Journal of Medical Economics

Journal of Medical Economics Downloaded from informahealthcare.com by Gazi Univ. on 08/20/14 For personal use only.

1369-6998 doi:10.3111/13696998.2014.912986

Vol. 17, No. 7, 2014, 499–507

Article 0013.R1/912986 All rights reserved: reproduction in whole or part not permitted

Original article Cost-effectiveness of metformin plus vildagliptin compared with metformin plus sulphonylurea for the treatment of patients with type 2 diabetes mellitus: a Portuguese healthcare system perspective

Daniel Viriato

Abstract

Novartis Farma SA, Porto Salvo, Portugal

Frederico Calado Jean-Bernard Gruenberger Siew Hwa Ong

Objective: To evaluate the cost-effectiveness of vildagliptin plus metformin vs generic sulphonylurea plus metformin in patients with type 2 diabetes mellitus, not controlled with metformin, from a Portuguese healthcare system perspective.

Novartis Pharma AG, Basel, Switzerland

Davide Carvalho Sa˜o Joa˜o Hospital, Faculty of Medicine, University of Porto, Portugal

Jose´ Silva-Nunes Curry Cabral Hospital, Lisbon, Portugal

Sukhvinder Johal HERON Commercialization – A PAREXEL Company, London, UK

Ricardo Viana Novartis Farma SA, Porto Salvo, Portugal Address for correspondence: Daniel Viriato, Health Economics and Outcomes Research, Novartis Farma – Produtos, Farmaceuticos S.A., Apartado 200, P-2711-901 Sintra, Portugal. Tel.: +351 21 0008693; Fax: +351 21 0008806; [email protected] Keywords: Diabetes mellitus – Vildagliptin – Cost-effectiveness – Portugal – Sulphonylurea – Metformin Accepted: 4 April 2014; published online: 23 April 2014 Citation: J Med Econ 2014; 17:499–507

Methods: A cost-effectiveness model was constructed using risk equations from the UK Prospective Diabetes Study Outcomes Model with a 10,000-patient cohort and a lifetime horizon. The model predicted microvascular and macrovascular complications and mortality in yearly cycles. Patients entered the model as metformin monotherapy failures and switched to alternative treatments (metformin plus basal-bolus insulin and subsequently metformin plus intensive insulin) when glycated hemoglobin A1c 47.5% was reached. Baseline patient characteristics and clinical variables were derived from a Portuguese epidemiological study. Cost estimates were based on direct medical costs only. One-way and probabilistic sensitivity analyses were conducted to test the robustness of the model. Results: There were fewer non-fatal diabetes-related adverse events (AEs) in patients treated with metformin plus vildagliptin compared with patients treated with metformin plus sulphonylurea (6752 vs 6815). Addition of vildagliptin compared with sulphonylurea led to increased drug acquisition costs but reduced costs of AEs, managing morbidities, and monitoring patients. Treatment with metformin plus vildagliptin yielded a mean per-patient gain of 0.1279 quality-adjusted life years (QALYs) and a mean per-patient increase in total cost of E1161, giving an incremental cost-effectiveness ratio (ICER) of E9072 per QALY. Univariate analyses showed that ICER values were robust and ranged from E4195 to E16,052 per QALY when different parameters were varied. Limitations: The model excluded several diabetes-related morbidities, such as peripheral neuropathy and ulceration, and did not model second events. Patients were presumed to enter the model with no diabetes-related complications. Conclusion: Treatment with metformin plus vildagliptin compared with metformin plus sulphonylurea is expected to result in a lower incidence of diabetes-related AEs and to be a cost-effective treatment strategy.

! 2014 Informa UK Ltd www.informahealthcare.com/jme

Cost-effectiveness of vildagliptin in Type 2 diabetes Viriato et al.

499

Journal of Medical Economics

Volume 17, Number 7

July 2014

Journal of Medical Economics Downloaded from informahealthcare.com by Gazi Univ. on 08/20/14 For personal use only.

Introduction Diabetes mellitus (diabetes) is a progressive, chronic disease characterized by frequent episodes of hyperglycemia. Type 2 diabetes (T2D) is the most prevalent form of the disease, accounting for nearly 90% of all cases worldwide1. The treatment of diabetes imposes a large economic burden on the individual, national healthcare system, and economy. In Portugal, for example, in 2011, the total costs associated with T2D represented 0.8–1.0% of the Gross Domestic Product of Portugal and 8–10% of total national health expenditure2. Complications have a significant impact on the costs of managing T2D3, and the direct medical costs of treating T2D are considerably higher for individuals who have fair or poor glycemic control than for those who have good glycemic control4,5. The main treatment objective is good long-term glycemic control, with glycated hemoglobin A1c (HbA1c) being the gold standard follow-up assessment6. The American Diabetic Association and European Association for the Study of Diabetes recommend a target of HbA1c 57.0% in most patients but recognize the value of individualizing treatment targets according to a variety of patient/disease characteristics7. Metformin remains the most widely used first-line T2D drug7. If lifestyle interventions and the maximum tolerated dose of metformin monotherapy alone do not achieve the HbA1c goal, treatment intensification with a second oral anti-hyperglycemic agent, such as a sulphonylurea, a glucagon-like peptide-1 (GLP-1) receptor agonist, a dipeptidyl peptidase-4 (DPP-4) inhibitor, or a basal insulin is recommended7–9. Vildagliptin, a DPP-4 inhibitor, has been approved for use in patients with T2D within the European Union (EU) since 200810,11. The safety and efficacy of vildagliptin, either as monotherapy or in combination with metformin, has been established in multiple studies10,12–15. Vildagliptin added to metformin has demonstrated an efficacy comparable with that for alternative treatments for T2D, including glimepiride (a sulphonylurea) and pioglitazone (a thiazolidinedione)14,16. Although sulphonylureas have a lower acquisition cost than vildagliptin, their use is more expensive to society than other medications after the management of sideeffects has been considered17. For example, the use of vildagliptin is associated with a significantly reduced risk of hypoglycemia compared with glimepiride10,14. Patients’ fear of hypoglycemic episodes can lead them to maintain blood glucose levels within a perceived ‘safety margin’, at higher levels than normal, which is associated with an increased risk of complications and associated healthcare costs18–20. The aim of this study was to compare, from a Portuguese healthcare system perspective, the cost-effectiveness of 500

