General approach

The cross-sectional study was conducted in two main phases. In the first phase, participating clinical centres mailed an invitation package to all their patients diagnosed with PAH. The package included a questionnaire, patient information sheet, informed consent form, and pre-addressed/stamped envelopes for document return. One clinical centre opted for on-site recruitment during routine follow-up visits. To minimize selection bias, all PAH patients attending these visits were consecutively invited to participate and received the study materials to complete at home. The questionnaire collected information on sources of indirect costs associated with PAH, particularly productivity losses, as well as healthcare utilization involving general practitioners or specialists outside the treating centres. To limit recall bias, the questions referred to the four weeks preceding questionnaire completion [16] Patients were instructed to return the signed informed consent form to their treating hospital (by mail or during a subsequent visit) and to send the completed questionnaire directly to the study coordination centre (EBPI, University of Zurich). In the second phase, data on healthcare use due to PAH were extracted from the medical records of all patients who returned the questionnaire. This extraction covered the six months prior to study inclusion.

Ethical approval was obtained from the responsible ethics committees on 23/04/2024 (BASEC-Nr. 2023-02352). Patient recruitment started on 26/04/2024 and ended on 10/12/2024.

Study population

The study included adult patients (≥18 years) living in Switzerland with a confirmed diagnosis of PAH (World Health Organization [WHO] group 1) established at least six months prior to study inclusion, who were receiving care at one of six participating clinical centres. The time point of enrolment was defined by the date patients completed the study questionnaire.

Inclusion criteria were diagnosis of PAH confirmed by right-heart catheterization, treatment at a participating PAH centre, provision of informed consent, and ability to complete the questionnaire in German, French, Italian, or English. Included PAH subtypes were: idiopathic, heritable, drug- and toxin-induced, and PAH associated with connective tissue disease, HIV, portal hypertension, congenital heart disease, schistosomiasis, long-term response to calcium channel blockers, and those with features of pulmonary veno-occlusive disease and/or pulmonary capillary haemangiomatosis. Patients were excluded if they had pulmonary hypertension falling into WHO groups 2–5.

Patient survey

The questionnaire used was based on the Productivity Cost Questionnaire (iPCQ), developed by the Institute for Medical Technology Assessment (iMTA) at Erasmus University, Rotterdam [16]. The iPCQ is a standardized tool designed to measure health-related productivity losses and comprises 18 items. Nine of these items gather demographic data and information on respondents’ employment status—such as weekly hours of paid work, working days per week, and productivity loss over the past four weeks.

For the purposes of this study, the questionnaire was adapted to the specific context of PAH and expanded to include additional items. The adaptations concerned presenteeism (i.e., reduced productivity while at work) due to PAH, time lost for household chores, and the provision of formal and informal care by relatives both below and above retirement age.

The final adapted instrument consisted of 17 questions and captured a range of indirect cost domains, including reduced employment levels, absenteeism, presenteeism, reliance on formal care, informal caregiving (by both working-age and retired individuals), and limitations in performing household chores within the preceding four weeks. Additionally, the questionnaire recorded healthcare resource use outside the participating clinical centres, specifically the number of ambulatory visits to general practitioners or specialists within the preceding four weeks. The estimated time to complete the questionnaire was 10 minutes. The iMTA was informed about the questionnaire adaptations. Although the changes did not affect the questionnaire integrity, iMTA can’t guarantee the accuracy or validity of the additionally collected data.

Medical records data

Six clinical centres located in German-speaking and French-speaking Switzerland participated in the study. Clinical data extracted from medical records included patient characteristics (age, sex, WHO functional class [WHO-FC], risk stratification according to the guidelines developed in 2022 by the European Society of Cardiology (ESC) and the European Respiratory Society (2022 ESC/ERS Guidelines), and major co-morbidities) as well as healthcare resource utilization (hospitalizations related to PAH, length of stay, outpatient visits, emergency department visits, diagnostic procedures, and PAH-specific treatments) [7].

