Determining the value contribution of selexipag for the treatment of pulmonary arterial hypertension (PAH) in Spain using reflective multi-criteria decision analysis (MCDA)

The study was designed according to MCDA methodology [13, 14], using the criteria from the MCDA framework specifically adapted for the appraisal of drugs indicated to treat rare diseases in Spain [15]. Inhaled iloprost was the chosen comparator [1]. A literature review was conducted to obtain relevant information on the disease and its current management in Spain as well as relevant evidence for both compared products. Information was structured into an evidence matrix. This matrix was scored by a broad multidisciplinary panel of Spanish stakeholders involved in healthcare decision-making. Scores were analysed quantitatively. Comments and reflections behind experts’ scores were collected in a qualitative manner.

A literature review was conducted to identify available evidence for selexipag and inhaled iloprost. Available evidence was used to create the MCDA evidence matrix.

Two types of documents were searched: published evidence in biomedical databases and specific product evaluations for selexipag and inhaled iloprost by official healthcare evaluation bodies.

Published evidence was searched in PubMed and MEDES (Medicina en Español) [16] databases in order to answer the following search question: What is the available evidence on epidemiology, health outcomes, unmet needs and economic consequences for the evaluation of drugs indicated for treatment of PAH in Spain? The PICOTS (population, intervention, comparison, outcomes, time span and studies) search strategy was used [17, 18]. The literature review in biomedical databases included published studies from 2007 to 2017.

Specific product evaluations for selexipag and inhaled iloprost were searched in official European and Spanish healthcare evaluation bodies’ webpages [e.g. European Medicines Agency (EMA) [19], Spanish Medicines Agency (AEMPS) [20] and Spanish regional and hospital evaluations [21]]. All evaluations found were included, regardless of the date of publication.

At the time of the study start, the reimbursed price for selexipag had not yet been established in Spain, so a hypothetical price was used. The approved list price [22] was later used in the second phase of the study.

The MCDA framework used was the one adapted and adopted for the appraisal of drugs indicated to treat rare diseases developed by the Catalonian Regional Healthcare Service (Catsalut) [15]. This adapted MCDA framework is based on the EVIDEM framework (version 4.0) [23] and composed of a total of 16 criteria (Table 1). These criteria are structured into two sections: the MCDA Core Model composed of 12 quantitative criteria focused on product evaluation and the MCDA Contextual Tool composed of 4 contextual criteria focused on the consideration of the context surrounding decision-making).

To compare selexipag with inhaled iloprost in the comparative criteria of the MCDA framework, data from randomised control trials from both interventions were used and compared in a descriptive manner.

The study was conducted with a multidisciplinary panel of 28 people involved in the management of PAH treatments and decision-making in Spain, including evaluators, clinicians, regional decision makers, hospital pharmacists from different regions and a patient representative in order to collect insights from a broad range of perspectives.

The study was conducted in two phases: a first face to face meeting with seven experts in the evaluation and management of PAH treatments in Spain, where participants received prior training on reflective MCDA methodology. The experts scored the evidence matrix and the reflections behind the score results were collectively discussed to assess the value contribution of selexipag. This first phase was conducted in November 2016.

The second phase of the study, performed in June 2017, had the objective of validating the results obtained in the first phase and improve the robustness of the results by increasing the sample size. This second part involved 21 participants and was performed using an online platform (Jotform) [24]. Prior to study start, participants received on-line training on reflective MCDA methodology.

Data were collected from each participant, transferred to a common database and analysed with Microsoft Excel. A descriptive analysis of each criterion was conducted. The evaluation of quantitative criteria was performed using a direct rating scale, which varied depending on the type of criteria: from 0 to + 5 for non-comparative criteria and from − 5 to + 5 for comparative criteria. Results of quantitative criteria scoring were given in the form of mean ± standard deviation (SD). The value contribution (VCx) of each quantitative criterion was calculated as the product of its normalised weight (Wx, ∑ Wx = 1) and standardised score (Sx = score/5). Total value estimate (VT) is the sum of all criteria value contributions:

The evaluation of contextual criteria was performed on a qualitative scale with three options (positive, neutral or negative impact) and scores were transformed into a numerical scale (+ 1, 0 and − 1 points respectively).

