Dasiglucagon demonstrates reduced costs in the treatment of severe hypoglycemia in a budget impact model

Abstract

BACKGROUND:

Approximately 7.3 million people with type 1 or type 2 diabetes (T1D/T2D) are treated with insulin, placing them at higher risk of severe hypoglycemia (SH). SH requires assistance of another individual and often necessitates the prompt administration of intravenous glucose, injectable glucagon, or both. Untreated, SH can progress to unconsciousness, seizures, coma, or death. Before 2018, all glucagon rescue treatments required reconstitution. The complexity of reconstitution is often a barrier to successful administration during a severe hypoglycemic event. Studies suggest successful administration of glucagon emergency kits range from 6%-56% of the time. Second-generation glucagon treatments and glucagon analogs do not require reconstitution and have caregiver administration success rates ranging from 94%-100%. Dasiglucagon is a glucagon analog administered via autoinjector or prefilled syringe and has been shown to result in rapid hypoglycemia recovery. Moreover, the autoinjector can be administered successfully 94% of the time by trained caregivers. Previous evaluation of costs in budget impact models (BIMs) demonstrated the potential for second-generation glucagon treatments to reduce the cost of SH events (SHEs). The current model expands on those findings with a treatment pathway and accompanying assumptions reflecting important aspects of real-world SHE treatment.

OBJECTIVE:

To evaluate the economic impact of dasiglucagon compared with available glucagon treatments for SHE management, considering direct cost of treatment and health care resource utilization.

METHODS:

A 1-year BIM with a hypothetical US commercial health plan of 1 million lives was developed with a target population of individuals with diabetes at risk of SHE. The treatment pathway model included initial and secondary treatment attempts, treatment administration success and failure, plasma glucose (PG) recovery within 15 minutes, emergency medical services, emergency department (ED) visits, and hospitalizations. A 1-way sensitivity analysis was conducted to assess the sensitivity of the model to changes in parameter values.

RESULTS:

In a 1 million-covered lives population, it was estimated that 12,006 SHEs would occur annually. The higher rate of initial treatment success and PG recovery within 15 minutes associated with dasiglucagon treatment resulted in lower total health care costs. Total SHE treatment costs with dasiglucagon were estimated at $13.4 million, compared with $16.7 million for injectable native glucagon, $20.7 million for nasal glucagon, $35.3 million for reconstituted glucagon, and $43.8 million for untreated individuals. Compared with untreated people, the number needed to treat (NNT) with dasiglucagon was 6 individuals to avoid 1 hospitalization. NNT for this same comparison was 59 for injectable native glucagon and 27 for nasal glucagon.

CONCLUSIONS:

Treatment of SH with dasiglucagon decreased total direct medical costs by reducing health care resource utilization (emergency calls, emergency transports, ED visits, and hospitalizations) and accompanying costs associated with the treatment of SH.

What is already known about this subject

• Approximately 7 million people with diabetes in the United States are treated with insulin and as such are at risk of severe hypoglycemia (SH).

• First-generation glucagon rescue treatments are successfully administered approximately 6%-56% of the time, and second-generation treatments of SH, including stable glucagon preparations and glucagon analog, can be successfully delivered 94%-100% of the time.

• Previous budget impact analyses with second-generation treatments do not consider rates of plasma glucose recovery or probability of attempting a second treatment attempt in the absence of symptom resolution.

What this study adds

• The current budget impact model estimates that higher rates of plasma glucose recovery, defined as the number of patients recovering within 15 minutes when using dasiglucagon, compared with other first- and second-generation glucagon rescue treatments, to treat SH events can result in significantly lower total annual treatment costs ($3.3 and $7.3 million in savings per 1 million people per year compared with treatment with native injectable glucagon or to nasal glucagon, respectively).

• Treatment with dasiglucagon reduced the cost of emergency department visits by $13.5 million and hospitalizations by $18.8 million when compared with no rescue treatment; $8.8 and $12.2 million, respectively, when compared with native glucagon that required reconstitution; $2.9 and $4.0 million, respectively, when compared with nasal glucagon; and $1.3 and $1.9 million, respectively, when compared with injectable glucagon.

