Ranolazine for the symptomatic treatment of patients with chronic angina pectoris in Greece: a cost-utility study

The model used is a decision tree constructed on Microsoft Excel (Fig. 1), which has two main cohorts of patients representing both treatment alternatives. On the one hand, the ranolazine branch represents the treatment using ranolazine as an add-on therapy for the symptomatic treatment of angina pectoris. On the other hand, the “base treatment” branch represents the standard treatment in the same type of patients. The ranolazine branch is divided into two sub-branches, which shows the possibility of adherence or non-adherence by the patients to the treatment. Patients who do not adhere to the treatment with ranolazine are treated immediately with standard treatment, however for the purpose of the model and following the principle of intention to treat, these patients remain in the ranolazine branch. Whether or not the patient adheres to the treatment, four possible scenarios are considered, related to the frequency of angina that might be experienced at this stage of the model: minimal, mild, moderate or severe frequency. The transition probabilities depend on whether the patient has continued the treatment with ranolazine, whether the treatment has been stopped by the patient, or whether the patient belongs to the cohort of patients with standard treatment. Once the patient is in one of these potential angina frequency groups, hospitalization may or may not be required. When hospitalization is required, then an intervention might be needed or not. Most frequently required interventions at this stage are coronary artery bypass graft (CABG), or percutaneous coronary intervention (PCI). We did not consider mortality in this model related to stable angina pectoris, because, it has been shown in different studies that there are not significant differences in the mortality of patients under evaluation [23, 28].

The clinical inputs considered in the model were: a) transition probabilities, including the probability of experiencing minimum, mild, moderate or severe angina frequency, the probability of being hospitalized, and the probability of being subjected to a revascularization procedure, and b) utility values related to non-hospitalization, hospitalization without revascularization, and hospitalization with revascularization. The above clinical inputs considered in the present analysis were obtained from relevant clinical and epidemiological trials [22, 23, 25‐27]. It is worth noting that the same published studies were used in a previous publication of the model [24].

As mentioned above, the transition probabilities considered in the present model are presented in detail elsewhere [24]. In brief, the probability of adherence to ranolazine treatment was estimated at 72 % [23] while the odds of experiencing at least a minimal frequency of angina (out of minimal, mild, moderate or severe) episode was estimated according to the treatment received [22] (Table 1). At this point it should be noticed that the frequency of angina classification depended on the self-administered questionnaire answers. The questionnaire used was the Seattle angina questionnaire [29]. It led to grouping of patients onto different categories according to their responses. To be more precise, patients with <0.5 episodes/week, 0.5-3 episodes/week, 3–5.5 episodes/week and > =5.5 episodes/week were classified to experience minimal, mild, moderate and severe angina, respectively.

Moreover, the 6-month probability of hospitalization have been estimated at 20, 20, 27 and 30 % for minimal, mild, moderate or severe angina frequencies respectively, based on data obtained from Merlin-TIMI 36 clinical trial [25] (Table 1).

Of note, it has been found that the probability of hospitalization is independent of the type of treatment (ranolazine or SoC) but it depends on the severity of angina. When hospitalization is required, an intervention might be needed (or not) such as CABG or PCI/Stent [26] (Table 1). All these data were reviewed and validated by 3 local experts (C.V., J.P., J.K.) in order to increase the credibility and validity of the model inputs.

QALYs were selected as the measure of effectiveness in present study. Utility weights are a measure of a patient’s preference of a particular health state and generally range from 0 (death) to 1 (perfect health). Utility values were obtained from the literature [27] (Table 1).

Since the analysis was conducted from the public third-party-payer perspective (EOPYY), only health care costs reimbursed by the payer were included in the model. The total reimbursement cost reflected and encapsulated all the possible healthcare resource consumption of patients during the horizon of analysis (6 months). The average resource utilization for a Greek patient suffering from stable angina was extracted from 3 the local experts using a questionnaire developed to serve the purpose of the present study. The questionnaire was developed to reflect the disease management in Greece, and the local experts participating in the present study confirmed its content validity. In this context, the cost of anti-anginal drug acquisition, hospitalizations, vascular interventions and monitoring tests including outpatient visits, laboratory tests and imaging diagnostic examinations were considered in the analysis. All costs reflect the year 2014 (€).

Drug acquisition costs of ranolazine and SoC were calculated using the latest price bulletin issued by the Ministry of Health (26.11.2014) [30], as well as the corresponding reimbursement prices (Positive List for the reimbursement of medicines, FEK 3376/16.12.14). The reimbursement prices were reduced by the patient relevant patient co-payment (25 %) and relative rebates as they suggested by the corresponding legislation (Official Government Gazzete, FEK 64/16.1.2014). In particular, the rebate of 9 % given by manufacturers to get into the positive list was considered in the analysis. For ranolazine only, an additional rebate of 2 % was considered as it is alone in its cluster (Official Government Gazzete, FEK 64/16.1.2014). At this point, it should be pointed out that the volume-related rebate ranging from 2 to 12 % could have been considered in the analysis. Due to lack of volume –related data for all drugs, it was not taken into account in the base case analysis.

Standard care drug costs have been modelled using the overall proportions of patients using each therapeutic class and the mean daily dose for each drug as obtained from the 3 local experts and the relevant drug acquisition cost, calculated as mentioned above. For each therapeutic class, the active substances (INN) considered in the analysis reflected the most commonly prescribed drug locally, as obtained from the 3 local experts. Hence, a weighted cost was calculated for each therapeutic class based on the distribution of patients to different active substances. Based on those mentioned above, the total 6-month acquisition cost for SoC was calculated at 159.01€.

