Cost‐effectiveness analysis of Dexcom ONE+ real‐time continuous glucose monitoring versus self‐monitoring of blood glucose in people with type 2 diabetes in Aotearoa New Zealand

Abstract

Aims

Real‐time continuous glucose monitors (rt‐CGM) have been found effective and economical for the treatment of diabetes in many countries. The objective of this study was to provide a cost‐effectiveness analysis of rt‐CGM versus self‐monitoring of blood glucose (SMBG) from the perspective of a healthcare payer in New Zealand.

Materials and Methods

The cost‐effectiveness of rt‐CGM in patients with type 2 diabetes (T2D) who require intensive insulin therapy was analysed using the IQVIA Core Diabetes Model (CDM), providing outputs including life expectancy, quality‐adjusted life years (QALYs), direct costs, incremental cost‐effectiveness ratios (ICERs), and incidence rates of complications. A lifetime (50 years) time horizon was used with an annual discount rate of 3.5%.

Results

In the base case, rt‐CGM was associated with a gain of 0.488 QALYs and incremental costs of NZD 5633 compared with SMBG, resulting in an ICER of NZD 11 533 per QALY. This corresponded to a gain of 87 QALYs per NZD 1 million invested. Scenario analyses suggested that CGM is potentially cost saving at earlier ages of rt‐CGM initiation and among high‐risk populations, such as Māori and Pacific Peoples, yielding higher gains in QALYs at lower total direct costs. Reductions were predicted in the risks of ophthalmic, renal, peripheral, and cardiovascular complications.

Conclusions

This analysis provides insight into the cost‐effectiveness of rt‐CGM versus SMBG from the perspective of payers in New Zealand, demonstrating reductions in the risks of complications in addition to reductions in their associated costs.

1. INTRODUCTION

For patients with type 2 diabetes (T2D), the maintenance of glycaemic control is crucial to prevent microvascular complications, quality of life impacts, and premature mortality.^{1, 2, 3} The New Zealand Society for the Study of Diabetes (NZSSD) recommends a target HbA1c for most patients of <53 mmol/mol (<7%) to prevent long‐term complications and early death. ⁴ However, less than half of New Zealanders with T2D reach the target HbA1c of <53 mmol/mol despite the availability of funded self‐monitoring of blood glucose (SMBG) levels.^{5, 6} Furthermore, studies have shown that despite diabetes treatment and glucose monitoring with SMBG, people with T2D still experience hypoglycaemic events, which can contribute to the development of micro‐ and macrovascular complications.^{7, 8, 9}

The achievement of adequate glycaemic management can be supported by the use of real‐time continuous glucose monitoring (rt‐CGM), with devices that can alert users about hyper‐ or hypoglycaemic events in addition to providing tools to analyse trends. ¹⁰ The benefits of rt‐CGM in reducing HbA1c levels and improving glycaemic management have been demonstrated in numerous studies^{11, 12, 13, 14, 15} and it has been shown to be a cost‐effective intervention in many countries.^{16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27} As a result, internationally, several bodies have recommended or reimbursed the use of continuous glucose monitoring (CGM), particularly for people requiring insulin treatment.^{28, 29, 30, 31, 32} In Aotearoa New Zealand, NZSSD currently recognizes that rt‐CGM can be very beneficial when necessary information cannot be obtained from capillary blood glucose monitoring for whatever reason, posing a risk to patient safety or affecting timely treatment decisions. ⁴ However, CGM is only funded for people with type 1 diabetes (T1D), type 3c, and other less common forms of diabetes. ³³ Furthermore, the cost‐effectiveness of rt‐CGM in the management of people with T2D in Aotearoa New Zealand is yet to be established.