Cost-effectiveness of vildagliptin in Type 2 diabetes Viriato et al.

vildagliptin vs generic sulphonylureas as a second-line therapy when used in combination with metformin after failure of metformin monotherapy in T2D patients. The methodology and preliminary performance of this model have been presented previously21.

Methods Model design A cost-effectiveness model was developed to demonstrate the long-term cost-effectiveness of metformin therapy in combination with vildagliptin or sulphonylurea in patients with T2D in the Portuguese healthcare setting. The model was constructed as a patient-level simulation model, utilizing the risk equations from the UK Prospective Diabetes Study (UKPDS) Outcomes Model22 to predict microvascular and macrovascular complications and mortality (both disease-specific and all-cause) over a lifetime horizon, taken as 40 years. The model simulated a cohort of 10,000 patients in yearly cycles, which enabled a reliable estimate of the expected costs and benefits, and was adapted to include weight gain and hypoglycemic events specific to each therapy. An overview of the patient simulation model is presented in Figure 1. A patient-level simulation was chosen because diabetes is a complex disease with multiple, simultaneous complications. The risk calculation engine was taken from the UKPDS Outcomes Model and was designed to capture the association between different types of complications at an individual patient level22. The model’s underlying risk equations are used widely for modeling T2D. This structure allows patients to develop multiple complications within each model cycle and over the simulation period. Patients entered the model with certain ‘static’ patient characteristics (see Table 1) and other ‘dynamic’ patient covariates that were determined by the risk equations and updated at the end of every cycle. At the start of the model, a patient was presumed to have no diabetic complications. In each subsequent cycle, patients could experience one of the following diabetes-related complications, according to their risk profile: ischemic heart disease, myocardial infarction, chronic heart failure, renal failure, stroke, lower limb amputation, or blindness in one eye. In addition, three equations were specified for each of the following events: death following an event in the first year, diabetes-related death following a diabetes-related event, and death due to other causes. Risk equations were inter-linked such that, if a patient experienced an event, he or she may then be at a higher risk of experiencing a different, linked event in subsequent cycles. Specifically, in each year cycle, a patient could experience one of a number of diabetes-related complications determined by his or her risk profile. If an event occurred, the www.informahealthcare.com/jme ! 2014 Informa UK Ltd

Journal of Medical Economics

Initial Condition Module

Treatment Module

• Current treatment • History of diabetes events • Risk profile

• Check HbA1c is above or below threshold (7.5%). • If the HbA1c is above the threshold then switch to next drug in the pathway • Check if patient is intolerant to new treatment

Volume 17, Number 7

July 2014

Risk Factor Module • Update risk factors dependant on events in previous cycle

Diabetes Events • UKPDS risk engine simulates diabetes complications

Journal of Medical Economics Downloaded from informahealthcare.com by Gazi Univ. on 08/20/14 For personal use only.

Yes

Stop

Costs and Utilities

No Is Patient Alive ?

• Store result • Simulate new patient

• Update costs and QALYs

Figure 1. An overview of the model structure (adapted from UKPDS) used to estimate the cost-effectiveness of vildagliptin compared with sulphonylurea added on to metformin for the treatment of patients with type 2 diabetes mellitus from a Portuguese healthcare system perspective. HbA1c, glycated hemoglobin; QALY, quality-adjusted life-year, UKPDS, United Kingdom Prospective Diabetes Study.

Table 1. Baseline characteristics of the model cohort*. Variable

Value

Age, mean (years) Male (% patients) Duration of T2D, mean (years) Smokers (% patients) Serum HbA1c, mean (%) Systolic blood pressure (mm/Hg) Serum TC:HDL ratio, mean BMI, mean (kg/m2)

63 52 9.13 21 7.2 148 5:1 31.39

BMI, body mass index; HbA1c, glycated hemoglobin; HDL, high-density lipoprotein; T2D, type 2 diabetes; TC, total cholesterol. *Data extracted from a cross-sectional study of 2434 T2D patients who received pharmacological treatment in Portugal24.