• The number needed to treat, compared with untreated people, demonstrated that only 6 individuals would need to be treated with dasiglucagon to avoid 1 hospitalization.

In the United States, approximately 7.3 million individuals are treated with insulin (1.6 million with type 1 diabetes [T1D] and 5.7 million with type 2 diabetes [T2D]).^1-4 However, insulin therapy significantly increases the risk of hypoglycemia.^4-9 Severe hypoglycemia (SH) is characterized by low blood glucose concentrations (< 54 mg/dL) and altered mental and/or physical functioning that requires assistance from another person for recovery.¹⁰ If untreated, SH can result in serious adverse health outcomes that include unconsciousness, seizures, coma, or death.^9,11-19 SH can also lead to significant direct medical costs and indirect productivity costs and has been shown to result in an impaired quality of life; however, the occurrence and coding of SH events (SHEs) are also underreported, thereby underestimating the actual economic impact.^14,20,21

Total US health care expenditures attributed to diabetes in 2017 was $237 billion, of which $69.7 billion resulted from hospital inpatient care, and $8.3 billion resulted from emergency and ambulance services, which may be attributable to SH.²¹ Beyond direct medical costs, the fear of hypoglycemia among people with diabetes contributes to insulin avoidance by delaying initiation of insulin therapy or by avoiding adherence to insulin regimens and glycemic targets, further affecting quality of life and diabetes disease management efforts.²² The average individual at risk may experience multiple SHEs in their lifetime, since these events occur at an estimated incidence of 0.39 events per person-year in people with T1D and 0.80 events per person-year in people with T2D treated with insulin.²³ Recurrent SHEs lead to an even greater risk of future events as a result of the reduced ability to detect the onset of symptoms.^7,23-27

The American Diabetes Association (ADA) recommends that all individuals who are at risk of Level 2 (defined as < 54 mg/dL or 3.0 mmol/L) and Level 3 hypoglycemia be prescribed glucagon so that it is readily available should a severe event occur.⁵ Among individuals with initial insulin prescriptions and at least 1 year of follow-up data, only 49.3% of individuals with T1D and 2.4% with T2D have been shown to fill glucagon prescriptions, and less than 10% of individuals with T1D or T2D fill glucagon prescriptions following an emergency department (ED) visit for hypoglycemia.^28,29

Currently available first-generation glucagon kits include the Glucagon Emergency Kit (Eli Lilly and Company) and GlucaGen Hypo-Kit (Novo Nordisk); these kits require reconstitution before emergency administration. The complexity of the reconstitution process with the first-generation glucagon treatments can be a barrier to successful preparation and administration during an SHE.^30-32 Multiple studies have reported that reconstituted glucagon treatments have low rates of successful administration that range from 6% to 56%.^30-32 These treatment failures can lead to increased risk of SH-related adverse outcomes, such as loss of consciousness, seizure, coma and death.⁵

Second-generation glucagon treatments, defined as treatments that do not require reconstitution, eliminate the complex preparation process necessary with the first-generation glucagon kits. Introduced to the US market in 2019, these second-generation treatments include Gvoke (Xeris Pharmaceuticals), a native glucagon in an autoinjector or prefilled syringe approved for use by patients with diabetes aged 2 years and above, and nasal glucagon powder Baqsimi (Eli Lilly and Company) for patients with diabetes aged 4 years and above. More recently, the first and only glucagon analog, dasiglucagon (Zegalogue) was approved by the US Food and Drug Administration (FDA). Dasiglucagon has improved aqueous solubility and physical stability compared with native human glucagon for injection.³³ Dasiglucagon is available in an autoinjector or prefilled syringe and is indicated for the treatment of SH in pediatric and adult patients with diabetes aged 6 years and above.³³ In each of 3 pivotal phase 3 trials, dasiglucagon demonstrated a rapid and reliable median time to plasma glucose (PG) recovery of 10 minutes in both adults and pediatric subjects.^33-36 Within the first 15 minutes after taking a dose, up to 99% of adult and 95% of pediatric subjects experienced PG recovery (defined as the first increase in PG of ≥ 20 mg/dL from the time of administration without rescue intravenous glucose).^33-36 Adverse events occurring in 2% or more of study subjects included nausea, vomiting, headache, diarrhea (adults only), and injection site pain.^33-36 Importantly, when dasiglucagon was compared with firstgeneration glucagon emergency kits, dasiglucagon and the other second-generation glucagon treatments demonstrated higher rates of successful administration in user simulation studies, ranging from 94%-100%.^30,31,37