All data were recorded using standardized Microsoft Excel templates. Pilot data extractions in two participating centres confirmed the templates to be suitable. Medical record data were pseudonymously linked to the questionnaire responses using unique codes assigned by the enrolling centres. In cases of significant missing or implausible information in the medical record data, the study coordination centre (EBPI) contacted the respective clinical site to verify and clarify the entries. Similarly, when substantial data inconsistencies were identified in patient questionnaires, the enrolling centre was asked to follow up briefly with the patient for clarification.

Participating centres received a reimbursement per enrolled patient for whom complete data were submitted. Patients who completed the questionnaire were compensated with gift card (CHF50, i.e., about €52) as a sign of appreciation.

Costs

The cost calculation was based on the resources used (e.g., PAH-specific treatment, outpatient visits, hospitalizations) and productivity losses (i.e., workdays lost) multiplied with corresponding unit costs and extrapolated to one year.

Hospitalization costs were estimated based on the length of stay in intensive care units (ICU), intermediate care, or general wards due to PAH. Due to the lack of PAH-specific hospitalization costs stratified by type of care, estimates from a report commissioned by the Swiss Confederation—based on data from four major health insurers (Helsana, SWICA, Groupe Mutuel, and CSS) and focusing on COVID-19 hospitalizations—were used as a proxy. According to the report, average daily costs were €1,746 for general wards (range: €1,629–1,862) and €4,188 for ICUs (range: €3,723–4,653) [17]. For intermediate care stays, a midpoint value between general ward and ICU costs was assumed (€2,967 per day).

Outpatient visits at the participating centres were valued at €145.39 per visit, based on the Swiss outpatient medical tariff system (TARMED) [18]. Emergency department visits were assigned a cost of €480 per visit, as estimated by the Swiss Health Observatory [19]. Diagnostic tests and other procedures conducted during patient visits were costed separately. Only the 15 most frequently performed tests and procedures were included in the analysis, with detailed unit cost information provided in the supporting information (S1 Table in S1 File). Diagnostic tests that were rarely performed (e.g., nocturnal oximetry) were not considered in the analyses.

Costs for oxygen therapy (including equipment rental, oxygen refills, maintenance, and consumables) were obtained from the Swiss List of Means and Objects (MiGeL) and drug costs (per tablet or vial) from the Specialty List (Spezialitätenliste) [20,21].

For ambulatory consultations with general practitioners, costs were estimated based on a standard 30-minute consultation using the TARMED system, amounting to €106.43 per visit. This estimate did not include potential additional charges for diagnostic services, preparation of medical reports, or updates to patient records.

Specialist consultations outside the treating centres were assigned an average cost of €157 per visit. This estimate reflects the typical range of specialist consultation fees in Switzerland, which generally fall between €157 and €209 per visit [22,23].

Indirect costs were calculated using the human capital approach, which considers all hours lost due to illness and treatment—by both patients and informal caregivers—as productivity losses [24]. To estimate these costs over a four-week period, the reported times lost were multiplied by the approximate earnings of the individuals, derived from national statistics. For informal care, the proxy good method was applied. This approach assigns a monetary value to unpaid caregiving by equating it to the wage of a professional caregiver or equivalent substitute (e.g., a domestic worker) [25,26]. This method allows for monetization of services for which no direct financial transaction occurs. Specifically, we used age- and gender-specific average monthly wages from the Swiss Federal Statistical Office (SFSO) for the year 2020 [27]. These salaries reflect full-time employment (100% full-time equivalent). To convert monthly salaries to an hourly rate, we divided the figures by 173, in accordance with SFSO standards (based on a 40-hour work week over 4⅓ weeks per month). For employed patients, productivity losses were valued using the average wage of Swiss workers of matching age and gender. For patients not working due to PAH, we applied the same average wage over a four-week period. This value was adjusted based on the average level of employment of the working patients, estimated at 93% of a full-time equivalent. The value of time lost to household chores was based on estimates from a Swiss study, which set this at €28 per hour in 2022. Formal care was valued at €55.08 per hour, corresponding to the hourly cost of professional home care services (Spitex), which provide domestic and support services [28]. More details on unit costs used for the indirect cost calculation are provided in the supplementary materials (S2 Table in S1 File).

