens, frequent injection requirements, and suboptimal adherence to insulin therapy remain significant challenges.7, 8
Insulin icodec, the first once‐weekly basal insulin analogue, received concurrent market approval in China, the European Union, and other countries in 2024, with subsequent inclusion in Chinese diabetes guidelines.9 Characterized by a prolonged half‐life of 196 h, which supports stable and continuous glucose control,10 its efficacy and safety were robustly demonstrated in the global ONWARDS Phase 3a trials.11 Compared with insulin degludec, a daily basal insulin with the longest half‐life, icodec illustrated superior haemoglobin A1c (HbA1c) reduction with a low incidence of clinically significant hypoglycemia.12 These therapeutic advantages were consistently observed in insulin‐naive Chinese T2D patients within ONWARDS 3.13 By reducing the frequency of basal insulin injections from at least 365 per year to just 52, icodec holds significant potential to enhance patient satisfaction, improve adherence, and ultimately lead to better health outcomes.
Icodec was incorporated into China's National Reimbursement Drug List (NRDL) in 2024 based on health technology assessment (HTA),14 positioning China among the first batch of countries to cover this innovative drug in social health insurance. Several economic evaluations have suggested potential value for icodec under projected price scenarios. For example, a 1‐year simulation within Italian national health service context indicated dominance versus degludec at expected prices,15 while another China‐focused analysis estimated a cost‐effective price range using global clinical trial data.16 More recently, Xie et al. applied actual prices and reported that icodec was cost‐effective versus degludec among insulin‐naïve patients but not in individuals already treated with basal or basal‐bolus regimens.17 Despite these advances, important evidence gaps remain that limit the relevance of existing studies for reimbursement decision‐making.
First, the financial consequences of reimbursing icodec for the healthcare system have not been evaluated, as no budget impact analysis (BIA) has been conducted to quantify payer affordability. Second, prior cost‐utility analyses were constrained by methodological limitations, including short simulation horizons, incomplete modelling of treatment intensification, omission of utility gains associated with reduced injection frequency, and the use of global trial data rather than clinical evidence specific to the target population. Consequently, they may fail to capture the comprehensive long‐term value of icodec.
To address these gaps, this study aims to evaluate both the long‐term cost‐utility and the budget impact of icodec from the perspective of the Chinese healthcare system, utilizing the NRDL price and clinical data from the Chinese cohort of the ONWARDS 3 trial.
METHODS
Modelling approach
This study employed the Swedish Institute of Health Economics Diabetes Cohort Model (IHE‐DCM) (version 4.4.2) to evaluate the long‐term cost‐utility of once‐weekly icodec compared to once‐daily degludec in patients with T2D in China from the Chinese healthcare system perspective.
The validated IHE‐DCM, a peer‐reviewed and widely implemented simulation tool in diabetes economic evaluations18, 19, 20, 21, 22, 23 employs Markov health states to quantify diabetes‐related complications (Figure 1). Developed in Microsoft® Excel 2013 with Visual Basic for Applications, the model projects microvascular and macrovascular complication incidence through annual cycles over a 40‐year maximum horizon. Long‐term outcomes were projected based on baseline characteristics, treatment regimens and their effects, costs of treatment and complications, and health utility values associated with diabetes complications.
The model further quantifies the contribution of individual factors to outcomes. We incorporated mortality risk equations from the United Kingdom Prospective Diabetes Study (UKPDS) 6824 and UKPDS 82.25 Macrovascular risk equations were derived from UKPDS 68, UKPDS 82, the Swedish National Diabetes Register,26 or the Australian Fremantle Diabetes Study.27 For microvascular risks, a set of equations sourced from other established diabetes models was used.28, 29, 30
A base‐case analysis was conducted with a 40‐year time horizon to capture outcomes over lifetime as much as possible. Future outcomes were discounted at an annual rate of 5%, consistent with the Chinese Guidelines for Pharmacoeconomic Evaluation.31 To ensure transparency and reproducibility, this study adhered to the Consolidated Health Economic Evaluation Reporting Standards (CHEERS)32 (see Table S1, Supporting Information).
Clinical data
Baseline cohort characteristics were primarily derived from the Chinese subgroup (n = 145) of the multinational, head‐to‐head ONWARDS 3 trial comparing icodec versus degludec in insulin‐naive adults with T2D.13, 33 This 26‐week trial provided key baseline parameters including mean baseline age (55.43 years), female proportion (42.80%), and diabetes duration (9.60 years) (Table 1). Other baseline parameters were sourced from published studies,25, 34 as detailed in Table S2. Selection criteria for source of baseline characteristics is detailed in Data S1.
