Cost-effectiveness of insulin detemir versus insulin glargine for Thai type 2 diabetes from a payer’s perspective

Abstract

Aims: An economic evidence is a vital tool that can inform the decision to use costly insulin analogs. This study aimed to evaluate long-term cost-effectiveness of insulin detemir (IDet) compared with insulin glargine (IGlar) in type 2 diabetes (T2DM) from the Thai payer’s perspective.

Methods: Long-term costs and outcomes were projected using a validated IMS CORE Diabetes Model, version 8.5. Cohort characteristics, baseline risk factors, and costs of diabetes complications were derived from Thai data sources. Relative risk was derived from a systematic review and meta-analysis study. Costs and outcomes were discounted at 3% per annum. Incremental cost-effectiveness ratio (ICER) was presented in 2015 US Dollars (USD). A series of one-way and probabilistic sensitivity analyses were performed.

Results: IDet yielded slightly greater quality-adjusted life years (QALYs) (8.921 vs 8.908), but incurred higher costs than IGlar (90,417.63 USD vs 66,674.03 USD), resulting in an ICER of ∼1.7 million USD per QALY. The findings were very sensitive to the cost of IDet. With a 34% reduction in the IDet cost, treatment with IDet would become cost-effective according to the Thai threshold of 4,434.59 USD per QALY.

Conclusions: Treatment with IDet in patients with T2DM who had uncontrolled blood glucose with oral anti-diabetic agents was not a cost-effective strategy compared with IGlar treatment in the Thai context. These findings could be generalized to other countries with a similar socioeconomics level and healthcare systems.

Keywords:

• Insulin glargine
• Insulin detemir
• Cost
• Cost-effective
• Thailand

Introduction

The advent of the aging society, increasingly unhealthy lifestyles with poor diets, and growing prevalence of excess body weight and obesity potentially influence the growth in the number of people with diabetes worldwide. This phenomenon has been observed in both developed and developing countries^1,2, including Thailand^3–5. According to the 4th Thai National Health Examination Survey (NHES IV), 7.5% of Thai adults aged 20 and older had diabetes; one-third of these groups were undiagnosed⁶. In addition, data from diabetes clinics in 11 tertiary hospitals from different parts of Thailand reported that 94.6% were type 2 diabetes (T2DM), and 5.4% were type 1 diabetes (T1DM). Of those T2DM, only 30.7% had HbA1c below the optimal value of 7.0%⁷.

As the number of people with diabetes grows, a proportion of national healthcare expenditure for persons with diabetes is expected to rise. Data from the Ministry of Public Health in Thailand indicate that the financial burden of diabetes was on the rise, and that the rate of diabetes-related hospitalization increased from 795.04 per 100,000 in the year 2000 to 1,081.25 per 100,000 in the year 2013⁸. High costs of diabetes and its complications also impose dramatic impacts on patients and their families. Effective interventions or initiatives are, therefore, urgently needed to reduce the incidence and financial impact of diabetes complications in the Thai population. These interventions would also help improve life expectancy and quality-of-life of patients with diabetes.

According to the Thai Clinical Practice Guideline for Diabetes⁹, insulin is recommended to be an add-on therapy when patients with T2DM receiving the first-line oral anti-diabetic drug, i.e. metformin, are unable to achieve glycemic control. In addition to insulin, other add-on therapies include thiazolidinedione, dipeptidyl peptidase-4 inhibitors, alpha-glucosidase inhibitors, and repaglinide. The target population in the study, therefore, included patients with T2DM who had uncontrolled blood glucose with oral anti-diabetic drugs and received a combination therapy.