The impact of the treatment-related health care costs for severe hypoglycemia when using different glucagon treatments has been explored in previous budget impact models (BIMs).^18,19 However, a number of critical aspects of the care pathway were not included in these models.^18,19 In the BIM reported by Pohlmann et al, which compared nasal glucagon with reconstituted glucagon, the framework did not consider rates of PG recovery or the probability of attempting a second treatment in the absence of symptom resolution.¹⁸ Furthermore, this model did not include the probability of subsequent hospitalizations following transport to the ED.¹⁸ A second model published by Leinwand et al compared injectable native glucagon with reconstituted glucagon in a standard glucagon emergency kit.¹⁹ Importantly, this model did not consider rate of PG recovery, the probability of a second treatment attempt, or the probability of transportation to the ED by a private vehicle.

The current model assessed the budget impact of SH treatment while accounting for each of the factors excluded from previous published BIMs, including probability for successful treatment administration and the probability of PG recovery. The BIM reported in this analysis was developed to examine the difference in costs of care associated with treatment with dasiglucagon compared with first- and second-generation glucagon treatments, as well as with no treatment.

Methods

ANALYTICAL FRAMEWORK AND PERSPECTIVE

A 1-year BIM for a hypothetical US commercial health plan population of 1 million lives was developed to evaluate costs associated with treating SH using dasiglucagon compared with other first- and second-generation treatments. Although prescription of glucagon is recommended for all people with diabetes, real-world data suggest that a large proportion of people with diabetes do not have a prescription for glucagon or are not filling their glucagon prescriptions.^28,29 Therefore, “no glucagon treatment” was included as a likely scenario. The target population was individuals with diabetes at risk for severe hypoglycemic events, including all people with T1D and T2D using insulin. A thorough review of published literature, available Medicare claims data, and treatment guidelines was conducted to develop a clinical decision tree to simulate possible patient pathways from immediate treatment-related care, to emergency response, and to hospital care (Figure 1).^5,18,19

FIGURE 1.

Severe Hypoglycemic Event Clinical Decision Pathway

PATIENT PATHWAY

Figure 1 displays the possible pathways of care modeled and considers successful and unsuccessful glucagon administration, probability of PG recovery within 15 minutes, emergency medical services (EMS) or private transportation to the ED, ED visit, and hospitalization and care. This patient pathway gives rise to multiple model parameters and encompasses clinical and health care resource use, as well as cost inputs. Each decision node was associated with a probability of occurrence, and each aspect of care was associated with a cost. Costs and probabilities were based on claims data from the Centers for Medicare & Medicaid Services, review of peer-reviewed literature, and assumptions based on endocrinology and diabetes expert clinical opinion.