To calculate the economic burden at national level, the estimated total number of PAH patient in Switzerland was combined with the estimated total costs per patient. According to recently published estimates, the age-standardized prevalence of PAH in Europe ranges between 22.8 and 54.7 cases per million population. [9,10] Applying a mean of 38.8 cases per million to the Swiss population of 9 million yields an estimated 349 PAH cases nationwide (range: 205–492 cases). The total cost range reported in the results combine the prevalence range with the 95% confidence intervals of the total costs.

For international comparison, cost results are reported in Euro (€), using an average exchange rate of CHF1.00 = €1.047 for the year 2024.

Sample size

No formal sample size calculation was performed, as the study relied on descriptive statistics; however, based on clinical centre estimates, a target sample size of approximately 120 patients was anticipated.

Statistical analysis

The demographic and clinical characteristics of the included population were described using counts and percentages for categorical variables, while continuous variables were described with means and standard deviations (SD). Chi-square tests and Fisher’s exact tests were used for categorical variables. Kruskal-Wallis tests and Mann-Whitney U tests were used for continuous variables. To investigate the association between patient characteristics with total costs, generalized linear model (GLM) regressions with a log link function and assuming a gamma distribution of errors were used. Considering the uncertainty associated with annualizing costs derived from patient-reported outcomes collected over relatively short observation periods (4 weeks or 6 months), sensitivity analyses were conducted. For variables collected over 4 weeks and extrapolated to one year, multiplication factors ranging from 10 to 14 were used, with 13 applied in the main analysis. For variables collected over a 6-month period and extrapolated to one year, the multiplication factor used in the main analysis was varied by ±20%.

Missing data was handled using complete case analysis for each variable, without performing imputation. When participants provided qualitative responses to questions intended to collect quantitative data, these responses were classified as missing. The main database was prepared in Microsoft Excel. Statistical analyses were performed with IBM SPSS Statistics 29.0.2.0.

Study participants

Between April and December 2024, a total of 177 patients diagnosed with PAH were contacted for study participation. Of these, 124 consented to participate and provided their personal data, yielding a participation rate of 70%. Approximately two-thirds of the participants were female, with a mean age of 62.1 years (median: 66 years). Non-responders had a mean age of 58.3 years (p = 0.162 for difference between responders and non-responders) and 59% were female (p = 0.393).

Given the statutory retirement age in Switzerland (65 years, excluding early retirement), over half of the cohort (53.2%) were retired. A total of 32 patients (25.8%) were classified as disabled (i.e., received money from the invalidity insurance), and only 17 (13.7%) were employed or self-employed at the time of the study.

Table 1 presents the distribution of participants including WHO-FC and ESC/ERS 2022 risk stratification, alongside data on comorbidities. The average time since diagnosis was 6.2 years. Patients in lower severity categories tended to have longer times since diagnosis. Specifically, mean ± SD time since PAH diagnosis by WHO-FC was 5.3 ± 4.6 years for Class 1, 7.3 ± 7.6 for Class 2, 5.4 ± 5.3 for Class 3, and 2.6 ± 2.7 for Class 4 (p = 0.284). Similarly, mean time since PAH diagnosis by ESC/ERS risk stratum was 7.1 ± 5.9 years for Stratum I (low risk), 6.0 ± 7.7 for Stratum II (intermediate-low risk), 5.1 ± 3.6 for Stratum III (intermediate-high risk), and 1.8 ± 1.0 for Stratum IV (high risk) (p = 0.167). For four patients, information on WHO-FC nor ESC/ERS risk was not available.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Demographic and clinical characteristics of the study population.

https://doi.org/10.1371/journal.pone.0348190.t001

Healthcare resources used

Table 2 summarizes healthcare resource utilization based on medical record data (covering 6 months) and patient-reported information (covering 4 weeks). Over the 6-month observation period, 23 participants (18.5%) experienced at least one PAH-related hospitalization. Among these, 15 patients were hospitalized once, 4 patients twice, another 4 patients three times, and 1 patient four times.

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Healthcare resource use according to medical record data and patient survey.

https://doi.org/10.1371/journal.pone.0348190.t002

When averaged across the full study sample, the hospitalization rate was 0.29 ± 0.72 admissions per patient. The mean length of hospital stay was 2.55 ± 10.30 days. As shown in Table 2, patients classified in higher WHO-FC or ESC/ERS risk strata were hospitalized more frequently and had longer hospital stays.