Long-acting human insulin analogs, such as insulin glargine (IGlar), have shown potential benefits in patients in lowering severe nocturnal hypoglycemia¹⁰. Included in the National List of Essential Medicine (NLEM), insulin glargine (IGlar) is indicated for patients with type 1 diabetes who have frequent severe hypoglycemia or nocturnal hypoglycemia from Neutral Protamine Hagedorn (NPH) insulin. Recently, insulin detemir (IDet) has become available in Thailand; however, the cost of this medication is high. The decision to use and pay for this medication would depend on whether potential incremental benefits gained from this product can outweigh its costs. Previous studies have reported the value for money of IDet compared to IGlar; however, their findings were inconsistent. Guillermin et al.¹¹ suggested that treating T2DM patients with IDet was associated with comparable HbA1c change, but higher costs compared with IGlar in Canada. On the other hand, changing from IGlar to IDet was found to be a cost-saving option in the German¹² and Chinese¹³ settings. Given this inconsistent evidence, there is a need to estimate the long-term cost-effectiveness of IDet compared with IGlar for the treatment of patients with T2DM to guide the IDet reimbursement decision-making process from a payer’s perspective. This study was, therefore, conducted to evaluate the long-term cost-effectiveness of IDet compared to IGlar from the perspective of the payer in Thailand. The selection of the comparator (IGlar) for this study was made based on the recommendation by Thai stakeholders, including endocrinologists, policy-makers, and patients who were interested in identifying a new and cost-effective insulin analog that can be used as an alternative to NPH insulin.

Methods

IMS CORE diabetes model

We used a validated computer simulation model of diabetes, namely The IMS Center for Outcomes Research (IMS CORE) model version 8.5, to estimate the long-term costs and clinical outcomes. Details of this model are described elsewhere^14,15. The IMS CORE model composed of 15 sub-models which simulate the diabetic complications of angina, myocardial infarction, congestive heart failure, stroke, peripheral vascular disease, neuropathy, foot care, retinopathy, macular edema, cataract, nephropathy, hypoglycemia, ketoacidosis, lactic acidosis, and non-specific mortality. Each sub-model encompasses a Markov model using Monte Carlo simulation which incorporates a treatment sequence. All-cause mortality rate used in the model was adjusted with the age-specific mortality rate of Thai people¹⁶.

A series of inter-connected time-dependent models represent patient characteristics, clinical parameters, risk factors, treatment effects, complications, and economic parameters over time. This model has been validated and tested in various settings to ensure the reliability of simulated outcomes^15,17.

Simulation cohorts

A hypothetical simulation cohort was patients with T2DM. Baseline characteristics and clinical data were derived from Thai Diabetes Registry (TDR) Reports in 2003¹⁸ and 2006¹⁹, the National Health Security Office (NHSO)²⁰, the National Statistical Office²¹, the Center for Alcohol Studies²², and other published studies^23–27 (Table 1).

Table 1. Baseline demographics and complications of patients in a simulated cohort.