The model assumed that, of the 1 million covered lives modeled, only those with T1D (4,835 lives) or people with T2D treated with insulin (17,374 lives) were at risk of SH.^1,4 All at-risk individuals were assumed to have had their glucagon prescriptions filled at the time of the SHE. Following an SHE, glucagon was either successfully or unsuccessfully administered (Figure 1); the likelihood of successful treatment administration was based on reported rates of caregiver full-dose administration success rates, since caregiver administration success rates were consistently reported across studies.^30,31,37 If glucagon was successfully administered by a caregiver, people could experience PG recovery within 15 minutes or not based on published data for each therapy.^33,36,38,39 In keeping within the ADA guidelines and FDA-approved labeling for glucagon rescue treatments, the model assumed that a recommended second dose of glucagon was given after 15 minutes if there was no resolution of symptoms.^33,40-43 Further actions following initial administration failure and initial successful administration included guideline-recommended contact with EMS, private transportation to the ED, or, for those with PG recovery, no further intervention. Upon arrival of EMS, people may be either cared for and released on scene or transported to the ED. If transported to the hospital, the individual was assumed to be admitted to the ED and subsequently admitted to the hospital or discharged from the ER. If admitted to the hospital, the patient was assumed to be either discharged alive or to have died before discharge.

KEY MODEL PARAMETERS

The prevalence of people with T1D and people with T2D treated with insulin was estimated from 2020 US Census data and Centers for Disease Control and Prevention reports shown in Table 1.⁴⁴ Published literature sources were used to estimate the incidence of SHEs in people with T1D and T2D.^1,4,23-26 Two different event rates for people with T1D and 4 incidence rates for people with T2D were identified, as shown in Table 1. An average of the values was used for the model’s incidence estimation. The average values calculated in the shaded rows of Table 1 were used as base-case estimates.

TABLE 1.

Epidemiology Base-Case Analysis

Parameter	Input	Source
Prevalence of T1D	4.83	Per 1,000a
US adults with T1D	1,400,000	CDC1
US children with T1D	187,000	CDC1
Prevalence of T2D on insulin	17.37	Per 1,000a
US adults and children with T2D	25,313,000	CDC1,b
Percent of people with T2D on insulin	22.53%	Selvin et al4,c
Incidence of SHE in people with T1D	0.387	Per person-year, average
In people on 3+ daily injections or pump therapy	0.408 per person-year	Gubitosi-Klug et al23
In people on 1-2 daily injections	0.366 per person-year	Gubitosi-Klug et al23
Incidence of SHE in people with T2D treated with insulin	0.583	Per person-year, weighted averaged
Median reported among people on basal regimens	0.04 per person-year	Seufert et al24
In people using MDI	0.23 per person-year	Herman et al25
In people using CSII	0.08 per person-year	Herman et al25
In all people with T2D on any insulin	1.05 per person-year	Edridge et al26

^a Assuming US population of 328,239,523 (US Census 2019). ⁴⁴

^bTotal reported with diabetes less total reported with T1D, as above.

^c Selvin et al report that 27.1% of people with diabetes are insulin treated; assuming all patients with T1D are insulin treated, the remaining use of insulin is attributed to the use of insulin in patients with T2D.4

^d Final row is weighted at 50% of total; all others are equally weighted.

CDC = Centers for Disease Control and Prevention; CSII = continuous subcutaneous insulin infusion; MDI = multiple daily injection; SHE = severe hypoglycemic event; T1D = type 1 diabetes mellitus; T2D = type 2 diabetes mellitus.

PATIENT PATHWAY PROBABILITY AND COST INPUTS

Effective use of glucagon to treat SH can be broken down into 2 steps: successful administration and achieving PG recovery. Successful administration rates for dasiglucagon, injectable native glucagon, nasal native glucagon, and reconstituted native glucagon were identified from previous publications (Table 2).^30,31,37 PG recovery rates within 15 minutes of all glucagon rescue treatments were identified from published trial data (Table 2).^34,38,39

TABLE 2.