The impact of such interventions to improve glycaemic control could be substantial in Aotearoa New Zealand as diabetes is increasingly common in the country, with a study of national‐level data finding a total of 268 248 cases between 2015 and 2019 (age‐standardized rate 3.9%) and projecting that this number will rise to 502 358 (5.0%) by 2040–2044. ³⁴ The incidence of T2D in particular has been found to be rising rapidly, with an approximately 5% per year increase in incidence in children between 1995 and 2015. ³⁵ Such increases in patient numbers have contributed to a substantial nationwide economic burden, with a recent report estimating the total annual cost of T2D to be NZD 2.1 billion, representing 0.67% of the country's total gross domestic product (GDP). ³⁶

In Aotearoa New Zealand, Māori and Pacific Peoples are at a greater risk of T2D and are more frequently impacted by diabetes‐related complications than New Zealanders of European descent.^{37, 38} A study conducted between 1994 and 2018 found that Māori patients have consistently higher hospital admission rates than European patients. ³⁷ Māori people with T2D also have higher rates of all‐cause mortality, cardiovascular mortality, and cancer mortality. ³⁷ Higher rates of end‐stage renal disease due to T2D have also been observed among Māori and Pacific Peoples. ³⁹ It has been hypothesised that such differences may be driven by increased hyperglycaemia among these populations. ³⁸ Approaches to reduce these disparities are therefore of great importance. ³⁸ While the funding of CGM for people with T1D and other less common forms of diabetes is an important step towards improving care for these groups, a significant unmet health need remains for people with T2D. The Pharmaceutical Management Agency (Pharmac) has acknowledged this need and encouraged the submission of funding applications for T2D, as there is clinical evidence supporting the use of CGM among individuals with the disease. ⁴⁰

Despite the available clinical evidence, the cost‐effectiveness of rt‐CGM for the management of patients with T2D receiving insulin therapy in Aotearoa New Zealand is currently unestablished. The objective of this analysis was therefore to evaluate the cost‐effectiveness of rt‐CGM when compared with SMBG in patients with T2D receiving intensive insulin therapy in the country. In this article, we demonstrate that rt‐CGM is cost‐effective from the payer perspective compared with the standard of care in our base case analysis and almost all scenarios analysed.

2. MATERIALS AND METHODS

2.1. Model structure

The analysis of the cost‐effectiveness of rt‐CGM was conducted using the IQVIA Core Diabetes Model (CDM, v10.0).^{41, 42, 43} The model structure has previously been described elsewhere.^{41, 42, 43} The CDM consists of 17 interdependent Markov sub‐models which predict the long‐term health outcomes and costs of diabetes interventions. The progression of risk factors is modelled using equations derived from studies such as the Framingham Heart Study and the United Kingdom Prospective Diabetes Study.^{44, 45} Model outputs include life expectancy, QALYs, direct costs, indirect costs, incremental cost‐effectiveness ratios (ICERs), and complication incidences.

2.2. Baseline cohort characteristics

Baseline characteristics were informed by a US retrospective cohort study. ¹¹ The cohort had a mean age of 64.5 years and a duration of diabetes of 16 years. A baseline HbA1c of 67 mmol/mol was assumed. ¹¹ Based on the EPiC data available in June 2024, a mixed population consisting of New Zealanders of European descent (50.3%), Māori (16.8%), and Asian & Pacific Peoples (32.9%) was included to reflect the ethnic distribution in Aotearoa New Zealand. ⁴⁶ Other baseline characteristics are summarized in Table S1. Baseline utility scores were informed by the UK Prospective Diabetes Study (UKPDS). ⁴⁷

2.3. Clinical effectiveness

Treatment effect parameters were sourced from a real‐world evidence (RWE) cohort study reported by Karter et al., with an HbA1c effect of −0.56% (6.1 mmol/mol). ¹¹ In addition to the treatment effect, an annual increase in HbA1c of 1.64 mmol/mol was included for the SMBG arm after Year 1 and for the rt‐CGM arm after Year 3, reflecting what has been observed in longitudinal studies showing that improvements can last up to 10 years. ⁴⁸

Hypo‐ and hyperglycaemic event rates were based on the Karter et al. cohort study in which relative differences between rt‐CGM initiators and non‐initiators of 4 events per 100 patient years and 2.5 events per 100 years, respectively, were reported. ¹¹ As these were relative differences in event rates, they were applied to the SMBG arm while the corresponding rates of the rt‐CGM arm were set to zero.

2.4. Costs

Costs of medication use reflect current subsidies on the Pharmaceutical Schedule (as of July 2024). Other cost estimates were inflated to 2024 NZD. Screening costs for retinopathy, microalbuminuria, and gross proteinuria were assumed to be represented by one general practitioner (GP) visit at NZD 80 per episode, based on the current unit cost recommended by the Cost Resource Manual. Other specific cost values used, in addition to their sources, are provided in Table S2.