risk profile of the patient was updated, and this altered the risk of future events. Patients were able to experience more than one non-fatal event during each cycle, but were only able to experience one of the same type of event for the duration of the model. That is, the model only predicted the first event in any single category of diabetes-related complications, and did not allow series of events such as sequential amputations or MI to be modeled directly. This is consistent with the UKPDS model. The risk of incurring a complication was based on Weibull regression, and a Gompertz regression model, a functional form widely used to model mortality, was used to calculate the risk of dying of natural death or diabetesrelated causes22. Non-diabetes-related death was able to occur at any point in the model. The risk of ! 2014 Informa UK Ltd www.informahealthcare.com/jme

non-diabetes-related death was a function of the cohort’s age, gender, and smoking status. To account for competing risks in the UKPDS simulation (for example, if a patient died within a cycle of the model, they could have no subsequent events), all equations were run in a random order. A pre-defined treatment algorithm allowed for a patient to switch treatments when the HbA1c level increased above a pre-specified threshold (7.5%), the patient developed intolerance to a treatment (i.e., the patient experienced hypoglycemia), or the patient failed to comply with current therapy. According to the advice of clinical experts, a Hb1Ac threshold of 7.5% was chosen to reflect current clinical practice among Portuguese physicians, which in turn reflects local population characteristics. Current Portuguese guidelines recommend a Hb1Ac threshold of 6.5% for effective glycemic control, but allow less restrictive HbA1c thresholds depending on a patient’s history of disease control, prognosis, complications, and comorbidities23. When a patient switched treatment, a drug-specific reduction in HbA1c over the first year and an annual increase in HbA1c in the following years (coefficient of treatment failure [CoF]) were applied. Patients were also able to experience side-effects such as hypoglycemic events (defined as symptoms suggestive of hypoglycemia and confirmed by self-monitored plasma glucose 53.1 mmol/L and severe hypoglycemic events (defined as any episode requiring the assistance of another party)10. The model estimated costs and quality-adjusted life years (QALYs) associated with each treatment strategy using direct medical costs only. Outputs included point Cost-effectiveness of vildagliptin in Type 2 diabetes Viriato et al.

501

Journal of Medical Economics Downloaded from informahealthcare.com by Gazi Univ. on 08/20/14 For personal use only.

Journal of Medical Economics

Volume 17, Number 7

July 2014

and probabilistic estimates for cost-effectiveness. Two treatment strategies were compared in the analysis: initial therapy with metformin plus sulphonylurea vs initial therapy with metformin plus vildagliptin, each to be followed with replacement therapy according to the rules described above. For both strategies, replacement therapies were considered to be, in the first instance, metformin plus a basal-bolus insulin regimen and, in the second instance, metformin plus an intensive or multiple-dose insulin regimen. After exhausting these options, patients were treated with the multiple-dose insulin regimen for the rest of their lifetime.

Unlike the original UKPDS model, the model developed here was able to handle adverse events. The model included two acute events (hypoglycemia and edema) and one chronic event (hip fractures). However, no treatmentspecific data on edema and fracture were available and, hence, the only treatment-related adverse event considered in the model was hypoglycemia. The risk of hypoglycemia and severe hypoglycemia for each treatment is shown in Table 2, together with the main efficacy and safety parameters of modeled interventions.

Study population

As there are no specific utilities estimates from Portugal for diabetes and its complications, baseline utility (utility for a diabetic patient without any complication) was modeled using the mean European Quality of Life-5 Dimensions (EQ-5D) value reported in an analysis of quality-of-life data from the UKPDS26. Utility decrements associated with complications were also applied and were taken from the UKPDS study26, with the exception of renal failure27, hypoglycemia28, and weight gain29. Utility decrements for each of the different complications and the utility of a diabetic patient without complications were obtained from the literature: no complications, 0.780; ischemic heart disease, 0.090, myocardial infarction (non-fatal), 0.055; congestive heart failure, 0.108; stroke (non-fatal), 0.164; lower limb amputation, 0.280; blindness in one eye, 0.07426; renal failure, 0.37927; symptomatic hypoglycemia, 0.014; severe hypoglycemia, 0.04728; and weight gain (per unit gain in BMI), 0.00529.

The patient population comprised patients who failed to achieve glycemic control with metformin monotherapy. Baseline cohort characteristics regarding demographic and clinical variables were derived from a Portuguese epidemiological study24 (Table 1).

Treatment effects and adverse events The clinical parameters considered in the cost-effectiveness model included serum HbA1c levels, weight gain, serum lipid ratio, and incidence of hypoglycemic events (Table 2). Parameter values were extracted from a headto-head clinical trial of metformin plus vildagliptin vs metformin plus sulphonylurea10. When patients were switched to a new agent, they experienced a ‘drug-specific’ reduction in HbA1c over the first year and an annual increase in HbA1c in the following years. A CoF was developed to help calculate time to treatment failure25, defined as the slope of the least-squares regression line of a glycemic index vs time plot calculated for each individual patient on constant therapy. The CoF was estimated from the Week 25 to Week 52 data of the head-to-head clinical trial10, which showed that the CoF in the vildagliptin arm was approximately one half of that in the sulphonylurea arm.

Quality-of-life

Costs Only direct medical costs were considered in the model. Annual drug acquisition costs were weighted average cost adjusted by volume, calculated based on 2013 unit costs of Portuguese list prices and recommended dose or defined daily dose (Table 3). The weighted average cost of SU was

Table 2. Main efficacy and safety parameters of modeled interventions. Variable

HbA1c reduction, Year 1 (%) Coefficient of failure* (%) Hypoglycemia (%) Severe hypoglycemia (%) Weight gain (kg/year) Change in TC:HDL (%/year)

Value

Reference

MET þ SU

MET þ VIL

0.53 0.37 16.2 0.72 1.56 2.37y

0.44 0.20 1.7 0 0.23 5.18y

10

Value

Reference

MET þ basal insulin

MET þ intensive insulin

1.1 0.2y 16.2 3.3 1.7 –

1.1 0.2y 16.2 3.3 1.7 –

38 38 38 38



HbA1c, glycated hemoglobin; HDL, high-density lipoprotein; MET, metformin; SU, sulphonylurea; TC, total cholesterol; VIL, vildagliptin. *Coefficient of failure was defined as the slope of the least-squares regression line of HbA1c vs time from Week 24 to Week 52 of Year 1 of a head-to-head clinical trial of metformin plus vildagliptin vs metformin plus sulphonylurea10. yNovartis, unpublished data.