In the base‐case analysis, patients initiated treatment with either insulin icodec or degludec. Treatment intensification to basal‐bolus insulin (glargine U100 plus aspart) occurred if HbA1c exceeded 8.31% for both groups.8 Clinical efficacy parameters, including changes in HbA1c, body mass index (BMI), and hypoglycaemia incidence, were exclusively derived from the Chinese ONWARDS 3 subgroup. After applying first‐year treatment effects, HbA1c was assumed to increase annually by 0.14% throughout the time horizon, consistent with previous economic evaluations.22, 23 Other biomarkers (e.g., BMI) remained stable after the first year.
The effects of intensification therapy were derived from ONWARDS 4.8
Cost and utility data
All costs were expressed in 2024 US dollar using an exchange rate of 1US$ = 7.1884 CNY (31 December 2024). Annual treatment costs incorporated insulin expenses, needle utilization, and complication management. For icodec (700 U, 1 mL), drug costs were calculated based on its 2024 NRDL price, whereas degludec and glargine U100 costs were calculated by prices sourced from China's 2024 national volume‐based procurement (VBP) data.
Needle costs were included according to injection frequency, with dosing regimens derived from clinical trials under the assumption of 100% treatment adherence. Costs for diabetes complications and hypoglycaemia events were obtained from published cost studies specific to Chinese T2D populations, as detailed in Table S3.
Health utilities were derived from published studies of Chinese insulin‐treated T2D populations35 and an Asian diabetes utility meta‐analysis,36 capturing quality‐of‐life impacts from complications. Injection‐related disutility was quantified using preference data specific to Chinese patients.37
Sensitivity analyses
As inherent uncertainties in projecting long‐term outcomes from short‐term trial data, the robustness of the base‐case results was evaluated through wide‐ranging sensitivity analyses, including time horizon, discount rate, treatment algorithm, HbA1c drift rates, cost of degludec, treatment‐related costs of complications, disutility value. Notably, we evaluated icodec's 1050 U (1.5 mL) formulation, available in China at different unit prices, versus degludec in one‐way sensitivity analysis. All one‐way analyses varied individual inputs and assumptions.
Probabilistic sensitivity analysis employed nonparametric bootstrapping with 1000 Monte Carlo simulations to quantify joint parameter uncertainty.
Budget impact analysis
A budget impact analysis was conducted to estimate the financial implications of including icodec in NRDL over a 3‐year time horizon from 2025 to 2027. The analysis compared two scenarios: a reimbursement scenario in which icodec is covered by social health insurance and a counterfactual scenario in which icodec is not reimbursed. The difference in total insurance expenditures between the two scenarios represents the budget impact attributable to the inclusion of icodec.
The size of the target population was estimated based on national epidemiological data on diabetes in China and related projections (see Table S4). Substitution patterns between icodec and other major basal insulins were estimated based on current market shares and forecasted market uptake after reimbursement. We considered drug costs, injection needle costs, and costs associated with hypoglycaemic events.
Reimbursement rates were determined according to the rules of the NRDL and published literature.38
RESULTS
Base case analysis
The 40‐year simulation indicated similar overall mortality (RR = 1.00) and cardiovascular‐specific mortality (RR = 0.99) for both once‐weekly icodec and once‐daily degludec. However, icodec was associated with a cumulative reduction in both microvascular and macrovascular complications, particularly in eye diseases and kidney diseases (see Table S5).
The discounted life expectancy was 13.420 years for the icodec and 13.395 years for degludec, with corresponding discounted quality‐adjusted life expectancy of 8.857 QALYs and 8.641 QALYs, respectively. Compared to degludec, icodec was associated with an increase of 0.025 years in life expectancy and a gain of 0.216 QALYs, which was mainly attributable to better efficacy and lower injection frequency (see Table S6).
Over the time horizon, the total costs were $1450.73 lower with icodec versus degludec. This reduction primarily stemmed from lower expenditures in three categories: anti‐hyperglycaemic treatment (−$834.70), macrovascular complications (−$563.09), and microvascular complications (−$52.94) for icodec (Table 2). Notably, the largest cost savings were observed in anti‐hyperglycaemic treatment (see Figure S1).
The cost‐utility analysis demonstrated that icodec was dominant over degludec in patients with T2D in China, offering greater long‐term clinical benefits at a lower cost.
Sensitivity analysis
One‐way sensitivity analyses supported the robustness of base‐case results, with icodec consistently dominant over degludec across variations in key parameters (Table 3). The highest incremental quality‐adjusted life expectancy for icodec (0.464 QALYs) occurred when disutility values for injection frequency were drawn from South Korean and Taiwanese patients with T2D.39 The greatest cost savings ($4166.30) occurred under the pre‐VBP degludec pricing scenario, of which $3550.30 was due to lower costs of anti‐hyperglycaemic treatment. When the annual drug cost of icodec increased by 35%, it remained dominant.