The costs of SMBG were based on a mean of 3.8 fingerstick tests per day, as reported from the DIAMOND T2D study, ¹⁴ with each test costing NZD 0.21 per strip. No other resource use was considered in relation to SMBG. The costs of rt‐CGM were based on the local listing price of NZD 2448 per annum.

2.5. Utilities

A baseline utility value of 0.785 was used, based on the UKPDS. ⁴⁷ Disutilities were sourced from a systematic review and meta‐analysis reported by Beaudet et al. 2014. ⁴⁹ For hypoglycaemic events, disutilities were sourced from Evans et al. 2013 ⁵⁰ and adjusted to reflect the T2D population. ⁴⁹ Specific utility values used are listed in Table S1. Due to findings that the disutility associated with each non‐severe hypoglycaemic event decreases with each event occurring throughout the year, ⁵¹ an adjustment that reduces the utility decrement over time was included in the model. This was based on the algorithm developed by Lauridsen et al. 2014. ⁵¹

As fingerstick testing is associated with burdens such as pain, time loss, and unwanted attention, ⁵² a utility benefit of 0.03 was included to represent the benefits of avoiding finger sticks, sourced from a study reported by Matza et al. ⁵³

2.6. Discounting, time horizon, and willingness‐to‐pay

Modelling was carried out from the perspective of the national healthcare payer in Aotearoa New Zealand. A remaining lifetime (50 years) time horizon was used with an annual discount rate of 3.5% applied to future cost and effectiveness outcomes.

In Aotearoa New Zealand, Pharmac does not use a willingness‐to‐pay (WTP) threshold. ⁵⁴ However, previous analyses have derived estimates of incremental spending using back calculations. A cost‐utility analysis of diabetes remission in people recently diagnosed with T2D, carried out in 2022, adopted a threshold of NZD 30 000 based on such approaches ⁵⁵ and this threshold has been adopted in this analysis as a conservative estimate of cost‐effectiveness.

2.7. Sensitivity analyses

A one‐way sensitivity analysis was conducted to assess the robustness of the outcomes. Parameters varied in these analyses included the costs of rt‐CGM, the HbA1c treatment effect, changes in the ethnic distribution, age at initiation, discount rate, and time horizon. The magnitude of the impact of rt‐CGM on HbA1c was varied by ±30%, and set to −0.30% (−3.28 mmol/mol) based on data reported from the DIAMOND T2D randomized controlled trial (RCT). ¹⁴ The sensitivity analysis also incorporated an additional utility representing reductions in the fear of hypoglycaemia (FoH) that patients may experience. This utility value was based on an analysis of the DIAMOND RCT reported by Polonsky et al. ⁵⁶ This FoH utility (0.02536) was additive to the utility for the avoidance of fingerstick testing (0.03), giving a total utility benefit of 0.05536. Patients in the SMBG arm were assumed to have no FoH benefit. An additional scenario using the Fremantle Cardiovascular Risk Equation ⁵⁷ and the Combined Western Australian Mortality Risk Equation ⁵⁸ was tested to assess the impact of a more specific risk engine, despite differences between the Australian and the New Zealand populations.

2.8. Projected clinical outcomes

Absolute risk reductions (ARR) and relative risks (RR) were calculated for all diabetes‐related complications included in the modelling based on the UKPDS 82 Risk Equation. These calculations were carried out for the mixed population, a 100% European population, a 100% Māori population, a 100% Pacific Peoples population, and a 50% Māori/50% Pacific Peoples population. The CDM unfortunately does not offer Māori ethnicity as an option for risk calculations; for the purpose of the current analysis, Māori people are allocated into risk estimations developed for Black people to capture the higher complication risks expected for this patient population segment.

3. RESULTS

In the base case, rt‐CGM was associated with a gain of 0.488 QALYs, in addition to a life expectancy gain of 0.240 years, compared with SMBG (Table 1). Although the costs of glucose monitoring were higher with rt‐CGM (NZD 25 336 vs. 2992), this was largely offset by reductions in costs associated with renal complications, ophthalmic complications, and the management of severe hypoglycaemia/hyperglycaemia, including diabetic ketoacidosis (DKA), resulting in a difference in total direct medical costs of NZD 5633. Overall, this resulted in an ICER of NZD 11 533 per QALY and a gain of 87 QALYs per NZD 1 million invested.