502

Cost-effectiveness of vildagliptin in Type 2 diabetes Viriato et al.

www.informahealthcare.com/jme ! 2014 Informa UK Ltd

Journal of Medical Economics

Table 3. Annual cost of treatment in patients with T2D. Cost (E)

References

MET þ generic SU MET þ vildagliptin MET þ basal-bolus insulin regimen MET þ multiple-dose insulin regimen Insulin regimen monitoring costs

76.70 662.74 515.00 763.37 542.62

39,40 39,40 39,40 39,40 41,42

MET, metformin; SU, sulphonylurea; T2D, type 2 diabetes.

Journal of Medical Economics Downloaded from informahealthcare.com by Gazi Univ. on 08/20/14 For personal use only.

Table 4. Costs of complications.

Ischemic heart disease Myocardial infarction (non-fatal) Congestive heart failure Stroke (non-fatal) Lower limb amputation Blindness in one eye Renal failure Hypoglycemia Severe hypoglycemia Myocardial infarction (fatal) Stroke (fatal)

July 2014

Model outcomes

Treatment

Event

Volume 17, Number 7

The primary outcomes of the economic analysis were costs and QALYs and the economic evaluation took the form of a cost utility analysis that calculated the incremental cost per QALY gained of metformin plus vildagliptin compared with metformin plus sulphonylurea. Model outputs also included the cumulative incidence of events over a lifetime of clinical events occurring over time and total life-years.

Sensitivity analysis Cost per event (E)

Reference

First year

Maintenance

1889.52 2402.55

720.66* 865.64*

43

3624.13 3963.98 6437.06 11,736.00 29,354.00 371.00 1370.00 3185.44

1484.81* 1177.34* 2342.45* 11,736.00 29,354.00y – – –

43

4852.95



43

43 43 44 43 45 45 43 43

*Values calculated from first-year values using proportion presented by Clarke et al.46. y Weighted approach based on surgical diagnosis-related groups for renal failure (2008–2010) plus the estimated cost per patient for hemodialysis (2010).

based on the costs of generic sulphonylureas (branded sulphonylureas were not considered). The cost of experiencing an event in the first year was modeled as being different to the cost of managing the events in subsequent years (maintenance). Unit costs of the management of each complication in the first year and in subsequent years are listed in Table 4. Maintenance costs were applied in all subsequent years until the end of the simulation horizon or until the patient died. Costs related to stroke and myocardial infarctions were classified into fatal or non-fatal, and hypoglycemia was treated as an acute event.

Study perspective, time horizon, and discounting The cost-effectiveness of implementing each intervention was developed from the perspective of the Portuguese healthcare system, taking into consideration recommendations from the Portuguese Medicines Agency30. All relevant costs and outcomes associated with T2D were captured in yearly cycles over a patient’s lifetime. Costs and utilities were discounted annually at 5.0% in agreement with Portuguese guidelines30. ! 2014 Informa UK Ltd www.informahealthcare.com/jme

To test the robustness of the model assumptions and specific parameters, one-way sensitivity analyses were performed to examine the effect of changing the value of several parameters for the two replacement therapies, metformin plus vildagliptin and metformin plus sulphonylurea. Parameters analyzed included HbA1c initial absolute 1-year change, CoF, drug costs, hypoglycemia risk, annual weight gain, utility for a patient with diabetes, utility decrements for diabetic complications, and discount rates for costs and benefits. The effect of varying individual parameters was examined using plausible ranges of values from the literature. Drug-specific parameters were varied by 25% and utility values were varied between 95% confidence intervals26. The discount rate was varied between 0 and 8%. A scenario in which the time horizon was changed to 20 years was also tested. A probabilistic sensitivity analysis (PSA) of the basecase cost-effectiveness results was also performed, in order to assess the impact of parameter uncertainty around major model inputs. Good modeling practice usually involves varying all uncertain model parameters at the same time within a PSA. However, the structure of the model within an Excel/Visual Basic framework meant that this would have been computationally expensive and would have taken too long to have been practical. For these reasons, a distribution was over-imposed on the drug-related specific parameters only because, by definition, these are the ones that differ between drugs. Accordingly, most of the incremental uncertainty between arms was assumed to be drug-specific. The parameters that were randomly sampled in the model were: initial HbA1c reduction, HbA1c increase over time, rate of adverse events, hypoglycemia and severe hypoglycemia, and weight gain. The ranges of variation of the parameters were implemented using a beta distribution for the rate of AEs and normal distributions for the other parameters.

Results The outcomes of the model for the 10,000 simulated patients over a lifetime are shown in Table 5. Over a Cost-effectiveness of vildagliptin in Type 2 diabetes Viriato et al.

503

Journal of Medical Economics

Volume 17, Number 7

July 2014

Journal of Medical Economics Downloaded from informahealthcare.com by Gazi Univ. on 08/20/14 For personal use only.