Parameter	Mean ± SD	Reference
Patient demographics
Mean age (years)	60.9 ± 11.5	TDR2006^19
Duration of diabetes (years)	10.5 ± 7.6	TDR2006^19
Proportion of male (%)	33.8	TDR2006^19
Risk factors
HbA1c level (%)	8.1 ± 1.8	TDR2006^19
Systolic blood pressure (mmHg)	143.6 ± 22.4	TDR2006^19
Total cholesterol (mg/dl)	197.1 ± 42.2	TDR2006^19
High density lipoprotein cholesterol (mg/dl)	53.4 ± 14.9	TDR2006^19
Low density lipoprotein cholesterol (mg/dl)	114.4 ± 35.5	TDR2006^19
Triglycerides (mg/dl)	153.1 ± 105.5	TDR2006^19
Body mass index (kg/m^2)	25.7 ± 4.3	TDR2006^19
eGFR (ml/min/1.73 m^2)	82.0 ± 27.5	Krairittichai et al.^23
Proportion of smokers (%)	3.9	NHSO2012^20
Number of cigarettes smoked (per day)	10.3	NSO^21
Alcohol consumption (oz/week)	4.55	CAS^22
Racial characteristics
Asian/Pacific Islander (%)	100	Assumption
Baseline cardiovascular disease complications
Myocardial infarction (%)	0.5	NHSO2012^20
Angina pectoris (%)	0.3	NHSO2012^20
Peripheral vascular disease (%)	0.23	hospital database^35
Stroke (%)	0.4	NHSO2012^20
Congestive heart failure (%)	0.4	NHSO2012^20
Atrial fibrillation (%)	0.2	NHSO2012^20
Baseline renal complications
Microalbuminuria (%)	18.0	TDR2006^19
Gross proteinuria (%)	26.1	TDR2006^19
End-stage renal disease (%)	0.1	Nitiyanant et al.^24
Baseline retinopathy complications
Background diabetic retinopathy (%)	22.0	TDR2006^19
Proliferative diabetic retinopathy (%)	9.4	TDR2006^19
Severe vision loss (%)	1.5	TDR2006^19
Baseline macular edema
Macular edema (%)	2.5	Supapluksakul et al.^25
Baseline cataract
Cataract (%)	42.8	TDR2006^19
Baseline foot ulcer complications
Uninfected ulcer (%)	5.9	TDR2003^18
Infected ulcer (%)	1.2	Nitiyanant et al.^24
Healed ulcer (%)	6.9	Nitiyanant et al.^24
History of amputation (%)	1.5	TDR2006^19
Baseline neuropathy
Neuropathy (%)	16.8	TDR2003^18, Chuengsamarn et al.^26
Baseline depression
Depression (%)	28.0	Thaneerat and Tangwongchai^27

CAS, Center for Alcohol Studies; NHSO, National Health Security Office; NSO, National Statistical Office; TDR, Thai Diabetes Registry.

Treatment effects

We obtained the treatment effect of IGlar from The Thai Insulin Therapy Assessment (TITAN) Program²⁸, which evaluated efficacy and adverse events of insulin use in 41 hospitals in Thailand. We performed a systemic search using MEDLINE, EMBASE, and Clinicaltrial.gov from their inception to August 2014. However, we did not identify any randomized controlled trials (RCTs) comparing IDet and IGlar in Thailand, but found a meta-analysis reporting weighted mean difference (WMD) of IDet conducted by Swinnen et al.²⁹. The systematic review included four RCTs comparing insulin detemir with insulin glargine, with a duration of 12 weeks or longer^30–33. We obtained the baseline rate of hypoglycemia of IGlar from the TITAN Program²⁸ and calculated the IDet hypoglycemia rate based on a meta-analysis by Swinnen et al.²⁹. Because none of the patients receiving IGlar in the TITAN Program experienced severe hypoglycemia²⁸, we derived this data from the report by the Canadian Agency for Drugs and Technologies in Health (CADTH)³⁴. The same approach was used to derive the rate of severe hypoglycemia in our previous study³⁵. Severe hypoglycemia of IDet was also calculated from the study conducted by Swinnen et al.²⁹. We pooled the change in body mass index (BMI) from baseline from four RCT studies^30–33, and found that patients receiving IDet had significantly less BMI change than those receiving IGlar (mean difference = −0.35) (Table 2). This is attributable to significantly less weight gain found in patients treated with IDet compared to patients treated with IGlar (mean difference = −0.91, 95% CI = −1.21 to −0.61)²⁹.

Table 2. Treatment efficacy and adverse events.

Description of efficacy and adverse event	Mean (95% CI)	Reference
HbA1c reduction at 6 months (%)
Glargine	−1.64 (–3.70–0.42)	TITAN^28 CADTH^34
Detemir	−1.57 (–1.40– –1.74)	Calculation^a
Weighted mean difference in HbA1c reduction (IDet vs IGlar) (%)	0.07 (–0.10–0.24)	Cochrane^29
Rate of severe hypoglycemia (times per 100 persons-year)
Glargine	0.69 (0.43–1.13)	TITAN^28
Detemir	0.61 (0.41–0.90)	Calculation^b
Rate ratio of severe hypoglycemia (IDet vs IGlar)	0.88 (0.59,1.30)	Cochrane^29
Rate of hypoglycemia (times per 100 persons-year)
Glargine	90.94 (65.73–116.15)	TITAN^28
Detemir	90.94 (81.85–100.94)	Calculation^c
Rate ratio of hypoglycemia (IDet vs IGlar)	1.00 (0.90,1.11)	Cochrane^29
Body mass index change from baseline
Glargine	0.98 (0.53–1.43)	RCTs^30–33
Detemir	0.63 (0.22–1.04)	RCTs^30–33