Probability and Cost Inputs in Base-Case Analysis ($USD)

Parameter	Input	Source
Probability of successful full-dose administration, %
Dasiglucagon	94.00	Bailey et al30
Injectable native glucagon (second-generation)	100.00	Newswanger et al37
Nasal glucagon (second-generation)	93.75	Yale et al31
Reconstituted glucagon (first-generation)	34.03	Yale et al31 Bailey et al30
No glucagon treatment	0.00	Assumption based on expert clinical opinion
Probability of PG recovery in 15 minutes, %
Dasiglucagon	99.00	Pieber et al34
Injectable native glucagon (second-generation)	78.90	Gvoke FDA submission38
Nasal glucagon (second-generation)	71.00	Rickels et al39
Reconstituted glucagon (first-generation)	95.00	Pieber et al34
No glucagon treatment	2.00	Pieber et al34
Probability of taking further action, %
Following initial administration failure	100.00	Assumption based on expert clinical opinion
Following successful initial administration and PG recovery	33.33	Assumption based on expert clinical opinion
Following successful initial administration but no PG recovery	75.00	Assumption based on expert clinical opinion
Probability of second treatment given lack of PG recovery	25.00	Assumption based on expert clinical opinion
When further action is taken, %
Probability of EMS call	57.46	Moffet et al45 Ginde et al46,a
Probability of driving patient to ED	42.54	Moffet et al45 Ginde et al46
EMS actions taken, %
Probability of release on scene following initial administration failure or lack of PG recovery	13.50	Moffet et al45
Probability of release on scene following initial administration success and PG recovery	100.00	Assumption based on expert clinical opinion
Probability of hospitalization after ED visit	25.00	Median reported46-48
Probability of death, %
Probability of death in the ED	0.40	Leinwand et al19
Probability of death during a hospitalization	1.38	Medicare claims49
Glucagon treatment, $
Dasiglucagon	309.00	Price Rx50
Injectable native glucagon (second-generation)	280.80	Price Rx50
Nasal glucagon (second-generation)	280.80	Price Rx50
Reconstituted glucagon (first-generation)	293.05	Price Rx50
No glucagon treatment	0.00	Assumption based on expert clinical opinion
EMS, without transport, $
Base HCPCS code, A0429	386.04	Based on Connecticut locality payment rates52
EMS, with transport, $
Base HCPCS code, A0429	419.76	Based on Connecticut locality payment rates52
ED visit for SHE, $	1,546.00	Median18,19,53-56
Hospital inpatient care for SHE, $	8,679.00	Medicare claims49

^a Moffet et al state that for a catchment area of 1.6 million, 8,332 encounters for hypoglycemia occurred over 3 years, resulting in an EMS call rate of 0.00174 per person-year for SHEs, of which 86.5% result in transport.45 Ginde et al report an ED visit rate of 0.034 per person-year for severe hypoglycemia.46 We assume that all SHE people arriving in the ED not accounted for by EMS transportation are transported by private vehicle.

ED = emergency department; EMS = emergency medical services; FDA = US Food and Drug Administration; HCPCS = Healthcare Common Procedure Coding System; PG = plasma glucose; SHE = severe hypoglycemic event.

Although all approved glucagon rescue treatments instruct users to call EMS after each administration, in practice not all SHEs result in EMS calls or transport. To illustrate real-world use of EMS services, inputs for EMS actions were calculated from values published in a study observing EMS services for SH in Alameda County, California, from 2013 to 2015 (Table 2).^45,46 Findings were used to estimate the number of EMS calls for SHEs in the United States, as well as those resulting in an ED transport. The estimate of the likelihood that people are privately transported to the ER was based on data from 2 separate sources (which refined assumptions made in earlier BIMs).¹⁸ These were converted to probabilities for the treatment failure and no treatment nodes of the care pathway. The 3 possible outcomes upon admission to the ED were identified from published literature (Table 2).^19,46-48 The probability of death after hospitalization was taken from analysis of 2018-2019 Medicare claims data (Table 2).⁴⁹ Detailed clinical parameter inputs are included in Supplementary Table 1 ^{(563.5KB, pdf)} (available in online article).