TABLE 1.

Summary of base case findings.

Base case costs and outcomes	rt‐CGM (Dexcom ONE+)	SMBG	Difference a
Glucose monitoring costs, NZD (list pricing)	25 336	2992	22 345
Management (preventative screening, medication) costs, NZD	1331	1292	39
CV complications, NZD	90 915	90 641	273
Renal complications, NZD	37 472	47 247	−9775
Ulcer/amputation/neuropathy complications, NZD	48 773	49 089	−317
Ophthalmic complications, NZD	31 294	34 575	−3282
Severe hypoglycaemia (requiring medical assistance)	0	1990	−1990
Hyperglycaemia and DKA, NZD	0	1661	−1661
Life expectancy	9.678	9.439	0.240
QALYs	7.114	6.626	0.488
ICER	NZD 11 553/QALY
QALYs per NZD 1 million invested	87 QALYs per NZD 1 million

Abbreviations: CV, cardiovascular; DKA, diabetic ketoacidosis; ICER, incremental cost‐effectiveness ratio; NZD, New Zealand dollar; QALY, quality‐adjusted life year; rt‐CGM, real‐time continuous glucose monitoring; SMBG, self‐monitoring of blood glucose.

^a Negative difference in costs favours rt‐CGM over SMBG, while a positive difference favours SMBG.

3.1. Sensitivity analyses

The robustness of the results was examined through sensitivity analyses. A scatter plot showing the results of a probabilistic sensitivity analysis, in addition to the deterministic base case result, is provided in Figure 1. A cost‐effectiveness acceptability curve from the probabilistic base case is displayed in Figure 2. The results of the one‐way analyses are summarized in Table S4 and displayed graphically in Figure 3. At a suggested willingness‐to‐pay threshold of NZD 30 000, rt‐CGM was cost‐effective in 67.3% of the iterations. Reducing the costs of rt‐CGM by 20% and increasing them by 20% resulted in a decrease in the ICER to NZD 1158 and an increase to NZD 21 926, respectively. The ICER was also somewhat sensitive to the magnitude of HbA1c effect, decreasing to NZD 6220 per QALY when it was increased by 30%. This corresponded to an increase of 161 QALYs gained per NZD 1 million invested and was largely driven by changes in the total direct costs (Table S4). When the effect was decreased, the ICER increased to NZD 17 379 per QALY for a 30% decrease and changed to NZD 19 860 per QALY when it was set to a value of 0.30% based on the results of the DIAMOND study. ¹⁴

FIGURE 1.

Cost‐effectiveness scatterplot from the probabilistic base case analysis. Scatter plot of results from the probabilistic base case sensitivity analysis. The green diamond represents the mean probabilistic sensitivity analysis result while the blue triangle represents the deterministic base case result. The green oval is a data ellipse at the 95% level based on a multivariate T‐distribution. PSA, probability sensitivity analysis; QALE, quality‐adjusted life expectancy; QALY, quality‐adjusted life year.

FIGURE 2.

Cost‐effectiveness acceptability curve from the probabilistic base case analysis. Plot of the likelihood of rt‐CGM being cost‐effective against hypothetical willingness‐to‐pay thresholds in Aotearoa New Zealand. QALY, quality‐adjusted life year; rt‐CGM, real‐time continuous glucose monitoring.

FIGURE 3.

Incremental cost‐effectiveness ratios from sensitivity analysis scenarios. Plots of ICERs for various scenarios considered in the sensitivity analyses. A conservative WTP‐threshold of NZD 30 000 was assumed as described in the Materials and Methods section. FoH, fear of hypoglycaemia; HbA1c, glycated haemoglobin; ICER, incremental cost‐effectiveness ratio; NZD, New Zealand dollars; QALY, quality‐adjusted life year; rt‐CGM, real‐time continuous glucose monitoring; WTP, willingness‐to‐pay.

