Cost-Utility of Triple Versus Dual Inhaler Therapy in Moderate to Severe Asthma

Jefferson Antonio Buendía; Diana Guerrero Patiño


BMC Pulm Med. 2021;21(398) 

In This Article

Materials and Methods

We conducted a probabilistic Markov model to estimate the cost and quality-adjusted life-years (QALYs) of patients with severe asthma treated with dual inhaled therapy and triple inhaled therapy in Colombia. The choice of time horizon was a lifetime, using a cycle length of 2 weeks following the natural history of the disease and previously published asthma economic evaluation models.[10–13] In this mathematical model, patients could transition between four mutually exclusive health states (symptom-free state or asthma-controlled, asthma exacerbation, asthma-related mortality, and all-cause mortality). During each cycle, patients in non-death health states could transit to any of three levels of asthma exacerbations: OCS burst (was defined as relatively major symptoms during the week and need of use of oral corticosteroids to achieve the control of symptoms), emergency department (patient that request treatment with systemic corticosteroids) and hospitalization. Asthma-related mortality following an exacerbation or all-cause mortality could also occur (Figure 1). We did this analysis from a societal perspective (including direct and indirect costs).. Half-cycle correction and an annual discounting rate of 5% were applied to both costs and QALYs, following the recommendations of the Colombian guide for health economic evaluations.[14] Treatment was considered cost-effective if the incremental cost-utility ratio was below $19,000 per QALY gained using the World Health Organization (WHO) recommendation of three times the GDP per capita to define the willingness to pay (WTP) in Colombia.

Figure 1.

Markov model

Parameters of the Markov model

Multiple parameters were derived from published research and local data, which are presented in Table 1. Data of relative risk (RR) on exacerbation rates were extracted from a recent systematic review of dual versus triple therapy in patients with severe asthma.[15] In this study, triple therapy was associated with a reduction in severe exacerbation risk (9 trials [9932 patients]; 22.7% vs 27.4%; RR, 0.85 [95% CI, 0.77 to 0.90]). The transition probabilities for moving between different health states were derived from clinical trials and local studies.[16,17] Data of utilities of each Markov state were extracted from a systematic review of utilities in asthma,[18,19]Table 1. This systematic review identifies 20 studies in asthma that report utilities in different severity states of asthma. Within these four studies (n = 330 patients) showed a median utility of 0.74 ± 0.029 for severe asthma, all estimated using a time trade-off or standard gamble or Asthma symptom utility index in the US and UK population. All these data (RR, transition probabilities, and utilities) were subjected to probabilistic sensitivity analysis as detailed below, and as recommended by Consolidated Health Economic Evaluation Reporting Standards (CHEERS) Statement.[20] In this sensitivity analysis, to build the range of RR to be used in this analysis, we use the CI 95% of RR published by clinical trials.[15] In the case of utilities and transition probabilities, the upper and lower ranges were estimated by adding or subtracting 25% of the value from the central value defined for the base case. The risk of asthma mortality and mortality from other causes was estimated using age- and gender-specific Colombian life tables mortality (2016 to 2020).[17,21] Based on previous studies of drug adherence, for dual and triple therapy, we assumed that 44% and 37% discontinued the treatment after 52 weeks of treatment respectively.[22,23] Sensitivity analysis of percentage of non-adherents and response rates were made by estimating the upper and lower range of each value by adding or subtracting 25% of the value defined previously.

All costs for each health state defined in the Markov model were extracted from a previously published Colombian-based study.[24] Briefly, this study identified the asthma-related direct and indirect costs of 1131 patients with severe asthma from January 1, 2004, through December 31, 2014, in Colombia, Table 1. Asthma severity classification was mainly based on the paper of Jacob et al..[25] Severe persistent asthma required to have more than six Short-Acting Beta-Agonists (SABA) fills per year, and the number of OCS fills per year, was greater than or equal to two or 4 or more exacerbations. Moreover, zero to six SABA fills and three or more SABA fills per year also constitute severe asthma. This criterion related to using rescue medication per year may be more accurate than using LABA + ICS given the high frequency of underuse and prescription of controller medications in Latin American countries.[26] This group of patients with severe asthma had an average of 1.4 ED visits per year, and 2.5 hospitalizations per year; rates that are comparable to those reported in clinical trials and observational studies in patients with severe asthma and tiotropium use.[7,27] Drug's cost and drug's share market of dual (included Budesonide/Formoterol 640/18 mcg daily, Fluticasone/Vilanterol 200/15 mcg daily, Fluticasone/Salmeterol 550/50 mcg daily) and triple therapy (included umeclidinium 50 mcg daily, glycopyrronium 20 mcg daily and tiotropium 5 mcg daily) was taken from the National Drug Price Information System.[28] All cost costs were transformed to 2020 costs using official inflation data in Colombia. We used US dollars (Currency rate: US$1.00 = COP$ 3,500) to express all costs in the study.[21]

Sensitivity Analysis

To explore parameter uncertainty of the model inputs, first, we conducted a deterministic sensitivity analysis using one-way sensitivity analysis with their tornado diagrams, respectively. In this analysis, we univariate evaluated the change in the incremental cost-effectiveness ratio by varying each parameter as described above. Also to explore parameter uncertainty of the model inputs, we conducted a probabilistic sensitivity analysis by randomly sampling from each of the parameter distributions (beta distribution in the case of relative risk and utilities, Dirichlet distribution for multinomial data in the case of transition probabilities, and gamma distribution in the case of costs). The expected costs and expected QALYs for each treatment strategy were calculated using that combination of parameter values in the model. This process was replicated one thousand times (i.e., second-order Monte Carlo simulation) for each treatment option resulting in the expected cost-utility. All analyses were done in Microsoft Excel®.