Table 5. Number of diabetes-related events: 10,000 simulations over a lifetime time horizon. Diabetes complication

MET þ SU

MET þ VIL

Difference

Macrovascular events* Microvascular eventsy Deathsz Total non-fatal diabetes-related adverse events

5931 (2409) 884 (386) 10,000 (2149) 6815 (2795)

5876 (2292) 876 (339) 10,000 (2089) 6752 (2631)

55 (117) 8 (47) 0 (60) 63 (164)

MET, metformin; SU, sulphonylurea; VIL, vildagliptin. All values are per cohort; values in parentheses were obtained at 5 years of the model run. *Macrovascular events included ischemic heart disease, myocardial infarction (non-fatal), congestive heart failure, and stroke. yMicrovascular events included amputation, blindness in one eye, and renal failure. zDeaths included both diabetes-related and non-diabetes-related deaths.

lifetime, the model estimated fewer non-fatal diabetesrelated adverse events in patients treated with metformin plus vildagliptin compared with patients treated with metformin plus sulphonylurea (6752 vs 6815, respectively). This is as expected, given the lower annual change in the total cholesterol:high-density lipoprotein ratio for patients treated with metformin plus vildagliptin and their lower CoF, which maintained a lower HbA1c while on treatment (although this was partially counteracted by patients on metformin plus sulphonylurea switching to insulin-based therapy earlier, which led to a larger reduction in HbA1c). Patients switched to insulin therapy at 5.0 and 4.0 years for metformin plus vildagliptin and metformin plus sulphonylurea, respectively. At 5 years of the model run there was a higher proportional difference in the incidence of diabetic complications between the two treatment arms compared with lifetime. One of the differences between both treatment arms was the probability of having hypoglycemia and severe hypoglycemia. The higher incidence of hypoglycemia led to increased hypoglycemia treatment costs per patient for the metformin plus sulphonylurea arm compared with the metformin plus vildagliptin arm (E234 vs E170, respectively). Over a lifetime time horizon, the addition of vildagliptin to metformin, compared with the addition of sulphonylurea to metformin, resulted in an increased mean cost per patient of drug acquisition and in decreased mean costs per patient of managing side effects, managing comorbidities, and patient monitoring (Table 6). Treatment with vildagliptin plus metformin yielded a mean per-patient gain of 0.1279 QALYs and an increase in total cost of E1161 compared with treatment with sulphonylurea plus metformin. The ICER for metformin plus vildagliptin compared with metformin plus sulphonylurea was E9072 per QALY. 504

Cost-effectiveness of vildagliptin in Type 2 diabetes Viriato et al.

Table 6. Base-case analysis: 10,000 simulations over a lifetime time horizon. MET þ SU Cost of comorbidities, mean (E) Cost of side effects, mean (E) Cost of drug acquisition, mean (E) Cost of monitoring, mean (E) Total cost, mean (E) Life years, mean QALYs, mean ICER (E/QALY)

MET þ VIL

Difference

7400

7224

177

715

422

293

2837

4785

þ1948

2297

1978

383

13,248 7.6591 5.6401 –

14,409 7.7486 5.7681 –

þ1161 þ0.0896 þ0.1279 þ9072

ICER, incremental cost-effectiveness ratio; MET, metformin; QALY, qualityadjusted life-year; SU, sulphonylurea; VIL, vildagliptin. All costs and QALYs are mean values per patient and are discounted.

The cumulative incidence of death during the first 5 years of the model was lower for patients treated with metformin plus vildagliptin than for patients treated with metformin plus sulphonylurea (2089 vs 2149, respectively), resulting in a greater mean life expectancy for patients treated with metformin plus vildagliptin (7.75 years vs 7.66 years, respectively). Compared with metformin plus sulphonylurea, treatment with metformin plus vildagliptin was associated with a lower incidence of diabetic complications, including hypoglycemia and weight gain.

Sensitivity analysis In a one-way analysis of the sensitivity of the incremental cost per QALY of metformin plus vildagliptin compared with metformin plus sulphonylurea to variations in several model parameters, the incremental cost per QALY was most sensitive to the CoF, drug costs, HbA1c, and discount rates for costs and benefits. Nevertheless, the cost-effectiveness results remained robust to plausible variations of the main parameters used in the model. Variation in utility decrements for complications had a relatively small impact on the ICER. In a scenario analysis, changing the time horizon of the analysis from 40 years to 20 years had only a marginal effect on the results. The incremental costs and QALYs gained for metformin plus vildagliptin were E878 and 0.0842, respectively, leading to an ICER of E10,428/QALY. This was as expected because most patients had died within 20 years from the start of the model. The PSA of 100 simulated interactions suggested that, for a willingness-to-pay threshold of E30,000 per QALY, treatment with metformin plus vildagliptin had a 79% probability of being cost-effective compared with metformin plus sulphonylurea (Figure 2). www.informahealthcare.com/jme ! 2014 Informa UK Ltd

Journal of Medical Economics Downloaded from informahealthcare.com by Gazi Univ. on 08/20/14 For personal use only.

Journal of Medical Economics

Volume 17, Number 7

July 2014

Figure 2. Cost-effectiveness acceptability curve. ICER, incremental cost-effectiveness ratio; QALY, quality-adjusted life-year.