Changing the ethnic distribution considered in the analysis from mixed to 100% European resulted in an increase in the ICER to NZD 19 841 per QALY, driven largely by reductions in costs for both interventions that were larger for SMBG than rt‐CGM. On the other hand, when a 100% Māori or a 50% Māori/50% Pacific Peoples cohort was considered, rt‐CGM became cost‐saving, incurring lower direct costs and yielding more QALYs. When considering the age of patients at the initiation of rt‐CGM, decreasing it to 55 years substantially reduced the difference in costs between the two interventions to NZD 2186, reducing the ICER to NZD 3535. If the age at initiation was further reduced to 45 years or 35 years, the total costs associated with rt‐CGM were found to be lower than those of SMBG (NZD −6104, and NZD −18 221, respectively), causing rt‐CGM to dominate.

Additional factors investigated included the discount rate, time horizon, and risk equations. Reducing the discount rate to 1.5% or 0% reduced the ICER to NZD 8745 and NZD 6804, respectively. Reducing the time horizon to 20, 10, or 5 years increased the ICERs to NZD 15 096, 28 321, and 44 091, respectively. The use of the Fremantle Cardiovascular Risk Equation and the Combined Western Australian Mortality Risk equation yielded slightly lower QALYs gained and higher cost‐savings, resulting in a dominant ICER.

3.2. Projected clinical outcomes

Projected complication rates were calculated for the mixed populations, the 100% European, 100% Māori, 100% Pacific Peoples, and 50% Māori/50% Pacific Peoples populations (Table 2). Regardless of ethnic distribution, RRs for ophthalmic, renal, and peripheral complications, in addition to cardiovascular complications including peripheral vascular disease, strokes, and myocardial infarction events were numerically less than one, indicating potential reductions with rt‐CGM. For the mixed population, the largest reductions were for end‐stage renal disease (RR 0.74, 95% confidence interval [CI] 0.68 to 0.80), proliferative diabetic retinopathy (RR 0.78, 95% CI 0.73 to 0.83), macular oedema (RR 0.81, 95% CI 0.77 to 0.86), and gross proteinuria (RR 0.81, 95% CI 0.76 to 0.85). Across populations, higher absolute reductions in proliferative diabetic retinopathy, gross proteinuria, end‐stage renal disease, congestive heart failure, and diabetic mortality were observed in the 100% Māori and 50% Māori/50% Pacific Peoples cohorts.

TABLE 2.

Projected diabetes complication incidence risk reductions when comparing rt‐CGM and SMBG for adult patients with T2D in New Zealand.

