Cost-Effectiveness of Januvia versus Avandia as Supplementary Treatment in Combination with Metformin for Patients with Type 2 Diabetes
1Ph Pharmerit it
Verheggen BG1, van der Steen A1, Heeg BMS1, Vos CBJ2, van Hout BA1 BV Th N th l d 2M k Sharp & Dohme BV, The Netherlands BV, The Netherlands, Merck Sh D h BV Th N th l d representing diabetes related complications (see poster PMC11). Estimates of disease progression, incidence in the sub models, disease related mortality, and all cause mortality, were derived from the United Kingdom Prospective Diabetes Study. It has been assumed that the drug efficacy and adverse event results of an 18-week head-to head comparison of Januvia versus Avandia supplementary to metformin is sustained during week 19-52. The analysis was conducted from a societal perspective. Direct and indirect costs were included. Effects were reported as (diseaseEff d (di free) life years and quality-adjusted life years (QALYs). Univariate and multivariate sensitivity analyses were carried out. The added uncertainty due to the assumption of sustained effectiveness is expressed by inflating the standard errors used in the multivariate sensitivity analyses by a factor 3. SEs were not inflated, multivariate sensitivity analyses showed that 81% of the iterations showed gains in QALYs and cost savings.
Objective
To assess the cost-effectiveness (CE) of Januvia versus Avandia as supplementary treatment in combination with metformin, for patients with type 2 diabetes in whom metformin (in addition to diet and exercise) does not provide adequate glycemic control. The model and the results were part of the reimbursement dossier of Januvia in the Netherlands. (1)
Conclusion
The base case results indicate that the use of Januvia is cost saving and generates an increase in QALYs per patient. The Dutch Health Insurance Board (CVZ) concluded that the model was valid but that the resulting ICUR was not robust. Based on the latter and the lack of long term safety data the CVZ concluded that the CE of Januvia was not adequately substantiated. q y
Results
Base case results are presented in tables 1 and 2 and in figure 2. The univariate sensitivity analyses revealed that results were most sensitive to changes in HbA1c drop, LDL and SBP change. The acceptability curve showed a CE p y probability of 74% at a willingness to pay of €20,000/QALY. When
Methods
The DELTA model was used to assess the costs and effects of Januvia 100mg in comparison to J i 100 i i Avandia 8mg (figure 1). This cohort model contains five sub models
Figure 1: Treatment comparisons and treatment switches
Strategy 1 2nd line tx 3rd line tx 4th line tx Metformin + Januvia Metformin + LD insulin Metformin + HD insulin Strategy 2 Metformin + Avandia Metformin + LD insulin Metformin + HD insulin
Table 1. Base case results (life time horizon)
Life years Januvia Avandia Incremental 16.615 16.585 0.030 QALY 12.735 12.706 0.029 Costs €40,908 €41,187 -€279
Table 2. incremental results for sub models Figure 2: Scattergram multivariate sensitivity analysis
No complication
€ 10,000 € 8,000 € 6,000 € 4,000 Difference in costs € 2,000 €0 0.000
Incremental QALYs1 0.052 -0 013 0.013 0.001 0.002 0.001 -0.013
Incremental costs2 €69 -€137 €137 -€51 €8 €11 €64
Quadrants: South East: North East: South West: North West:
59% 14% 14% 13%
Macrovascular disease Renal disease Blindness
-0.400
-0.300
-0.200
-0.100
0.100
0.200
0.300
0.400
0.500
0.600
Amputation Heart failure
1Incremental
-€ 2,000 -€ 4,000 -€ 6,000 -€ 8,000 -€ 10,000 Difference in effects
QALY for patients with >1 complication (-0.002) are not included in the table 2Incremental cost for drug treatment (-€262), hypoglycaemia (€0) and other death (€18) are not included g ( ), yp g y ( ) ( ) in the table References: 1) Pharmacoeconomic dossier: Januvia; submitted at Dutch Health Insurance Board, April 2007 2) Claxton K. The irrelevance of inference: a decision-making approach to the stochastic evaluation of health care technologies. J Health Econ. 1999; 18:341-364.21
Comment on NL reimbursement process of Januvia:
According to Bayesian decision analysis theory, decisions have to be made on point estimates and estimates concerning uncertainty should be used to decide on additional research.(2) When following that theory, provisional research (2) theory reimbursement should have been advised, potentially with the condition
of collecting additional data. It is not clear what decision strategy encompasses the decision to reject the pharmacoeconomic analysis of Januvia on the basis of “lack of robustness”. In the mean time, the Dutch ministry of Health, did decide to reimburse Januvia®.
This poster was presented at the ISPOR Congress, Dublin October 2007 This study was financially supported by Merck Sharp & Dohme BV