Note: Severe hypoglycemia required assistance from another person as a result of hypoglycemia, regardless of with/without blood sugar measurement, but the recovery was attributable to restoration of blood glucose to normal.

^a HbA1c reduction of IDet =0.07 + (–1.64) = –1.57; Upper 95% CI = –1.64 + (–0.10) = –1.74, Lower 95% CI = –1.64 + (0.24) = –1.40.

^b Rate of severe hypoglycemia of IDet =0.69*0.88 = 0.61; Upper 95% CI =0.69*1.30 = 0.90, Lower 95% CI =0.69*0.59 = 0.41.

^c Rate of hypoglycemia of IDet =90.94*1.00 = 90.94; Upper 95% CI =90.94*1.11 = 100.94, Lower 95% CI =90.94*0.90 = 81.85.

Costs and perspective

Considering the costs involved in the treatment of diabetes, from the perspective of the payer, only direct medical costs, namely the cost of insulin and costs associated with diabetes-related complications and management, were included. Insulin treatment costs were estimated from the daily dose and unit costs. Unit costs of IDet and IGlar were obtained from the Drug and Medical Supply Information Center (DMSIC)³⁶, Ministry of Public Health, which provides qualified and unbiased information about medicines and medical supplies to the Thai population. According to the Thai Health Technology Assessment Guidelines³⁷, the use of median drug price is recommended for a study on cost-effectiveness. If each drug has more than one brand name, the median drug price from all the listed original drugs should be used. Both IDet and IGlar had only one brand, with the median price of 2,801.26 THB (77.64 USD) and 2,792.70 THB (77.40 USD), respectively. The drug daily dose was pooled from four RCT studies^30–33. IGlar was used once daily, with 0.57 (95% CI =0.46–0.69) units per kilogram. IDet, on the other hand, was classified into mixed dose, once, and twice daily doses. The mixed dose of 0.83 (95% CI =0.77–0.89) units per kilogram was used for the base-case scenario. The once-daily dose was 0.50 (95% CI =0.11–0.89) units per kilogram, and the twice-daily dose was 0.93 (95% CI =0.88–0.98) units per kilogram. The IDet doses were varied in the sensitivity analysis as well. The average weight of 69.67 kilograms was calculated from 25.7 kilograms per square meter BMI of Thai diabetes¹⁹ and the average height of 1.65 meters³⁸. The daily doses of IDet and IGlar were 57.83 units per day and 39.71 units per day, respectively. Based on the daily dose and the unit price listed above, the acquisition cost was 539.97 THB (14.97 USD) per day, or 197,087.79 THB (5,462.52 USD) per year, for IDet, and 369.69 THB (10.25 USD) per day, or 134,935.85 THB (3,739.91 USD) per year, for IGlar. Costs associated with diabetes-related complications, management, and insulin adverse events were primarily derived from hospital databases and Thai published literature^39–44 (Table 3). Data regarding other oral anti-diabetic drugs were not taken into consideration, because we assumed that their costs were equal for patients receiving both insulin alternatives and, therefore, cancelled out from the analysis. Hospital databases contain four sub-databases, which are inpatient, outpatient, pharmacy, and financial databases. Information regarding age, gender, health insurance coverage, date of birth, and diagnosis codes (International Statistical Classification of Diseases 10th Revision; ICD-10) are found in the inpatient and outpatient databases. The pharmacy database includes the medication name, regimen for each prescription, and amount of medication prescribed per prescription. Lastly, the financial database contains charges for medication, laboratory, and medical services. All the databases are linked using a unique encrypted identification code for each patient. All the costs were taken for calculation after being adjusted for inflation with the consumer price index⁴⁵, and presented in the year 2015. The costs were converted at the rate of 36.08 THB per USD as of December 30, 2015⁴⁶.