Even when increased by 125%, icodec was no longer dominant but remained cost‐effective, with an ICER of $13085.62 per QALY, below China's per capita GDP threshold ($13319.93/QALY).
Probabilistic sensitivity analyses yielded results consistent with the base case. On average, icodec provided an additional 0.212 QALYs compared with degludec while reducing direct medical costs by $1320.46. At a willingness‐to‐pay threshold equal to China's per capita GDP, the probability that icodec was dominant over degludec was 100% (Figure 2).
Budget impact analysis
The inclusion of icodec in the NRDL would reduce insurance fund expenditures by $1.17 million, $2.35 million, and $2.58 million in 2025, 2026, and 2027, respectively. When only drug costs were considered, reimbursement of icodec was associated with modest increases in fund expenditures of $3.14 million, $6.27 million, and $6.90 million over the same period (see Table S7).
DISCUSSION
This study presents the first comprehensive economic evaluation of once‐weekly insulin icodec to integrate a long‐term cost‐utility analysis with a budget impact analysis under real‐world reimbursement conditions in China. Using NRDL price and clinical efficacy data from the Chinese subgroup of ONWARDS 3, our analysis demonstrates that icodec is a dominant strategy compared with once‐daily insulin degludec, delivering greater long‐term health benefits at lower overall costs for patients with T2D in China. The robustness of these findings was supported by extensive one‐way and probabilistic sensitivity analyses across a wide range of model inputs and assumptions.
Notably, the budget impact analysis revealed that the reimbursement of icodec is projected to reduce the financial burden on the social health insurance fund, thereby validating its value proposition from both efficiency and affordability perspectives.
Compared to once‐daily degludec, once‐weekly icodec was associated with 0.025 years increase in life expectancy and 0.216 QALYs gain. Superior glycaemic control was one of the key contributors, which exerted clinical effects through two distinct mechanisms. First, icodec's significantly greater reduction in HbA1c compared to degludec delayed initiation of intensification therapy by 2 years (icodec: 10 years; degludec: 8 years), consequently reducing exposure to hypoglycaemia risk inherent in intensification regimens.8 Second, better glycaemic control directly reduced cumulative complication incidence, including background retinopathy (34.75% vs.
39.28%), proliferative retinopathy (3.79% vs. 4.77%), proliferative retinopathy and macular oedema (3.54% vs. 4.61%), neuropathy (7.94% vs.
8.87%), macroalbuminuria (16.21% vs. 18.48%), end‐stage renal disease (8.87% vs. 9.89%), and cardiovascular disease (30.13% vs.
30.52%), thereby prolonging life expectancy and enhancing quality of life. The superior efficacy of icodec also translated into long‐term cost savings, including reduced costs on diabetes‐related complications and costs of intensification therapy. These savings fully offset the higher treatment costs associated with icodec itself, ultimately leading to lower overall cost for the icodec group versus degludec group ($39713.13 vs. $41163.86).
Reduced injection frequency constituted another key contributor to the icodec's long‐term clinical benefits, operating through direct quality‐of‐life improvements and enhanced treatment adherence. On the one hand, our modelling quantified a 0.101‐QALY improvement specifically to injection frequency reduction, establishing it as the primary driver of icodec's clinical advantage over degludec. This aligns with one‐way sensitivity analyses where the change in injection‐related disutilities generated the largest QALY gains.
On the other hand, weekly dosing potentially addresses the well‐documented adherence challenges in type 2 diabetes. Suboptimal adherence remains prevalent in insulin therapy for T2D.40 A systematic review reported adherence rates was only 62–64% for daily insulin regimens, with even lower rates approximating 33% among younger patients.41 Evidence from glucagon‐like peptide‐1 receptor agonists demonstrated superior adherence and glycaemic outcomes with weekly versus daily administration.42 Although direct evidence linking insulin injection frequency to adherence and improved glycaemic control was limited, it was reasonable to believe that similar benefits could be achieved with insulin regimens. Our analysis incorporated data from a randomized clinical trial and assumed 100% adherence in both groups.
This approach might underestimate real‐world adherence differentials, suggesting that modelled quality‐adjusted life year gains represent conservative estimates. Actual benefits could be amplified through improved adherence facilitated by weekly dosing in clinical practice.