A two-way ANOVA test of the score means for each criterion obtained from the two study phases was carried out to assess potential differences between scores from both phases of the study. All the quantitative criteria, except for “cost of intervention” (because the price used in both phases of the study was different) were considered in this analysis.

Scores attributed to human error or failure to understand the criteria (e.g.: a higher score assigned to the criteria “comparative cost of the intervention” for an intervention that is more expensive than the comparator) were considered as inconsistencies, excluded from the analysis and replaced using the mean substitution.

The evidence matrix was constructed with a total of 36 identified references (Fig. 1) found in biomedical databases (n = 15), documents from official sources (i.e.: EMA, AEMPS) (n = 2), regional and hospital evaluations (n = 1), clinical guidelines or protocols (n = 2) and online grey literature (n = 16).

The literature review for the first phase was conducted in October 2016 and then updated in May 2017 for the second phase. To calculate the budget impact of selexipag, and in the absence of any public data source, the number of target patients used was based on estimates provided by Actelion Pharmaceuticals Ltd. (personal communication; reference on file).

The scores resulting from both phases of the study for each criterion were analysed separately. There were no significant differences between the score means from both studies for any of the analysed criteria (95% CI p > 0.05). Thus, both samples were combined into a single database and further analysed as shown in Fig. 2. Only statistically significant differences reported across participants’ professional profiles were mentioned.

Selexipag’s “comparative effectiveness” versus iloprost scored 2.3 ± 1.9 on a scale of − 5 to 5. Most participants assigned a positive score to this criterion (19 out of 21). Experts pointed out that the positive score in the “comparative efficacy” criterion for selexipag was due to the robust data obtained in clinical trials compared to placebo (no direct comparative data for selexipag versus iloprost were available). There were statistically significant differences in opinions for this criterion between clinicians and hospital pharmacists (3.3 vs 1, respectively, 95% CI: 0.40 to 4.15; p = 0.01). No differences in terms of “comparative safety/tolerability” were reported, as this criterion scored 0.1 ± 2.1 on a scale of − 5 to 5, due to the similar rate of adverse events reported with selexipag and inhaled iloprost. “Comparative patient reported outcomes (PROs)” scored 2.4 ± 1.8 on a scale of − 5 to 5. Selexipag was perceived as a drug capable of providing value in terms of patient’s quality of life compared to inhaled iloprost due to the convenience of its route of administration and posology: orally for selexipag versus inhaled for iloprost, and the dosing schedule (two times per day vs 6–9 times per day, respectively [25, 26]). Some experts pointed out that the use of two different questionnaires in clinical trials (the CAMPHOR questionnaire [27] for selexipag and the EQ-5D VAS [28] for iloprost) did not allow the comparison of quality of life data for both drugs.

Study participants scored the criterion “preventive benefit” positively, as it scored 3.0 ± 1.1 on a scale of 0 to 5, due to its mechanism of action, with the potential to stabilise the disease, reduce morbidity-mortality events and decrease the need to progress to more inconvenient treatments. The “therapeutic benefit” of selexipag was also scored positively across all participants (3.0 ± 0.8 on a scale of 0 to 5), based on the impact on the delay of the progression of the disease, the therapeutic convenience due to its oral administration and the impact on reducing the number of hospitalisations.. The “comparative cost” of selexipag was analysed only considering the scores from the second phase of the study, when an official approved price for selexipag in Spain was available. “Comparative cost” scored − 1.6 ± 1.6 on a scale of − 5 to 5. This negative score for selexipag was due to the higher price of selexipag vs inhaled iloprost (incremental cost of selexipag 9.8%, ex-factory price per day 128.8 € vs 117.3 €, respectively, according to the mean dosage recommended in the summary of product characteristics of both drugs (2 doses/day for selexipag and 7.5 inhalations/day for iloprost) [25, 26]). Three scoring results from the second phase in which participants scored selexipag’s higher price positively were treated as inconsistences. Experts scored selexipag positively for the “comparative other medical costs” (2.3 ± 2.0 on a scale of − 5 to 5), reflecting their perception of the economic benefits obtained due to savings on healthcare resources resulting from treatment with selexipag compared to use of iloprost. However, hospital pharmacists noted these savings had not been quantified and are not usually considered at time of inclusion of a drug into a hospital formulary. Selexipag is perceived as a therapeutic option that can produce savings in “non-medical costs” compared to iloprost (this criterion scored 2.1 ± 1.9 on a scale of − 5 to 5), because it can improve the daily life of patients with PAH (e.g. better quality of life, reduced cost of carers). Thus, most participants considered that the cost difference between the two products would be acceptable for payers, due to the low prevalence of the disease, the potential cost-offsets derived from the reduction of other medical costs, the perceived improvement of patient’s quality of life and other expected clinical benefits resulting from treatment with selexipag.