Table 3. Costs of diabetes complications and management.

Description of complication or event	Mean (THB per year)	SD	Reference
Management cost
Cost of aspirin	185	122.38	hospital database^35
Cost of statin	1,748	4,571.43	hospital database^35
Cost of ACEI	1,121	2,691.21	hospital database^35
Cost of antidepressants	2,169	5,435.24	hospital database^35
Screening cost for microalbuminuria	322	—	Tertiary care hospital^39
Screening cost for gross proteinuria	60	—	Tertiary care hospital^39
Cost incurred in stopping ACEIs due to side-effects	0	—	Assumption
Eye screening	130	—	Pornpinatepong^40
Foot screening program	71	—	Standard cost list^41
Direct cost of CVD complications
Myocardial infarction, first year	106,663	130,023.29	hospital database^35
Myocardial infarction, second year onward	24,967	41,567.09	hospital database^35
Angina, first year	60,196	84,040.73	hospital database^35
Angina, second year onward	18,742	28,421.06	hospital database^35
Congestive heart failure, first year	58,523	79,794.29	hospital database^35
Congestive heart failure, second year onward	24,533	39,111.64	hospital database^35
Stroke, first year	71,875	71,875	hospital database^35
Stroke, second year onward	23,258	32,079.58	hospital database^35
Stroke, death within 30 days	38,464	42,078.42	hospital database^35
Peripheral vascular disease, first year	157,289	278,367.78	hospital database^35
Peripheral vascular disease, second year onward	49,594	49,497.75	hospital database^35
Direct cost of renal complications
Cost of HD first year	455,371	455,371	Teerawattananon et al.^42
Cost of HD second year onward	431,220	431,220	Teerawattananon et al.^42
Cost of PD first year	463,438	463,438	Teerawattananon et al.^42
Cost of PD second year onward	411,014	411,014	Teerawattananon et al.^42
Cost of RT first year	934,673	934,673	Tertiary care hospital^43
Cost of RT second year onward	432,327	432,327	Tertiary care hospital^43
Direct cost of acute events
Costs of major hypoglycemia	28,056	71,294.76	hospital database^35
Cost of ketoacidosis event	13,380	36,660.21	hospital database^35
Cost of lactic acidosis event	65,189	98,212.74	hospital database^35
Direct cost of eye disease
Cost of laser treatment	1,934	1,934	Pornpinatepong^40
Cost of cataract operation	7,050	7,050	NHSO^44
Cost of blindness—year of onset	30,213	17,162.95	hospital database^35
Cost of blindness—following years	17,550	33,122.45	hospital database^35
Direct cost of neuropathy and foot complications
Cost of neuropathy first year	23,228	37,762.17	hospital database^35
Cost of neuropathy second year onward	17,366	28,339.21	hospital database^35
Cost of amputation prosthesis	48,713	59,446.86	hospital database^35
Cost of gangrene treatment	77,177	95,754.63	hospital database^35
Cost of infected ulcer	52,558	75,022.62	Assumption (equal to that of uninfected ulcer)
Cost of standard uninfected ulcer	52,558	75,022.62	hospital database^35

NHSO, National Health Security Office; THB, Thai Baht; SD, standard deviation.

If there was no SD in some costs, we assumed SD as equaling the mean.

Discounting and time horizon

All the costs and outcomes were discounted at the rate of 3% per annum in accordance with the Thai Health Technology Assessment Guideline⁴⁷. To capture the long-term costs and outcomes, the time horizon was expanded to 50 years, as the starting age of our study cohort was 61 years.