Complication	Base case		100% European		100% Māori		100% Pacific peoples		50% Māori and 50% Pacific peoples
	ARR ± SE	RR (95% CI)	ARR ± SE	RR (95% CI)	ARR ± SE	RR (95% CI)	ARR ± SE	RR (95% CI)	ARR ± SE	RR (95% CI)
Ophthalmic complications
Background diabetic retinopathy	−5.25 ± 0.72	0.82 (0.78–0.87)	−5.27 ± 0.73	0.82 (0.78–0.87)	−4.82 ± 0.65	0.83 (0.79–0.87)	−5.31 ± 0.74	0.82 (0.78–0.87)	−4.94 ± 0.69	0.83 (0.79–0.87)
Proliferative diabetic retinopathy	−3.61 ± 0.48	0.78 (0.73–0.83)	−3.34 ± 0.45	0.76 (0.71–0.82)	−5.06 ± 0.62	0.82 (0.78–0.86)	−3.48 ± 0.46	0.76 (0.71–0.81)	−4.29 ± 0.54	0.80 (0.76–0.84)
Macular oedema	−4.41 ± 0.61	0.81 (0.77–0.86)	−4.57 ± 0.62	0.81 (0.76–0.86)	−3.88 ± 0.54	0.82 (0.78–0.87)	−4.60 ± 0.63	0.81 (0.77–0.86)	−4.25 ± 0.59	0.82 (0.77–0.86)
Severe vision loss	−2.98 ± 0.55	0.86 (0.81–0.90)	−2.95 ± 0.53	0.85 (0.80–0.90)	−3.57 ± 0.62	0.87 (0.83–0.91)	−3.04 ± 0.54	0.84 (0.80–0.89)	−3.25 ± 0.58	0.86 (0.82–0.91)
Cataract	−1.03 ± 0.24	0.90 (0.85–0.94)	−1.14 ± 0.24	0.89 (0.85–0.93)	−0.82 ± 0.21	0.91 (0.87–0.96)	−1.17 ± 0.25	0.89 (0.84–0.93)	−1.03 ± 0.23	0.90 (0.85–0.94)
Renal complications
Microalbuminuria	−6.10 ± 0.82	0.82 (0.78–0.87)	−6.12 ± 0.80	0.81 (0.77–0.86)	−6.99 ± 0.91	0.83 (0.79–0.87)	−6.16 ± 0.81	0.81 (0.77–0.86)	−6.40 ± 0.86	0.83 (0.79–0.87)
Gross proteinuria	−5.48 ± 0.73	0.81 (0.76–0.85)	−5.15 ± 0.69	0.78 (0.73–0.84)	−7.23 ± 0.95	0.86 (0.82–0.89)	−5.23 ± 0.70	0.78 (0.73–0.83)	−6.22 ± 0.81	0.83 (0.79–0.87)
End‐stage renal disease	−3.64 ± 0.51	0.74 (0.68–0.80)	−2.97 ± 0.43	0.70 (0.63–0.77)	−7.20 ± 0.96	0.79 (0.74–0.84)	−2.97 ± 0.43	0.70 (0.63–0.78)	−5.07 ± 0.68	0.77 (0.72–0.83)
Peripheral complications
Foot Ulcer	−0.70 ± 0.52	0.91 (0.80–1.04)	−0.74 ± 0.53	0.91 (0.79–1.04)	−0.54 ± 0.51	0.93 (0.81–1.06)	−0.69 ± 0.53	0.91 (0.80–1.05)	−0.66 ± 0.52	0.92 (0.80–1.05)
Recurring foot ulcer	−0.39 ± 0.36	0.98 (0.94–1.02)	−0.38 ± 0.36	0.98 (0.94–1.02)	−0.12 ± 0.34	0.99 (0.95–1.03)	−0.42 ± 0.37	0.98 (0.94–1.02)	−0.35 ± 0.35	0.98 (0.94–1.02)
Amputation from foot ulcer	−0.09 ± 0.29	0.98 (0.88–1.10)	−0.16 ± 0.29	0.97 (0.87–1.08)	−0.02 ± 0.27	1.00 (0.89–1.11)	−0.18 ± 0.29	0.97 (0.87–1.08)	−0.09 ± 0.28	0.98 (0.88–1.09)
Amputation from recurring foot ulcer	−0.12 ± 0.15	0.97 (0.88–1.06)	−0.15 ± 0.15	0.96 (0.87–1.05)	−0.09 ± 0.14	0.97 (0.89–1.06)	−0.18 ± 0.16	0.95 (0.87–1.04)	−0.13 ± 0.15	0.96 (0.88–1.05)
Neuropathy	−6.31 ± 0.91	0.89 (0.86–0.92)	−6.51 ± 0.93	0.89 (0.86–0.92)	−5.45 ± 0.80	0.87 (0.84–0.91)	−6.55 ± 0.94	0.89 (0.86–0.92)	−6.19 ± 0.86	0.88 (0.85–0.91)
Cardiovascular complications
Congestive heart failure event	0.27 ± 0.86	1.01 (0.93–1.10)	0.29 ± 0.86	1.01 (0.93–1.10)	0.86 ± 0.85	1.05 (0.96–1.14)	0.33 ± 0.86	1.02 (0.93–1.11)	0.58 ± 0.86	1.03 (0.95–1.12)
Peripheral vascular disease onset	−1.43 ± 0.29	0.90 (0.86–0.94)	−1.51 ± 0.29	0.89 (0.85–0.93)	−1.0 ± 0.24	0.92 (0.88–0.95)	−1.62 ± 0.30	0.89 (0.85–0.93)	−1.26 ± 0.27	0.91 (0.87–0.94)
Angina	0.44 ± 0.51	1.03 (0.96–1.10)	0.43 ± 0.52	1.03 (0.96–1.10)	0.74 ± 0.49	1.05 (0.99–1.12)	0.39 ± 0.53	1.02 (0.96–1.09)	0.47 ± 0.51	1.03 (0.97–1.10)
Diabetes mortality	0.21 ± 0.52	1.01 (0.97–1.04)	0.31 ± 0.50	1.01 (0.98–1.05)	−0.42 ± 0.54	0.99 (0.96–1.02)	0.18 ± 0.54	1.01 (0.97–1.04)	0.02 ± 0.54	1.00 (0.97–1.03)
Stroke event	−0.50 ± 0.54	0.96 (0.89–1.04)	−0.52 ± 0.55	0.96 (0.89–1.04)	−0.17 ± 0.54	0.99 (0.91–1.07)	−0.65 ± 0.55	0.95 (0.88–1.03)	−0.38 ± 0.54	0.97 (0.90–1.05)
Myocardial Infarction event	−2.14 ± 0.68	0.96 (0.93–0.98)	−2.00 ± 0.68	0.96 (0.93–0.99)	−2.23 ± 0.66	0.95 (0.93–0.98)	−2.00 ± 0.70	0.96 (0.93–0.99)	−2.22 ± 0.68	0.95 (0.93–0.98)