The total costs of treating 1 SHE was compared across all glucagon rescue treatments. Differences in time to treatment success across all treatments were analyzed to evaluate the impact on cost outcomes. This BIM assessed the population-level economic impact of introducing dasiglucagon by combining patient-specific severe hypoglycemic incidence rates, health care utilization, and cost estimates. Costs in the model associated with each aspect of treatment are shown in Table 2. Wholesale acquisition costs (WAC) for 2021 were used for cost inputs for initial and second treatment attempts at SH onset.^30,50 Per-mile rates for the median Medicare payment rate (in terms of locality) were used.^51,52 ED costs were estimated by reviewing 6 publications that reported ED costs for SH and adjusted for inflation (Table 2).^18,19,53-56 Inpatient care costs were taken from the average of estimated payments for 2020 claims (through June 2020) for SH-related inpatient stays.⁴⁹ Detailed cost parameter inputs are included in Supplementary Table 1 ^{(563.5KB, pdf)} .

SENSITIVITY ANALYSIS

To assess the sensitivity of the model to changes in parameter values, a 1-way sensitivity analysis was conducted. For purposes of this assessment, all parameter values were varied up to 15% higher and lower than the base value. A sensitivity analysis was also conducted based on PG recovery at 30 minutes.^34,42,43

Results

BASE CASE

A total of 12,006 annual SHEs were estimated in a population of 1 million US commercial covered lives. Transition probabilities for people experiencing SHEs were assumed to be the same, regardless of diabetes type, since diabetes type did not affect treatment success probabilities and mortality. Cost-offset results depended on treatment outcome probabilities and costs.

TOTAL COSTS

Total modeled SHE-related annual medical costs in a US commercial payer population were significantly lower with dasiglucagon compared with the other glucagon rescue treatments and with no treatment. Estimated total SHE-related annual medical costs were $13.4 million for dasiglucagon, $16.7 million for injectable native glucagon, 20.7 million for nasal glucagon, $35.3 million for reconstituted glucagon, and $43.8 million for no glucagon treatment. As such, dasiglucagon treatment resulted in $3.3 million lower total annual health care expenditures than the next lowest cost treatment (injectable native glucagon; Table 3). Overall, dasiglucagon treatment resulted in an estimated 20% reduction in total annual SH treatment costs compared with injectable native glucagon ($3.3 million less expenditure), 35% reduction compared with nasal glucagon ($7.3 million less expenditure), 62% reduction compared with reconstituted glucagon ($22.0 million less expenditure), and 69% reduction compared with no glucagon treatment ($30.4 million less expenditure).

TABLE 3.

Base-Case Annual Costs of SHE in a US Commercial Plan, with 1 Million Covered Lives by Treatment

	Dasiglucagon, $	Injectable native glucagon, $	Nasal glucagon, $	Reconstituted glucagon, $	No glucagon treatment, $
Total annual SHE costs	13,352,605	16,652,070	20,700,641	35,324,301	43,764,451
Initial treatment attempts	3,709,870	3,371,299	3,371,299	3,518,373	0
Second treatment attempts	8,718	177,836	229,143	14,966	0
EMS calls	1,011,013	1,262,434	1,481,496	2,089,347	2,663,364
EMS transport	13,965	42,457	67,282	136,167	201,218
ED visits	3,605,019	4,940,408	6,512,127	12,380,474	17,126,742
Hospitalizations	5,004,021	6,857,635	9,039,293	17,184,973	23,773,128
Cost per member per year for SHEs	13	17	21	35	44
Cost per member per month for SHEs	1.11	1.39	1.73	2.94	3.65
Cost per SHEs	1,114	1,387	1,724	2,939	3,645
Cost per patient with diabetes	601	750	932	1,581	1,971
Total annual pharmacy costs	3,718,588	3,549,135	3,600,442	3,533,340	0
Total annual medical costs	9,634,016	13,102,935	17,100,199	31,790,961	43,764,451

ED = emergency department; EMS = emergency medical services; SHE = severe hypoglycemic event.