The quality of the evidence supporting selexipag was perceived as very good by study participants (this criterion scored 4.0 ± 1.2; on a scale of 0 to 5). Participants recognised the quality of both the design and the analysis of the clinical variables from the GRIPHON pivotal trial. There were statistically significant differences in the scores between clinicians and hospital pharmacists (4.8 vs 3.6 respectively; 95% CI: 0.01 to 2.4; p = 0.04) and between clinicians and healthcare decision makers (4.8 vs 3 respectively; 95% CI: 0.01 to 0.4; p = 0.007). Some experts suggested the combined morbidity-mortality variable should be presented separately in study results, because the main unmet clinical need in PAH remains improvement in survival. Also, the comparison to placebo was considered as a limitation by some participants. The inclusion of selexipag in clinical practice guidelines as a reference for PAH treatment [1] was considered positively (this criterion scored 3.4 ± 1.1, on a scale of 0 to 5) (Table 2).

Figure 3 shows the overall value contribution of selexipag to each criterion for PAH treatment when adjusting the values of the scores obtained during the evaluation to the relative importance assigned to each criterion (weighting).

Assessment of the qualitative criteria from the adapted MCDA framework was positive for all criteria (Fig. 4). Participants scored the qualitative criteria in terms of positive, neutral or negative impact of selexipag for each criterion. Selexipag was perceived as a valuable therapeutic alternative that would have a positive impact in terms of “priority access for population”, according to 71% of study participants (7% perceived it would have a negative impact and 22% a neutral impact), as it would be expected that the classification of PAH as a rare disease and the severity of the disease would favour the incorporation of innovative treatments. However, even though rare disease treatment is a priority for the Spanish national healthcare system, access to new treatments in practice can be slow, according to the opinion of some participants. A total of 64% of experts agreed that selexipag was aligned with the “common goals and specific interests” item of the MCDA framework and that selexipag’s access could be supported by scientific societies, patient associations and other professional groups because it matches the unmet needs identified for the disease. Although participants noted there is multidisciplinary interest from healthcare professionals in the approval of new drugs for diseases with high unmet needs, 36% participants considered there was no impact on this criterion considering the limited prevalence of PAH, which would dilute the specific pressures for inclusion of selexipag into drug formularies. Unlike for other drugs administered by intravenous injection or inhaled, study participants considered treatment with selexipag will not require specific training for patients and healthcare professionals involved in PAH. Thus, the impact on “system capacity and appropriate use of intervention” was considered positively by 75% of participants and with neutral impact for the remaining 25%.

The experts perceived selexipag’s budget impact as negative, due to the higher acquisition cost of selexipag compared to iloprost and the potential positive impact on disease prognosis arising from treatment with selexipag, which will, in turn, increase its prevalence. Although the budget impact was perceived negatively, a total of 64% of participants agreed that the use of selexipag would have a positive impact on the “opportunity costs and affordability” criterion, because of treatment derived clinical benefits and savings in other medical resources (i.e.: less hospital visits, less hospitalisations due to increased efficacy, training from healthcare professionals to patients on the treatment administration). Twenty-five per cent (25%) of participants considered selexipag to have a neutral impact on this criterion and 11% considered selexipag to have a negative impact due to the perception that selexipag would increase the hospital pharmacy budget. In summary, experts believe that treatment of PAH with selexipag would be aligned with the Spanish national healthcare system priorities and the system would be prepared for management of selexipag in real clinical practice.