The average number of outpatient visits at the participating centres over the 6-month period was 5.13 ± 5.34 per patient. Commonly performed diagnostic procedures during these visits included B-type natriuretic peptide (BNP) or N-terminal pro-BNP (NT-proBNP) testing, 6-minute walk distance (6MWD) assessments, arterial blood gas analysis or pulse oximetry, and pulmonary function tests. The frequency of diagnostic testing tended to increase with worsening WHO-FC or risk category (Table 2).

Longterm oxygen therapy was administered to 45 patients (38%), with the proportion of patients requiring oxygen increasing markedly in WHO-FC 4 and ESC/ERS risk IV.

Regarding pharmacological treatment, the most frequently prescribed medication was macitentan (N = 46; 70%), followed by tadalafil (N = 57; 46%), riociguat (N = 21; 17%), treprostinil (N = 20; 16%), sildenafil (N = 17; 14%), selexipag (N = 14; 11%), and ambrisentan (N = 10; 8%). Other PAH-specific therapies were used in fewer than 7% of patients. The number of PAH drugs per patient ranged from 0 (N = 7, 6%) to 4 (N = 3, 2%). Most patients received at least 2 (N = 63, 51%) or 3 (N = 29, 23%) PAH drugs simultaneously.

In addition to outpatient visits at their treating PAH centres, patients reported an average of 0.60 ± 0.93 visits to a general practitioner and 1.50 ± 2.83 visits to medical specialists over the 4 weeks preceding study enrolment. The mean number of hours of formal care received per patient was 0.6 ± 1.7 hours during the same 4-week period.

Regarding productivity losses, the average number of work hours lost per patient—composed of losses due to absence from work and disability, and averaged across all study participants—was 28.1 ± 57.1 hours over a 4-week period. This corresponds to an estimated loss of approximately five full working days per patient per month. The mean number of hours of informal care averaged 4.7 ± 15.8 hours per patient, of which approximately 70% (3.3 hours) was provided by non-retired caregivers.

More details on the distribution and uncertainty of the collected data (e.g., in terms of median, minimum, maximum, and 95% confidence intervals) are provided in the supplementary materials (S3 Table in S1 File).

Estimated annual costs per patient

Table 3 presents the estimated annual direct and indirect costs per PAH patient, both for the overall study population and stratified by WHO-FC and ESC/ERS risk category (Figs 1 and 2). The mean total cost per patient was estimated at €138,958 per year. Of this, direct costs accounted for approximately 79%, while indirect costs represented the remaining 21%.

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Estimated direct and indirect costs per patient per year (EUR).

https://doi.org/10.1371/journal.pone.0348190.t003

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Estimated total costs of PAH per patient per year (EUR), by WHO functional class.

https://doi.org/10.1371/journal.pone.0348190.g001

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Estimated total costs of PAH per patient per year (EUR), by Risk stratification ESC/ERS.

https://doi.org/10.1371/journal.pone.0348190.g002

Pharmacological treatment emerged as the primary cost component, comprising 65% of the total annual costs and 83% of direct healthcare costs. In contrast, costs related to hospitalizations, diagnostics, and outpatient visits contributed only a small proportion to the overall economic burden. More details on the distribution and uncertainty of the cost estimates (e.g., in terms of median, minimum, maximum, and 95% confidence intervals) are provided in the supplementary materials (S4 Table in S1 File).

When stratified by WHO-FC, average total annual costs increased with disease severity: €81,957 for Class 1, €139,836 for Class 2, €160,366 for Class 3, and €166,569 for Class 4 (p = 0.011). A similar pattern was observed when stratifying by ESC/ERS risk categories, though with some variation: €130,970 for low risk, €131,393 for intermediate-low risk, €127,730 for intermediate-high risk, and €291,728 for high risk (p = 0.083). To assess the robustness of cost estimates across disease severity groups, additional distributional analyses were conducted (see S5 Table in S1 File). Generally, greater variability and wider confidence intervals were observed in higher-risk groups, reflecting small subgroup sizes and the influence of high-cost cases.