Regarding ranolazine, based on data provided by local experts it was considered that during the first two months 80 % of patients received 750 mg (375 mg twice daily) as a daily dose, while the remaining 20 % received 1000 mg (500 mg twice daily) as a daily dose and from the third month and onwards 50 % of patients received 750 mg (375 mg twice daily) as a daily dose, while 40 % of patients received 1000 mg (500 mg twice daily) as a daily dose and the remaining 10 % received 1500 mg (750 mg twice daily) per day. The total 2-month and 4-month acquisition cost for ranolazine was calculated at 89.27€ and 178.75€ respectively (Additional file 1: Table S1).

To estimate the total 6-month hospitalization cost (excluding hospitalizations related to vascular interventions), the reimbursement tariff per hospitalization obtained from the Diagnostic Related Groups (DRGs) tariffs’ list issued by the Greek Ministry of Health [31] was multiplied by the number of hospitalizations, as provided by the 3 local experts. The hospitalization cost at a 6-month period was found to range from €226.67 in patients experiencing minimal angina episodes to €2451 in patients experiencing severe angina episodes [Table 2].

Regarding the cost of vascular interventions, the proportion of patients with stable chronic angina undergoing revascularization (i.e. PCI or CABG), as extracted from the literature [26] and validated by the 3 local experts, was combined with the reimbursed tariff per revascularization obtained from the relevant DRGs tariffs’ lists issued by the Greek Ministry of Health [31]. The cost per CABG and PCI/stent were found to be €5772 and €1798 respectively. To account for the effect of angina severity to the hospitalization cost, the model allowed the proportion of patients requiring hospitalization (with and without revascularization) and the hospitalization frequency to depend on angina severity.

Routine monitoring costs include outpatient visits, laboratory tests (e.g. blood and biochemical tests) and diagnostic tests. The number of visits, laboratory tests and diagnostic tests (e.g. echo, MRI etc.) required as well as the proportion of patients undergoing each test were retrieved from the 3 local experts. The corresponding reimbursed unit costs were obtained from the Government Gazette (FEK A’262/16-12-2011) and from the official site of EOPYY respectively.

The 6-month cost for outpatient visits was found to be extremely low ranging from €7 in patients suffering from minimal angina episodes to €31.33 in those suffering from severe angina. With respect to laboratory and diagnostic tests, it was found that the 6-month laboratory test cost varies from €7.93 to €40.29 and the 6-month diagnostic tests cost from €40.72 to €123.79 [Table 2].

As with hospitalizations, the model allowed the proportion of patients undergoing to laboratory and diagnostic tests, as well as the number of tests and visits to depend on angina severity.

The aforementioned data were used to get mean estimates of QALY, and total direct costs related to each comparator (ranolazine + SoC vs. SoC alone). The cost-effectiveness of ranolazine plus SoC over SoC alone was evaluated by calculating the incremental cost per QALY gained (ICER). For a treatment to be considered cost-effective a willingness-to-pay (WTP) threshold of €34,000 per QALY was used in the current analysis. This is based on the WHO guidelines stating that a treatment should be considered cost-effective if the ICER is up to 3 times the GDP per capita of that country and a treatment is considered highly cost effective at less than 1 times the GDP per capita [32]. The GDP per capita in Greece was estimated at €17,000 taken from the IMF estimation of GDP per capita using current prices [33].

The majority of input data used in the current model are subjected to variation. Therefore, in order to deal with uncertainty, a probabilistic sensitivity analysis (PSA) was performed using a second-order Monte Carlo simulation. In this analysis, a distribution was assigned around each parameter (i.e. costs, transition probabilities etc.) and the aforementioned economic and health outcomes associated with simultaneously selecting random values from those distributions were generated. Distributions were selected based on the nature of variables [34]. In particular, a gamma distribution was used to represent the uncertainty in costs, because costs are constrained on the interval zero to positive infinity and are often highly skewed. Binomial parameters (i.e. the probability of hospitalization) and utility values are constricted on the interval zero to one and hence they were varied according to beta distribution. For multinomial data such as the probability of experiencing a minimal, mild, moderate or severe angina episode, a Dirichlet distribution was used. Then, 5000 estimates of costs, LY, QALYs, and incremental cost per QALYs were obtained by applying the bootstrapping technique. The mean values of the bootstrapped estimates represent unbiased estimations of the parameters under investigation. The bootstrap percentile method was used to obtain the uncertainty appropriate intervals for each parameter [35]. A cost-effectiveness acceptability curve (CEAC) was plotted, showing the proportion of simulations that are considered cost-effective at different levels of WTP per QALY gained.

Moreover, a one-way sensitivity analysis (OWSA) was undertaken to test the robustness of the results, by varying individual parameters between low and high values, in order to ascertain the key drivers of cost-effectiveness. The upper and lower bounds of all parameters were set at ± 20 % (assumption) of the base case values. The parameters evaluated in this analysis were the cost of ranolazine, the cost of SoC, the cost of hospitalization with no revascularization, the cost of hospitalization with CABG, the cost of hospitalization with PCI/Stent, the probability of patients continuing on ranolazine therapy, the probability of hospitalization with minimal angina, the probabilities of hospitalization, and the utility values. Additionally, one scenario was also considered for the time horizon of the model, assuming that it was 1 year instead of 6 months used in the base case analysis.

All statistical calculations performed using Microsoft Excel 2010.