The cost per member per year for SHEs was lowest for dasiglucagon at $13.35, as compared with $16.65-$35.32 for other treatment options and $43.76 for no treatment. For a population of 1,000,000 covered lives, treatment with dasiglucagon led to lower costs per SHE treated compared with other glucagon treatments or no treatment evaluated on a per-member per-year basis and per-patient with diabetes basis (Table 3). Although the total pharmacy costs for dasiglucagon ($3.72 million) were higher than injectable native glucagon ($3.55 million), nasal glucagon ($3.60 million), and reconstituted glucagon ($3.53 million), the pharmacy expenses were offset by direct medical savings. The total medical costs for dasiglucagon were estimated to be $9.6 million compared with $13.1 million for injectable native glucagon, $17.1 million for nasal glucagon, $31.8 million for reconstituted glucagon, and $43.8 million for no glucagon treatment (Table 3).

COST WITH INITIAL AND SECOND TREATMENT ATTEMPT

In this model, the higher rate of initial treatment success and PG recovery within 15 minutes with dasiglucagon resulted in a reduction in the number of EMS calls, leading to lower costs for EMS services (Table 3). Treatment with injectable native glucagon and nasal glucagon resulted in higher costs due in part to higher rates of second treatment attempts, hospitalizations, and ED visits when compared with dasiglucagon treatment.

COST OF EMERGENCY RESPONSE AND HOSPITALIZATIONS

Total annual costs for ED visits reflect patients brought to the ED by EMS or by private vehicles. Lower ED volumes resulting from treatment with dasiglucagon resulted in ED costs of approximately $3.6 million compared with ED costs of $4.9 million to $12.4 million annually for other glucagon options (Table 3). Similarly, lower rates of inpatient hospitalizations resulted in total annual hospitalization costs of $5.0 million for people treated with dasiglucagon compared with annual hospitalization costs of $6.9 million to $17.2 million for people treated with other glucagon options (Table 3).

NUMBER NEEDED TO TREAT

Improved PG recovery with dasiglucagon may ultimately lead to fewer hospitalizations (Table 4). When compared with people receiving no glucagon treatment, 6 persons would need to be treated with dasiglucagon to avoid 1 hospitalization. Corresponding number needed to treat was 27 for native nasal glucagon and 59 for injectable native glucagon.

TABLE 4.

Number Needed to Treat with Dasiglucagon, by Comparison Treatment

	Injectable native glucagon	Nasal glucagon	Reconstituted glucagon	No glucagon treatment
Number needed to treat with dasiglucagon to avoid 1 hospitalization	59	27	9	6

SENSITIVITY ANALYSIS RESULTS

The 1-way sensitivity analysis showed that the total annual treatment cost with dasiglucagon did not exceed the total annual treatment cost with nasal glucagon, reconstituted glucagon, or no glucagon treatment, regardless of changes to model parameters up to 15% of their baseline value. The total annual treatment cost with dasiglucagon exceeded the cost of treatment with injectable native glucagon in 3 instances. When the probability of initial successful treatment administration with dasiglucagon was varied to its lowest level, total annual treatment costs with dasiglucagon increased by $5,118,257, resulting in an estimated annual net savings with injectable native glucagon compared with dasiglucagon of $1,818,801. Additional details on sensitivity results are found in Supplementary Figures 1-4 ^{(563.5KB, pdf)} (available in online article).

When modeled using rates of PG recovery at 30 minutes, total costs of treatment were lowest with injectable native glucagon, followed by dasiglucagon, nasal glucagon, reconstituted glucagon, and no glucagon treatment ($10.9 million, $13.1 million, $12.8-13.2 million, $34.8 million, and $43.8 million, respectively).