Exploratory analyses indicated that treatment with intravenous Treprostinil was a major cost driver. Patients receiving Treprostinil (n = 20) incurred substantially higher costs compared to those not receiving it. The mean annual direct costs for Treprostinil-treated patients were €350,967, compared to €62,604 for those not on this therapy. Similarly, indirect costs were markedly higher in this group, averaging €54,370 versus €25,128. Among the patients who did not receive Treprostinil, the costs increased with disease severity: €62,312 for Class 1, €81,476 for Class 2, €104,547 for Class 3, and €166,570 for Class 4 (p = 0.006). A different pattern was observed when stratifying by ESC/ERS risk categories: €87,377 for low risk, €84,444 for intermediate-low risk, €82,794 for intermediate-high risk, and €209,166 for high risk (p = 0.046).

Another exploratory subgroup analysis was conducted among patients with connective tissue disease (CTD)–associated PAH to compare systemic sclerosis (SSc) and non-SSc cases. Of the 37 patients with CTD-associated PAH, 19 had SSc. Patients with SSc were more frequently female (84.2% vs. 61.1%) and older (mean age 69.8 vs. 62.4 years) compared with non-SSc patients. Mean annual direct and indirect costs were higher in non-SSc patients (€139,080 and €44,153, respectively) than in those with SSc (€64,981 and €25,985, respectively).

In a further analysis, patients were stratified according to the time since PAH diagnosis. If compared to prevalent cases (i.e., those diagnosed more than 1 year age), incident cases generally had higher costs related to diagnostic and outpatient visits to PAH centres, but considerably lower costs related to PAH treatment (see S6 Table in S1 File).

In the first GLM regression with total costs as dependent variable, males, ESC/ERS Risk Stratum IV and disease duration were significantly associated with increased costs. On the other hand, females, age, and being employed/self-employed were associated with decreasing costs (see S7 Table in S1 File). The second GLM aimed to investigate which factors among WHO-FC, ESC/ERS 2022 risk strata, and PAH clinical subgroup classification had the highest impact on total costs (see S8 Table in S1 File). Treatment with treprostinil was the strongest determinant of total costs (coefficient 1.432, p < 0.001). Higher WHO-FC was also significantly associated with increased costs, with the largest effect observed for WHO-FC 4 (coefficient 0.970, p = 0.035), followed by WHO-FC 3 (coefficient 0.745, p < 0.001) and WHO-FC 2 (coefficient 0.379, p = 0.022). In contrast, PAH clinical subtype and ESC/ERS 2022 risk strata were not significantly associated with total healthcare costs in the multivariable model. It is important to emphasize that the regression analysis was conducted in a relatively small sample, which may limit statistical power and the stability of model estimates. In addition, several covariates included in the model, particularly WHO-FC and ESC/ERS risk strata, partially overlap conceptually and clinically, as both reflect disease severity. WHO-FC and ESC/ERS risk strata were in fact significantly correlated (correlation coefficient 0.575, p < 0.001). This collinearity may have attenuated independent associations and should be considered when interpreting the relative contribution of individual predictors to healthcare costs.

In the sensitivity analysis, in which outcomes collected over 4 weeks and 6 months were annualized using lower and higher multiplication factors, estimated direct costs ranged from €87,114 to €130,227, indirect costs from €22,957 to €32,140, and total costs from €110,071 to €162,367. In general, the relative distribution of the costs components remained stable. Varying the multiplication factor for outcomes collected over a 4-week period had little impact on cost estimates. In contrast, varying the annualization of variables collected over a 6-month period (particularly treatment costs) had a moderate impact on the estimated costs (see S9 Table in S1 File). Graphical representations of additional subgroup analyses (e.g., by gender, age group, age since PAH diagnosis, PAH subtype, number of co-morbidities, and number of prescribed PAH drugs) are provided in the supporting information (Figs S1 to S15 in S1 File).

Estimated national burden

Assuming the study sample is representative of the broader PAH population in Switzerland with respect to healthcare utilization and productivity losses, the total annual economic burden of PAH is projected at €48.5 million (range: €23.2–81.0 million). Of this, €37.0 million (range: €17.7–64.8 million) would be attributable to direct costs, and €12.5 million (range: €4.4–18.9 million) to indirect costs.