Similarly, the reduced injection frequency and lower incidence of hypoglycaemia substantially contributed to the budget savings associated with reimbursing icodec under the social health insurance scheme. This finding further underscores the value of icodec and provides important support for its inclusion in the national reimbursement system, particularly in the context of ongoing concerns about the sustainability of the insurance fund. The dominant results of our analysis align with findings from the Italian evaluation15 but differ from those reported by Xie et al.,17 who concluded that icodec was cost‐effective in insulin‐naive individuals with type 2 diabetes.
Several methodological differences may account for these divergent outcomes. Our study used clinical data from the Chinese subgroup of the ONWARDS 3 trial, which may more accurately reflect treatment effects in the Chinese population, whereas Xie et al. applied global trial data. In addition, two important contributors to the incremental QALYs observed in our evaluation, the reduction in hypoglycaemia risk associated with delayed treatment intensification and the improvement in health‐related quality of life associated with fewer injections, were not incorporated in Xie et al.'s model.
The omission of these elements may partially explain the differences in long‐term cost‐utility results between the two studies.
Although recent reforms in drug pricing in China have substantially reduced insulin costs,43, 44 our findings are likely to be generalizable beyond this setting. Two observations support this inference. First, the price of the comparator drug, once‐daily degludec, decreased by 40% in 2024 under the VBP, compared with the cost used in a prior Chinese evaluation.16 Despite this, once‐weekly icodec remained a cost‐saving strategy in our model.
Second, as insulin prices in China are among the lowest globally,43 we conducted one‐way sensitivity analyses with higher prices for icodec. Even when the price of icodec was doubled, it remained cost‐effective compared with degludec. Therefore, these findings suggest that differences in drug pricing across health systems are unlikely to alter the primary inference: once‐weekly icodec represents a cost‐effective alternative to insulin degludec.
Strengths and limitations
This study has several key strengths. First, it is the first integrated evaluation of icodec, the first once‐weekly basal insulin analogue, by combining a comprehensive long‐term cost‐utility analysis with a budget impact analysis under real‐world reimbursement conditions in China. As China is the first country to include icodec in its social health insurance coverage based on HTA evidence, this analysis provides timely, valuable insights for the reimbursement decision of icodec in other health systems.
Second, the inclusion of Chinese subgroup data from the ONWARDS 3 trial and incorporation of the full spectrum of diabetes treatment enhance the credibility of the results. Third, by evaluating the contribution of each factor, we clearly identify the primary value drivers of icodec versus degludec.
As with all health economic modelling studies, this analysis has several limitations that should be considered when interpreting the results. First, long‐term modelling projections are inherently dependent on short‐term trial data and historical evidence, introducing a degree of uncertainty. To mitigate this, we conducted extensive sensitivity analyses.
Second, due to the relatively short time since the launch of insulin icodec, there is currently a lack of real‐world evidence on the impact of once‐weekly insulin on adherence and health outcomes, which limits our ability to fully assess the value of icodec. If relevant real‐world evidence were available, the long‐term health outcome results for insulin icodec would likely be more favourable. Third, our model relies primarily on changes in HbA1c and BMI as key inputs for complication risk equations.
While we acknowledge that other biomarkers may influence long‐term complication risk, HbA1c remains the gold standard for predicting glycaemia‐associated risks for the microvascular and macrovascular complications,45 and BMI is a well‐established determinant of cardiovascular risk and mortality.46 Fourth, the IHE‐DCM model was not specifically developed for the Chinese population and may not fully capture the epidemiological and clinical characteristics of Chinese T2D patients. Finally, the modelled cohort was constructed using three data sources that either included a limited number of Chinese individuals with type 2 diabetes or originated from a UK population, which may affect its representativeness for the broader Chinese setting.
CONCLUSIONS
Icodec represents a highly efficient therapeutic option for patients with T2D requiring basal insulin therapy. Its innovative once‐weekly administration has the potential to enhance patient adherence and improve health outcomes. Compared to once‐daily insulin degludec, once‐weekly insulin icodec is dominant at its NRDL price in China, offering improved long‐term health outcomes for T2D patients at a lower cost, while also generating potential annual savings.
FUNDING INFORMATION
This study is supported by the National Natural Science Foundation of China (Grant No. 72374049), and Novo Nordisk (China) Pharmaceutical Co. Ltd., Beijing, China.
CONFLICT OF INTEREST STATEMENT
Novo Nordisk (China) Pharmaceutical Co. Ltd., Beijing, China provides funding for financial compensation of the authors and for publication fees. Two authors are full‐time employees of Novo Nordisk (China) Pharmaceutical Co.
Ltd.; additionally, Fengyu Sun (Novo Nordisk (China) Pharmaceutical Co. Ltd.) contributed to the study design and manuscript revision in a non‐author capacity. Except for the contributions of these individuals, the sponsor had no role in the design, data collection, data analysis, and reporting of this study.
Supporting information