Abbreviations: ARR, absolute risk reduction; CI, confidence interval; HbA1c, glycated haemoglobin; RR, relative risk; SE, standard error; T2D, type 2 diabetes.

4. DISCUSSION

The base case analysis demonstrated a gain of 0.488 QALYs and a life expectancy gain of 0.240 years when using rt‐CGM rather than SMBG, in addition to an ICER of NZD 11 533 per QALY. Although Aotearoa New Zealand does not have a formal willingness‐to‐pay threshold, back calculations of incremental spend each year have been used previously to derive estimates, with a 2022 analysis suggesting a conservative threshold of NZD 30 000. ⁵⁵ When considering this threshold, rt‐CGM is cost‐effective in not only the base case, but also all the scenarios considered in the one‐way sensitivity analysis, other than reducing the time horizon to 5 years. This result was therefore robust regardless of changes in HbA1c effect size, ethnic distribution, age at rt‐CGM initiation, and discount rate. Similarly, the ICERs reported here are far below the GDP per capita in New Zealand, which was estimated to be NZD 75 671 in 2023. ⁵⁹

Despite the modest incremental life expectancy, it should be noted that rt‐CGM offers benefits that go beyond life extension; rt‐CGM use results in improvements in quality of life, reduces the incidence of micro‐ and macrovascular complications, and severe hypo‐ and hyperglycaemic events, as suggested by the present analysis and supported by previous publications.^{16, 17, 21, 23, 24} The results of scenario analyses suggest that ICERs are affected by ethnicity: when populations consisting of either 100% Māori or 50% Māori/50% Pacific Peoples were considered, rt‐CGM became a cost‐saving intervention. Importantly, rt‐CGM was found to be associated with reductions in costs associated with renal complications, ulcer/amputation/neuropathy complications, ophthalmic complications, severe hypoglycaemia, hyperglycaemia, and DKA. Consistent with this, reductions in risks were predicted for most ophthalmic and renal complications examined, in addition to a reduction in the risk of neuropathy. Populations consisting of either 100% Māori or 50% Māori/50% Pacific Peoples experienced even larger absolute reductions in the risks of proliferative diabetic retinopathy, gross proteinuria, end‐stage renal disease, congestive heart failure, and diabetic mortality. These scenarios were conducted in a one‐way sensitivity analysis context, and other potentially influencing factors were not accounted for, nor were clinical inputs derived from studies that explicitly included these groups. So further investigation is warranted to confirm that rt‐CGM is a potentially cost‐saving intervention to reduce disparities in insulin‐treated T2D. However, in support, rt‐CGM has been shown to be highly effective in predominantly Māori populations with T2D.^{60, 61}

The effectiveness of continuous glucose monitoring for patients with T2D has been supported by numerous studies, with a 2024 systematic review comparing it with SMBG finding HbA1c improvements regardless of whether patients were receiving insulin or oral agents only, and with rt‐CGM trending towards having larger benefits than intermittently‐scanned monitoring. ⁶² The glycaemic control benefit modelled in the present analysis was based on an RWE study including 36 080 insulin‐treated T2D patients. ¹¹ That study found improvements in glycaemic control with rt‐CGM as measured by drops in HbA1c, hypoglycaemia rates, increases in the proportion of patients with HbA1c <53 mmol/mol, and decreases in the proportion of patients with HbA1c >75 mmol/mol. Furthermore, rt‐CGM was associated with decreased resource use, including reductions in emergency department visits and hospitalizations for hypoglycaemia, in addition to overall reductions in the numbers of outpatient and telephone visits. Importantly, rt‐CGM has been shown to have glycaemic benefits in an Aotearoa New Zealand population. ⁶³ In a multicentre RCT including a mostly (54%) Māori population of patients with T2D requiring insulin, rt‐CGM was associated with improvements in time‐in‐range and HbA1c. ⁶³