Sensitivity analyses

A series of one-way sensitivity analyses were performed to examine the influence of key input parameters on the outcomes projected by the model, including change in HbA1c, rate of hypoglycemia, discount rate, daily dose of IDet, costs, and utilities. The results are displayed as a tornado diagram. Furthermore, a probabilistic sensitivity analysis was performed using a non-parametric bootstrapping approach. It is a re-sampling procedure that estimates an empirical sampling distribution for costs and effects. Bootstrap samples of the same size as the original data are drawn with replacement from the original sample. Repeating the process many times generates bootstrap replicates of costs and effects, which can be used to calculate the incremental cost-effectiveness ratio (ICER)⁴⁸. In this model, a cohort of 1,000 non-identical patients was set. Each patient with different baseline characteristics was simulated 1,000 times. The mean and the standard deviation of costs, life expectancy, and quality-adjusted life expectancy were calculated. The result was presented as a cost-effectiveness acceptability curve (CEAC). As recommended by the Health Economic Working Group (HEWG) under the Sub-committee for Development of the NLEM⁴⁹, a ceiling threshold of 160,000 THB per QALY (4,434.59 USD per QALY) gained was used to justify the cost-effectiveness of new intervention in Thailand in this study.

Results

Base-case analysis

Our study suggests that IDet incurred much higher total costs (3,262,268 THB vs 2,405,599 THB, or 90,417.63 USD vs 66,674.03 USD), but led to minimal improvement in life years (LYs) and QALYs compared with IGlar (13.119 LYs vs 13.116 LYs; 8.921 QALYs vs 8.908 QALYs), leading to an extremely high ICER of ∼1.7 million USD per QALY. However, if IDet was used as a single dose, the result became cost-saving, indicating the lower cost, but higher effectiveness of IDet compared with IGlar. On the contrary, the ICER was increasing when a double dose of IDet was used, as shown in Table 4. The large cost component was primarily driven by the medication cost. The acquisition cost of IDet was 2,701,682 THB (74,880.32 USD), which was greater than that of IGlar (1,849,320 THB or 51,256.10 USD). The costs of treatment complications were virtually similar between IDet and IGlar (Table 5).

Table 4. Summary of base-case analysis.

	Total cost (THB/USD)	Life year	QALY	ICER^a (THB/USD per life year)	ICER^a (THB/USD per QALY)
Insulin Glargine	2,405,599/66,674.03	13.116	8.908	–	–
Insulin Detemir (mixed dose)	3,262,268/90,417.63	13.119	8.921	285,556,370^b/7,914,533.54	62,990,376^b/1,745,852.99
Insulin Detemir (single dose)	2,188,104/60,645.90	13.119	8.921	Cost-saving^c	Cost-saving^c
Insulin Detemir (double dose)	3,587,769	13.119	8.921	394,056,526^d/10,921,744.07	86,924,234^d/2,409,208.26

ICER, Incremental cost-effectiveness ratio; QALY, Quality adjusted-life year; THB, Thai Baht.

^a ICER = (total cost of IDet – total cost of IGlar)/(Effect of IDet – Effect of IGlar). The ICER reported was slightly different from calculation, based on the formula, simply due to the rounding errors.

^b ICER comparing insulin detemir (mixed dose) with insulin glargine.

^c Cost saving indicated the lower cost, but higher effectiveness of insulin detemir (single dose) compared with insulin glargine.

^d ICER comparing insulin detemir (double dose) with insulin glargine.

Table 5. Cost component associated with insulin detemir and insulin glargine.

Cost component	Insulin detemir (THB per person)	Insulin glargine (THB per person)	Incremental difference (THB per person)
Medication	2,701,682	1,849,320	852,362
Management	25,431	25,437	−6
Cardiovascular disease	123,376	123,236	140
Renal complications	219,361	215,116	4,245
Ulcer/amputation/neuropathy	149,969	149,770	199
Eye complications	35,024	34,963	61
Hypoglycemia	2,329	2,691	−362
Anti-depression treatment	5,097	5,066	31
Total cost (THB)	3,262,268	2,405,599	856,669

Sensitivity analyses

The cost of IDet showed the greatest impact on the ICER estimates (Figure 1). When the cost of IDet declined from the base case by 34% (from 197,087.79 THB (5,462.52 USD) per year to 130,000 THB (3,603.10 USD) per year), IDet became a cost-effective option compared to the ceiling threshold of 160,000 THB per QALY (Table 6). When IDet was used as a once-daily dose (118,727.58 THB (3,290.68 USD) per year), it became a cost-saving strategy. Figure 2 illustrates that IDet is unlikely to be a cost-effective strategy, even if the ceiling ratio was set at 3 million THB.