Strengths and limitations

The study has several strengths, including the use of detailed, patient-level clinical and cost data from multiple Swiss centres, and the adoption of a societal perspective in cost estimation. Indirect costs were assessed through a validated, simple, and short questionnaire. Although the survey was not pilot tested, the participation rate was high and there were very few missing or implausible values, implying that the questionnaire was well accepted and understood by the study participants. However, several limitations should be noted. First, although we could include a substantial proportion of the Swiss PAH population, the absolute number of included PAH patients remained small due to the rarity of the condition, implying a risk of residual chance effects. Second, considerable variability was observed across multiple variables (e.g., length of stay, number of diagnostic tests, costs). This heterogeneity mainly reflected the limited sample size and substantial inter-patient differences in health care utilization. Third, to calculate the national burden of PAH, we applied published prevalence estimates ranging from 22.8 to 54.7 PAH cases per million in the absence of Swiss-specific data. According to clinical experts, the real number of PAH patients in Switzerland likely approximates the upper bound of this range. In the future, data from the Swiss pulmonary hypertension registry may enable a more accurate estimation. Fourth, some cost variables were based on outcomes collected over relatively short observation periods. Patient-reported outcomes were collected over 4 weeks, while other outcomes were extracted from medical records covering a 6-month period. Extrapolating these variables to one year by simple multiplication implies the assumption that the impact of PAH on patient health and resource use remains constant over time. However, seasonal effects, which could lead, for example, to more frequent exacerbations, cannot be excluded. The sensitivity analysis conducted to assess the uncertainty related to the annualization of costs showed that varying the multiplication factor for outcomes collected over 4 weeks (10–14 instead of 13 used in the main analysis) had a minimal impact on direct costs and only a small impact on indirect costs. For variables collected over a 6-month period, the multiplication factor used in the main analysis was varied by ±20% in the sensitivity analysis; this resulted in a moderate impact on cost estimates, particularly for treatment costs. Fifth, we cannot exclude non-response bias, also given that non-responders had a lower mean age and were more often male than responders (both age and gender were not statistically different between responders and non-responders). This may have led to non-representative results for healthcare resource consumption and productivity losses. Sixth, we are not able to verify self-reported loss of time and healthcare use outside the enrolling hospital: although we adhered to the timeframe specified by the iPCQ, some patients may have encountered difficulties recalling the exact amount of time lost or the number of visits in the last 4 weeks. Seventh, this study focused exclusively on medications indicated for PAH and did not account for treatments related to concomitant diseases. Given that most participants were polymorbid, some healthcare resource use and productivity losses may have been attributable to comorbidities rather than PAH itself. Consequently, the reported costs may both underestimate the true economic burden—by excluding therapies required for associated conditions (e.g., steroids or immunosuppressants in patients with connective tissue disease)—and overestimate PAH-specific costs by capturing resource utilization driven by coexisting diseases. Eighth, when patient recruitment started, three principles pathways for treatment of PAH were available (endothelin, nitric oxide, prostacyclin). Newer therapies such as activin-signalling inhibitors were not yet approved/available and their impact is therefore not considered in this article. Ninth, among the participating PAH patients, only one individual received epoprostenol in combination with macitentan, prescribed in the context of a late PAH diagnosis during pregnancy, where it represented the most appropriate therapeutic option. Epoprostenol is not included in the list of medications covered by mandatory basic health insurance in Switzerland, and no official Swiss public price is available. One option would have been to directly contact the pharmaceutical company to obtain the applicable cost. However, given that this concerns a single patient in our cohort, doing so could raise ethical concerns regarding patient rights. Therefore, for this patient, only macitentan costs were included in the analysis, while potential epoprostenol costs were omitted. Given that this concerns a single case, we expect this omission to have a limited impact on the overall cost estimates. Finally, cross-sectional data cannot capture longitudinal cost dynamics or treatment trajectories.

Despite these limitations, the findings underscore the high financial impact of PAH in Switzerland and highlight the importance of optimizing disease management and resource allocation for this rare, high-cost condition.