A key benefit of the modelling approach used here was the inclusion of a mixed population in addition to scenarios considering Māori and Pacific Peoples populations. In Aotearoa New Zealand, Māori and Pacific people have a higher risk of T2D, earlier ages of onset, and increased risks of complications when compared with New Zealanders of European descent.^{37, 64} The findings in this analysis suggesting that rt‐CGM is cost‐saving in these groups and that they may experience reductions in the rates of complications are thus of particular importance. However, because rt‐CGM was found to be cost‐effective across Aotearoa New Zealand ethnic groups, including those of European descent, a population approach to public funding for rt‐CGM is warranted to improve overall T2D management in Aotearoa New Zealand.

Key limitations of this study included a focus on direct costs, a lack of parameters and risk equations specific to the Aotearoa New Zealand population, and not considering rt‐CGM in the context of more complex interventions that may impact the effectiveness of rt‐CGM such as lifestyle interventions or education programmes. The exclusion of indirect costs in particular means that this analysis is likely quite conservative. A 2021 report estimated that of the NZD 2118 million total annual cost of T2D in New Zealand, lost personal income, tax revenue, and non‐salary labour accounted for NZD 1060 million. ³⁶ Similarly, the combination of rt‐CGM with additional interventions may be beneficial in diabetes management, meaning the exclusion of such approaches may prevent the model from fully capturing the benefit of real‐time monitoring programmes. A 2024 systematic review of interventions including education, nurse‐led programmes, exercise, diet, and other lifestyle interventions in Aotearoa New Zealand found that they are associated with modest improvements in HbA1c and weight at 6 months. ⁶⁵ Conversely, the use of risk equations that were not specifically developed for Aotearoa New Zealanders may lead to an under‐ or overestimation of the risks in this population. ⁶⁶ This is a result of widespread screening for T2D that has resulted in reductions in CV risks compared with European populations or even New Zealanders prior to the introduction of such screening. ⁶⁶ Unfortunately, the use of the PREDICT model, developed to account for this changed risk profile, ⁶⁶ was not possible as this equation is not included in the CDM. Similarly, the specificity of the analysis carried out here for the New Zealand population would benefit from the use of data collected from a New Zealand population, which is unfortunately currently lacking. While one RCT examining rt‐CGM has been reported from New Zealand, it had a relatively short follow‐up duration of 12 weeks and was focused on a predominantly Māori population. ⁶³ However, the patient characteristics reported by that study suggest that the baseline risk profile in New Zealand is less favourable than that considered in this analysis.

While further work to elucidate the potential benefits of rt‐CGM in patients with T2D in Aotearoa New Zealand would be beneficial, the analysis presented here provides insight into the cost‐effectiveness of such approaches. The reductions in the incidences and costs of a variety of complications, in addition to QALY and life expectancy gains, suggest that rt‐CGM has the potential to improve the lives of New Zealanders with T2D. Furthermore, these benefits can be realized with a relatively small ICER compared with SMBG that is far below even a conservative WTP threshold, thus representing a high‐value investment for the healthcare system.

CONFLICT OF INTEREST STATEMENT

Jessica Y. Matuoka and Gregory J. Norman are employees of Dexcom. Jessica Y. Matuoka and Gregory J. Norman hold stock or stock options in Dexcom. Hamza A. Alshannaq is a former employee of Dexcom. Ryan G. Paul has received speaking honoraria from Dexcom and Abbott and has received research grants from Dexcom. Richard F. Pollock is a shareholder, director, and full‐time employee of Covalence Research Ltd., which received consultancy fees from Dexcom to prepare the manuscript.

Supporting information

Table S1. Baseline patient characteristics for type 2 cohort population.

Table S2. Cost per diabetes complication or event, adjusted to NZD 2024.

Table S3. Baseline utility and diabetes complication disutility values used in the comparison between rt‐CGM and SMBG.

Table S4. Results of sensitivity analyses comparing rt‐CGM and SMBG.

DATA AVAILABILITY STATEMENT

Data supporting the findings of this study are available from the corresponding author upon reasonable request.