Figure 1. Tornado diagram of insulin detemir vs insulin glargine.

Figure 2. The cost-effectiveness acceptability curve of insulin detemir vs insulin glargine.

Table 6. Threshold analysis varying cost of insulin detemir.

Cost of insulin detemir (THB)		Percent reduction from base case	Incremental cost-effectiveness ratio (THB/QALY)
Cost per year^b	Cost per unit^a
40,000	568.53	80	Cost-saving
70,000	944.93	64	Cost-saving
100,000	1,421.33	49	Cost-saving
130,000	1,847.72	34	158,893
160,000	2,274.12	19	28,254,770
190,000	2,700.52	4	56,350,648

^a Cost per unit at base case was 2,801.26 THB (100 units per ml; 3 ml, five cartridges).

^b Cost per year was 197,087.79 THB.

Discussion

Our study reported the long-term effectiveness and cost-effectiveness of IDet compared with IGlar from the Thai payer’s perspective. We found that IDet was associated with higher costs and better QALYs than IGlar; however, its benefit, measured as QALY gained, did not offset its high acquisition costs. This conclusion is due mainly to the extremely high total cost of IDet compared with IGlar (3,262,268 THB vs 2,405,599 THB or 90,417.63 USD vs 66,674.03 USD) and a negligible gain in QALYs (8.921 vs 8.908 for IDet and IGlar, respectively). The difference in total costs was mainly from the medication. Our findings indicated that IDet has become a cost-saving alternative compared with IGlar when it is used as a once-daily treatment. Treatment effects, such as HbA1c reduction and the incidence of severe hypoglycemia, were not significantly different between IDet and IGlar.

Despite different healthcare system, our findings were in line with a previous study based on the CORE Diabetes Model, suggesting that treatment of -T2DM patients with IDet led to higher lifetime costs and insignificant change in HbA1c compared with IGlar, regardless of IDet regimens¹¹. Consistent findings may in part be due to that of Guillermin et al.¹¹, who pooled the relative efficacy and safety of IDet and IGlar from the same head-to-head RCTs^30–33 as our study.

Few studies reported that IDet was economically attractive compared to IGlar. IDet was cost-effectiveness from the perspective of American⁵⁰ and German payers¹². Similarly, Yang et al.¹³ suggested that switching from IGlar to IDet was a cost-saving option for Chinese T2DM patients. It should be noted that the studies reporting a favorable cost-effectiveness profile obtained the efficacy of IDet from a single RCT or observational study that showed the larger reduction in HbA1c, the lower increase in BMI, and the fewer rates of major hypoglycemic events in IDet than IGlar. The American study derived the efficacy of IDet from a head-to-head RCT of IDet + Insulin Aspart (IAsp) and IGlar + IAsp in type 1 diabetes patients, while the German¹² and Chinese¹³ studies obtained the treatment effect of IDet from the Predictable Results and Experience in Diabetes through Intensification and Control to Target: An International Variability Evaluation (PREDICTIVE) study, which was an international, multi-center, open-label observational study involving more than 35,000 patients with T1DM or T2DM transferred to an IDet regimen.