Discussion In the diabetes setting, there is increasing demand to manage the cost impact of treatment as treatment-related costs continue to increase31. As the prevalence of T2D is projected to continue to rise in Portugal32, cost-effective treatments with favorable adverse event profiles are required to minimize the associated economic burden. The current study utilized a patient-level simulation model to determine the cost-effectiveness of metformin plus vildagliptin vs metformin plus sulphonylurea for the treatment of patients with T2D in the Portuguese setting. Clinical efficacy and safety data were derived from a headto-head clinical trial of vildagliptin and sulphonylurea in combination with metformin10, and the cost-effectiveness analysis used the UKPDS algorithm to extrapolate efficacy end-points from the trial to outcomes beyond the trial period. Treatment with vildagliptin compared with sulphonylurea was projected to be cost-effective compared with generic sulphonylureas for the treatment of patients with T2D not controlled with metformin monotherapy, with an ICER within the cost-effectiveness threshold for the Portuguese setting33. A univariate sensitivity analysis showed that metformin plus vildagliptin was consistently a cost-effective therapy. The present study is, to our knowledge, the first published cost-effective analysis of vildagliptin as an adjunct to first-line metformin therapy. The results presented for the Portuguese healthcare system indicate that similar analyses in other countries may be appropriate. In the present study, initial treatment with vildagliptin was more expensive than with sulphonylurea. However, the cost of managing adverse events and comorbidities was significantly lower. This is particularly important as a study by Kanavos et al.34 showed that inpatient ! 2014 Informa UK Ltd www.informahealthcare.com/jme

costs, due to increased medical care associated with diabetes-related complications, are consistently higher than outpatient costs. Hypoglycemia is a common T2D-related adverse event and is considered the key limiting step for optimizing glycemic control. In the context of current clinical guidelines which recommend strict glycemic control, intensive therapy may be associated with an increased risk of hypoglycemia, which may lead to dysrhythmias as well as accidents and falls, dizziness, confusion, or infection. Accordingly, in at-risk individuals, blood glucose targets may need to be moderated, and drug selection should favor agents that do not precipitate such events7. The present model reported a higher incidence of hypoglycemia and increased hypoglycemia treatment costs for the metformin plus sulphonylurea arm compared with the metformin plus vildagliptin arm, which contributed to a favorable ICER for treatment with metformin plus vildagliptin. The model used in the present study has a number of strengths and limitations. The foremost strength of the model was the inclusion of the risk equations from the UKPDS Outcomes Model22, which allow simulation of a range of long-term outcomes while taking into account the association between different types of complications at an individual level35. Furthermore, the model was also robust in accounting for treatment strategies that patients with T2D typically follow over their lifetime. With regard to limitations, the UKPDS Outcomes Model22, from which the model was derived, does not explicitly model second events within any event category. However, the occurrence of second events is not ignored, in the sense that the associated excess mortality and excess resource use are indirectly captured by higher risks of death and higher healthcare costs associated with patients with a history of events36. Cost-effectiveness of vildagliptin in Type 2 diabetes Viriato et al.

505

Journal of Medical Economics Downloaded from informahealthcare.com by Gazi Univ. on 08/20/14 For personal use only.

Journal of Medical Economics

Volume 17, Number 7

July 2014

A further limitation of the model is that, at its start, a patient was presumed to have no diabetes-related complications. This characteristic is consistent with the entry criteria for the head-to-head clinical trial of vildagliptin compared with sulphonylurea, the study from which the main efficacy and safety parameters of the modeled interventions were taken10. In addition, the assumption is the same for both treatment arms and, therefore, might be expected to impact both arms similarly. Indeed, the exclusion of patients with complications at model entry may represent a conservative approach because the clinical profiles of vildagliptin and sulphonylurea suggest that such complications might disproportionately increase the risk of cardiovascular events in sulphonylurea-treated patients. The model did not incorporate a number of diabetesrelated morbidities, such as peripheral neuropathy and ulceration, which might be expected to generate additional reductions in patients’ quality-of-life, increased rates of hospitalization, and additional treatment costs. However, these were not included as major end-points in the UKPDS and could not easily be incorporated into the outcomes model22. Other complications, although included, were represented using a single state only, an approach which fails to represent disease progression and its consequences for treatment modality, costs, and quality-of-life. It should be noted that, since the design and implementation of this model, the UKPDS Outcomes Model has been re-estimated using follow-up data collected during the post-study monitoring period, providing an additional 10 years of data on risk factors, complications, resource use, and quality-of-life37. This version of the model includes an additional diabetes complication of lower extremity ulcer, additional risk factors such as heart rate, second event equations for MI and stroke, and other enhancements. This new model predicted fewer macrovascular events and higher survival over a 10-year period compared with the previous UKPDS model. Future research should concentrate on using parameters from the updated UKPDS model because it is based on data from a longer follow-up study of patients with T2D, captures more outcomes, and more comprehensively captures the progression of diabetes.

Conclusions For the Portuguese healthcare system, compared with sulphonylureas, the addition of vildagliptin is projected to be cost-effective for the treatment of patients with T2D not controlled with metformin monotherapy. Using our model, treatment with vildagliptin, compared with sulphonylurea, in combination with metformin generated a lower overall incidence of T2D-related adverse events, 506

Cost-effectiveness of vildagliptin in Type 2 diabetes Viriato et al.

resulting in an ICER within the cost-effectiveness threshold for the Portuguese healthcare system setting.

Transparency Declaration of funding Supported by funding from Novartis Pharma AG. The funding organization provided feedback on the design of the model and the inputs/sources used in the model. The funding organization also reviewed the manuscript. Declaration of financial/other relationships D. Viriato and R. Viana are employees of Novartis Farma SA. F. Calado, J.-B. Gruenberger and S. H. Ong are employees of Novartis Pharma AG. S. Johal is an employee of HERON Commercialization – A PAREXEL Company, a consulting company that provides services to several pharmaceutical companies, including Novartis Farma SA and Novartis Pharma AG, and received funding to complete this analysis. D. Carvalho and J. Silva-Nunes have received consulting/speaking fees from Novartis Farma SA. Acknowledgments Nick Rusbridge at HERON Commercialization – A PAREXEL Company provided editorial support in the preparation of this manuscript, funded by Novartis Pharma AG.