Discussion

The economic burden of diabetes and the specific costs tied to SH in the US health care system are substantial and well documented in the published literature. Even so, the incidence of SHEs are known to be significantly underreported, which considerably underestimates the actual economic impact of SH.²¹

The impact on health outcomes and health care costs of untreated and undertreated SH have also been widely studied and reported.^9,11-19 With closer adherence to the ADA guidelines, which recommend that all individuals who are at risk for Level 2 or Level 3 hypoglycemia are prescribed glucagon rescue treatment, the economic burden of SH can be reduced significantly; further work and awareness with primary care physicians on the importance of prescribing glucagon rescue treatments and dialogues with people with diabetes and their caregivers on the importance of carrying rescue treatments may be needed. Two budget impact studies have reported that use of second-generation glucagon rescue treatments may positively improve health outcomes and reduce the cost of care.^18,19

However, in contrast to previous models, this analysis highlights the increased cost savings of the total annual SHE treatment costs associated with dasiglucagon. The probability of PG recovery in the first 15 minutes for dasiglucagon was 99%; this high recovery rate, coupled with a 94% probability of successful full-dose administration, resulted in lower rates of EMS transport, ED visits, hospitalizations, and deaths compared with all glucagon rescue treatments. When compared with other second-generation rescue therapies, dasiglucagon resulted in an estimated 20% reduction in total annual costs compared with injectable native glucagon and a 35% reduction compared with nasal glucagon. This BIM, along with previous studies of second-generation rescue therapies, demonstrates that effective initial full-dose administration of glucagon leads to a significant reduction in downstream health care resource utilization, resulting in direct medical cost savings.

The BIM presented in this analysis showed a significantly lower total and per-patient cost of care with dasiglucagon treatment. The model possesses several unique features when compared with other published models. The current analysis is the first to compare all available second-generation glucagon rescue treatments, rather than a comparison only to reconstituted glucagon emergency kits and untreated SH. Our analysis also includes a robust 1-way sensitivity analysis on key parameter values, in which parameter values were varied by 15% from their default value.

Variables previously not analyzed in other published BIMs, including probability of PG recovery following initial administration, probability of attempting a second treatment attempt, and hospitalization, were included in this analysis. Importantly, several more conservative assumptions were used in the development of the current model. The incidence of SHEs may be underestimated, since previous BIMs have used incidence of SHEs of 2.5 events per person-year based on data reported in Canada, whereas in this model, the average incidence of SHEs among at-risk people was considered to be 0.54 events per person-year based on multiple sources in the published literature.^23-26,57

Hospitalization costs used in our model represented average Medicare payment rates. However, it is widely recognized that private payers pay more than Medicare (reportedly an average of 189% of Medicare rates for inpatient services and 264% of Medicare rates for outpatient services), suggesting that our findings represent a conservative estimate of total potential costs/cost savings.⁵⁸ Conversely, WAC was used to estimate the cost of glucagon in this model; actual costs to commercial payers may be lower because of negotiated prices. It was also assumed that 50% of people transported were urban dwellers (living an average of 1.17 miles from nearest hospital), and the remainder were rural dwellers (living an average of 5.07 miles from the nearest eligible hospital); in reality, more people may be urban dwellers.

LIMITATIONS

Limitations of the current model include the assumption that all people with TID and T2D using insulin in the population have access to glucagon. Similar to the other published BIMs, this model does not consider indirect impacts, including caregiver burden, loss of productivity, and long-term disability.²¹ The BIM was developed from a commercial payer perspective but used Medicare reimbursement rates, and the analysis conducted from the perspective of the US commercial health plans may not be applicable to other payer types.

Conclusions

This model showed that the higher rates of successful full-dose administration and PG recovery by 15 minutes associated with dasiglucagon compared with other glucagon rescue treatments led to lower costs per member, per event, and per patient with diabetes. Dasiglucagon is a first-in-class glucagon analog that does not require reconstitution before administration and is available in an autoinjector and prefilled syringe. Across 3 pivotal phase 3 trials (2 adult and 1 pediatric), dasiglucagon showed a consistent and rapid median time to PG recovery of 10 minutes.^33-36 This model demonstrated that dasiglucagon may have a favorable budget impact for US commercial payers because of its high rate of successful administration and subsequent PG recovery rates, resulting in a lower rate of emergency calls, emergency transports, ED visits, and hospitalizations, leading to overall medical cost savings.