Our study was conducted to evaluate the cost-effectiveness between IGlar and IDet within the Thai context. The baseline demographics, treatment effects, and complications were mainly derived from the TDR database, which enrolled 9,419 diabetes patients from different parts of Thailand. Besides drug acquisition costs and diabetes complications, treatment costs were estimated as closely to costs incurred by the Thai healthcare payer as possible. Those costs were mostly derived from hospital databases, which reflect the actual practice of diabetes treatment in Thailand. Even though the treatment effects of insulin analogs were derived from a meta-analysis of RCTs conducted in other countries²⁹, the efficacy and adverse events of IGlar were derived from a large Thai cohort study²⁸.

Our study adds the following contributions to the existing literature. First, as a previous study demonstrated, IDet leads to lower weight gain than IGlar²⁹: we incorporated the change in the BMI from baseline into the CORE diabetes model. Second, several scenarios of data analysis were performed based on the different daily injection doses of IDet. The findings indicated that ICER was sensitive to the daily dose of IDet. Finally, the total lifetime direct costs were estimated for each insulin treatment, which reflected the insulin costs incurred by the payer. The key strength of our study was that the findings from this study were considered as part of the economic evidence that can be used to support the selection of long-acting insulin analogs in Thailand. In addition, findings from our study can be used to inform the decision-making process in other countries, with similar socioeconomics levels and healthcare systems to Thailand, as they might have comparable costs of treatments, baseline demographic characteristics, and willingness to pay values to our study.

In Thailand, the medicines that are high cost but are important for particular groups of patients, economic evaluation, and budget impact analysis information is required for the sub-committee for the development of the NLEM to consider in decision-making. The health technology assessment (HTA) process for pharmaceutical reimbursement involves prioritizing the list of HTA topics according to the burden of disease and the degree of severity by the sub-committee for the development of the NLEM, performing a health economic evaluation by a Health Economic Working Group commissioned by the sub-committee, appraising the methodological quality of a health economic evaluation report by experts in the field, and considering the health economic evaluation evidence during the reimbursement decision by the sub-committee⁵¹.

This study encompasses a certain number of limitations. First, given the lack of local efficacy data, we assumed that the treatment effects of IDet observed in the published RCTs were generalizable to the Thai T2DM patients. We acknowledged that the daily dose of both types of insulin in other countries might be slightly higher than that being used in general practice in Thailand. Second, the utility values were mostly derived from the data from other countries; however, the result of sensitivity analysis suggested that the impact of utility on cost-effective findings was negligible. Third, we assumed that T2DM patients would use the same insulin for lifetime; however, it is possible that patients may switch to another insulin for a variety of reasons in real practice setting. Finally, our study did not take into account the more recent meta-analysis by Freemantle et al.⁵², showing that IGlar 300 units/mL (Gla-300) was associated with a reduction in HbA1c, but an increase in body weight over IDet; however, none of these changes were statistically significant. Similarly, this meta-analysis also shows that IGlar-300 was associated with a numerically lower rate of documented symptomatic and nocturnal hypoglycemic events. This recent meta-analysis was omitted from our study because it was not available at the time of our systematic search. If the findings of Freemantle et al.⁵² had been used in this study, IGlar might have shown a cost-saving option compared to IDet, due to the lower rate of hypoglycemia.

Based on the findings of the study, IDet is not economically attractive compared to IGlar from the Thai payer’s perspective. However, the decision to make an appropriate choice needed to consider other factors such as drug allergy, adherence, frequency of administration, ease of use, and availability and accessibility of the insulin among Thai T2DM patients who are usually advanced in age.

Conclusions

From the Thai payer’s perspective, treatment with IDet in patients with T2DM who had uncontrolled blood glucose with oral anti-diabetic drugs was not a cost-effective insulin analog compared with IGlar treatment. Cost of IDet was the most influential factor that played a major role in the study findings. Our findings could be generalized to other countries, with similar socioeconomics levels and healthcare systems.

Transparency

Declaration of funding

This study was supported by the CMU Mid-Career Research Fellow Program, Chiang Mai University, Thailand.

Declaration of financial and other relationships

The authors report no conflicts of interest.

Acknowledgments

The authors gratefully acknowledge the IMS Health team for support and training in the use of the IMS CORE Diabetes Model.