References 1. World Health Organization. World Health Organization-Diabetes. Geneva: World Health Organization, 2013. http://www.who.int/mediacentre/factsheets/fs312/en/index.html. Accessed 15 August 2013 2. Gardete Correia L, Boavida JM, Fragoso de Almeida JP, et al. Diabetes: Factors e nu´meros 2012. Relato´rio anual do observato´rio nacional da diabetes. Portugal. Lisboa: Sociedade Portuguesa de Diabetologia, 2013 3. Andayani TM, Ibrahim MIM, Asdie AH. Assessing the impact of complications on the direct medical costs of Type 2 diabetes mellitus outpatients. Int J Curr Pharm Res 2010;2:32-5 4. Menzin J, Korn JR, Cohen J, et al. Relationship between glycemic control and diabetes-related hospital costs in patients with type 1 or type 2 diabetes mellitus. J Manag Care Pharm 2010;16:264-75 5. Oglesby AK, Secnik K, Barron J, et al. The association between diabetes related medical costs and glycemic control: a retrospective analysis. Cost Eff Resour Alloc 2006;4:1 6. Hanas R, John G. 2010 consensus statement on the worldwide standardization of the hemoglobin A1C measurement. Diabetes Care 2010;33:1903-4 7. Inzucchi SE, Bergenstal RM, Buse JB, et al. Management of hyperglycemia in type 2 diabetes: a patient-centered approach: position statement of the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD). Diabetes Care 2012;35:1364-79 8. Blonde L. Current antihyperglycemic treatment guidelines and algorithms for patients with type 2 diabetes mellitus. Am J Med 2010;123(3 Suppl):S12-18 9. Nathan DM, Buse JB, Davidson MB, et al. Medical management of hyperglycemia in type 2 diabetes: a consensus algorithm for the initiation and adjustment of therapy: a consensus statement of the American Diabetes Association and the European Association for the Study of Diabetes. Diabetes Care 2009;32:193-203 10. Ferrannini E, Fonseca V, Zinman B, et al. Fifty-two-week efficacy and safety of vildagliptin vs. glimepiride in patients with type 2 diabetes mellitus

www.informahealthcare.com/jme ! 2014 Informa UK Ltd

Journal of Medical Economics

11. 12.

13.

Journal of Medical Economics Downloaded from informahealthcare.com by Gazi Univ. on 08/20/14 For personal use only.

14.

15.

16.

17. 18. 19.

20.

21.

22.

23.

24.

25. 26.

27.

inadequately controlled on metformin monotherapy. Diabetes Obes Metab 2009;11:157-66 Profit L, Chrisp P, Nadin C. Vildagliptin: the evidence for its place in the treatment of type 2 diabetes mellitus. Core Evid 2008;3:13-30 Dejager S, Razac S, Foley JE, et al. Vildagliptin in drug-naive patients with type 2 diabetes: a 24-week, double-blind, randomized, placebo-controlled, multiple-dose study. Horm Metab Res 2007;39:218-23 Garber AJ, Schweizer A, Baron MA, et al. Vildagliptin in combination with pioglitazone improves glycaemic control in patients with type 2 diabetes failing thiazolidinedione monotherapy: a randomized, placebo-controlled study. Diabetes Obes Metab 2007;9:166-74 Matthews DR, Dejager S, Ahren B, et al. Vildagliptin add-on to metformin produces similar efficacy and reduced hypoglycaemic risk compared with glimepiride, with no weight gain: results from a 2-year study. Diabetes Obes Metab 2010;12:780-9 Scherbaum WA, Schweizer A, Mari A, et al. Efficacy and tolerability of vildagliptin in drug-naive patients with type 2 diabetes and mild hyperglycaemia. Diabetes Obes Metab 2008;10:675-82 Bolli G, Dotta F, Rochotte E, et al. Efficacy and tolerability of vildagliptin vs. pioglitazone when added to metformin: a 24-week, randomized, double-blind study. Diabetes Obes Metab 2008;10:82-90 Rodbard HW, Jellinger PS. A critique of the 2012 ADA/EASD position statement. Diabetologia 2012;55:2850-2 Amiel SA, Dixon T, Mann R, et al. Hypoglycaemia in Type 2 diabetes. Diabet Med 2008;25:245-54 Liu S, Zhao Y, Hempe JM, et al. Economic burden of hypoglycemia in patients with Type 2 diabetes. Expert Rev Pharmacoecon Outcomes Res 2012;12:47-51 Willis WD, Diago-Cabezudo JI, Madec-Hily A, Aslam A. Medical resource use, disturbance of daily life and burden of hypoglycemia in insulin-treated patients with diabetes: results from a European online survey. Expert Rev Pharmacoecon Outcomes Res 2013;13(1):123-30 Calado F, Gruenberger J-B, de Nigris E, et al. Cost-effectiveness of vildagliptin compared to generic sulphonylureas added on to metformin from a Portuguese healthcare system perspective. Poster #1199-P. Philadelphia, PA: 72nd American Diabetes Association Congress, 2012 Clarke PM, Gray AM, Briggs A, et al. A model to estimate the lifetime health outcomes of patients with Type 2 diabetes: the United Kingdom Prospective Diabetes Study (UKPDS) Outcomes Model (UKPDS no. 68). Diabetologia 2004;47:1747-59 Direc¸a˜o-Geral de Sau´de. Norma 052/2011 da Direc¸a˜o-Geral de Sau´de. Abordagem Terapeˆutica Farmacolo´gica na Diabetes tipo 2. 2013 Jul 30. Report No.: 052/2011 Cabrita J, Lopes MM, de Oliveira Martins S, et al. Perfil de prescric¸a¨o e avaliac¸a¨o do grau de cumprimento de orientac¸o¨es terapeˆuticas em diabe´ticos tipo 2: Estudo transversal em medicina geral e familiar. Rev Port Farmacoter 2011;3:7-13 Wallace TM, Matthews DR. Coefficient of failure: a methodology for examining longitudinal beta-cell function in Type 2 diabetes. Diabet Med 2002;19:465-9 Clarke P, Gray A, Holman R. Estimating utility values for health states of type 2 diabetic patients using the EQ-5D (UKPDS 62). Med Decis Making 2002;22:340-9 Manns B, Meltzer D, Taub K, et al. Illustrating the impact of including future costs in economic evaluations: an application to end-stage renal disease care. Health Econ 2003;12:949-58

! 2014 Informa UK Ltd www.informahealthcare.com/jme

Volume 17, Number 7

July 2014

28. Currie CJ, Morgan CL, Poole CD, et al. Multivariate models of health-related utility and the fear of hypoglycaemia in people with diabetes. Curr Med Res Opin 2006;22:1523-34 29. Erhardt W, Bergenheim K, Duprat-Lomon I, et al. Cost effectiveness of saxagliptin and metformin versus sulfonylurea and metformin in the treatment of type 2 diabetes mellitus in Germany: a Cardiff diabetes model analysis. Clin Drug Investig 2012;32:189-202 30. Infarmed. Orientac¸o`es metodolo´gicas para estudos de avaliac¸a¨o econo´mica de medicamentos. Lisbon: Infarmed – Autoridade Nacional do Medicamento e Produtos em Sau´de I.P., 1998 31. Zhang P, Zhang X, Brown JB. Economic impact of diabetes. IDF Diabetes Atlas. 4th ed. Brussels: International Diabetes Federation, 2013 32. International Diabetes Federation. Diabetes Atlas 2012. Brussels: International Diabetes Federation, 2013. http://www.idf.org/diabetes-atlas2012-update-out-now. Accessed September 23, 2013 33. Yazdanpanah Y, Perelman J, Dilorenzo MA, et al. Routine HIV Screening in Portugal: clinical impact and cost-effectiveness. PLoS One 2013;8:e84173 34. Kanavos P, van den Aardweg S, Schurer W. Diabetes expenditure, burden of disease and management in 5 EU countries. London: London School of Economics, 2012 35. Schwarz B, Gouveia M, Chen J, et al. Cost-effectiveness of sitagliptin-based treatment regimens in European patients with type 2 diabetes and haemoglobin A1c above target on metformin monotherapy. Diabetes Obes Metab 2008;10(1 Suppl):43-55 36. Tilden DP, Mariz S, O’Bryan-Tear G, et al. A lifetime modelled economic evaluation comparing pioglitazone and rosiglitazone for the treatment of type 2 diabetes mellitus in the UK. Pharmacoeconomics 2007;25:39-54 37. Hayes AJ, Leal J, Gray AM, et al. UKPDS Outcomes Model 2: a new version of a model to simulate lifetime health outcomes of patients with type 2 diabetes mellitus using data from the 30 year United Kingdom Prospective Diabetes Study: UKPDS 82. Diabetologia 2013;56:1925-33 38. Monami M, Marchionni N, Mannucci E. Long-acting insulin analogues versus NPH human insulin in type 2 diabetes: a meta-analysis. Diabetes Res Clin Pract 2008;81:184-9 39. IMS Health. IMS Health Moving Annual Target (MAT) for Antidiabetics in Portugal from June 2012 to June 2013. Porto Salvo: IMS Health, 2013 40. Infarmed. Autoridade Nacional do Medicamento e Produtos de Sau´de I.P. Lisbon, 2014 41. Ministe´rios da Economia e do Emprego e da Sau´de. Despacho no. 4294-A/ 2013. Ministe´rios da Economia e do Emprego e da Sau´de. Lisbon, 2013 42. Pereira R, Gouveia M, Martins AP. Ana´lise custo-efectividade de sitagliptina quando adicionada e metformina em doentes com diabetes tipo 2 em Portugal. Revista Portuguesa de Diabetes 2012;7:13-23 43. Ministe´rio da Sau´de. Portaria 163. Portaria n. 163. (2013) Dia´rio da Repu´blica. Lisbon, 2013 44. Portugal S, Martins B, Ramos L, et al. Estudo de Avaliac¸a˜o do Impacto dos Custos Financeiros e Sociais da Deficieˆncia. Relato´rio Final. Coimbra: Centro de Estudos Sociais, Instituto Nacional de Reabilitac¸a˜o, Ministe´rio do Trabalho e da Solidariedade Social, 2010 45. Hammer M, Lammert M, Mejias SM, et al. Costs of managing severe hypoglycaemia in three European countries. J Med Econ 2009;12:281-90 46. Clarke P, Gray A, Legood R, et al. The impact of diabetes-related complications on healthcare costs: results from the United Kingdom Prospective Diabetes Study (UKPDS Study No. 65). Diabet Med 2003;20:442-50

Cost-effectiveness of vildagliptin in Type 2 diabetes Viriato et al.

507

Cost-effectiveness of metformin plus vildagliptin compared with metformin plus sulphonylurea for the treatment of patients with type 2 diabetes mellitus: a Portuguese healthcare system perspective.

To evaluate the cost-effectiveness of vildagliptin plus metformin vs generic sulphonylurea plus metformin in patients with type 2 diabetes mellitus, n...
206KB Sizes 0 Downloads 3 Views