Assessing the fiscal consequences of novel and existing treatments for triple negative breast cancer in Switzerland by applying a government perspective framework

Abstract

Background

Triple Negative Breast Cancer (TNBC) is an aggressive subtype of breast cancer that can impact patients’ employment and workforce participation. This study estimates how the employment effects of TNBC impact government tax revenue and public benefits expenditure in Switzerland, representing the fiscal burden of disease (FBoD), and likely consequences of introducing new treatment options.

Methods

A four-state cohort model was used to calculate fiscal effects for two treatments: Neoadjuvant pembrolizumab plus chemotherapy followed by adjuvant pembrolizumab monotherapy (P + C→P) and neoadjuvant chemotherapy alone (C). Lifetime present values of tax revenue, social benefit payments, and healthcare costs were calculated for the average population and those undergoing treatment to assess the FBoD.

Results

An average TNBC patient treated with C and P + C→P is expected to generate CHF128,999 and CHF97,008 less tax than the average population, respectively, and require increased social benefit payments. Compared to C, 75% of the incremental healthcare costs of P + C→P are estimated to be offset through tax revenue gains.

Conclusions

This analysis demonstrates that 75% of the additional costs of a new TNBC treatment option can be offset by gains in tax revenue. Fiscal analysis can be a useful tool to complement existing methods for evaluating new treatments.

Keywords:

• Cancer
• economic evaluation
• fiscal burden
• public health
• triple negative breast cancer

JEL CLASSIFICATION CODES:

• I18
• I1
• I
• H51
• H5
• H
• H00
• I00

Introduction

Breast cancer is one of the most commonly diagnosed cancers globally, representing 11.5% of cancer cases diagnosed in 2022 and causing over 666,000 deaths^1,2. There were 7,292 new cases of breast cancer in Switzerland in 2020, with over 1,500 deaths¹, and breast cancer represents the most common cause of cancer-related death in Swiss women³. Triple negative breast cancer (TNBC) is a subtype of breast cancer lacking expression of the estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor-2 (HER2)⁴. TNBC represents 10–15% of all breast cancer diagnoses, with a typically younger age of disease onset than other breast cancer subtypes^5,6 and is also associated with a poorer prognosis and high rates of recurrence^7,8. Historically there have been limited effective therapies for TNBC, due to lack of expression of the various receptor sites as well as the aggressive nature of the disease^4,8. The economic and humanistic burden of TNBC is also substantial^5,9 and includes treatment costs and healthcare resource use (HCRU), which increase as the disease progresses^9,10.

Due to the younger age at diagnosis, TNBC typically affects women when they are active in the workforce⁶, and can reduce patients’ ability to work, leading to productivity and work loss⁹. Following diagnosis, breast cancer patients are likely to stop working and it can take years to return to work (RTW), if they return at all^11,12. A study analyzing RTW in France showed that, among breast cancer patients, only 33% had returned to their work within 6 months and 73.2% within 2 years after diagnosis¹³. A similar study reported median time to RTW, following an early breast cancer diagnosis, of 242 days¹⁴. Even when patients return to work, they are more likely to have sickness absences or require disability payments in the years following diagnosis¹⁵. Patients may even leave the workforce entirely, either through permanent disability or early retirement¹⁶, and these economic effects only worsen with disease recurrence^16–18. The economic effects of TNBC may extend to patients’ spouses and families, who can experience income loss and work absences following diagnosis^19,20. The overall effect of TNBC on employment likely contributes to reductions in household income that can persist for many years²¹.

It is important to note that the economic consequences of TNBC are not limited to the patient, their family or even the healthcare system, but can influence government finances through reduced tax revenue and increased social benefit payments²². When a person is unable to work due to receiving a breast cancer diagnosis and subsequent treatment, they are likely to pay less taxes due to employment disruptions, while simultaneously increasing demand on public benefit programs, such as sick leave paid by the government. As such, the short-term and long-term effects of TNBC extend beyond the healthcare system, and a broader perspective is required to understand the true disease burden. Given the impact of TNBC on government finances, treatments that improve clinical outcomes for patients may also have a quantifiable fiscal effect. For example, treatments which can reduce the proportion of patients experiencing disease recurrence can reduce the fiscal effect associated with TNBC recurrence by, for example, reducing employment interruptions¹⁶. Hence, assessing the efficacy of new treatment options through a broader government perspective allows a wider range of economic benefits to be captured.

Recent advancements in TNBC treatment indicate that improved outcomes and economic gains can be achieved. Notably pembrolizumab, a high-affinity monoclonal antibody which binds to programmed cell death protein 1 was approved by the Food and Drug Administration (FDA) and the European Medicines Agency (EMA) in July 2021 and April 2022, respectively, in combination with chemotherapy as a neoadjuvant treatment and then continued as a single agent adjuvant treatment for high-risk, early-stage TNBC^23,24. This approval was based on results of the pivotal KEYNOTE-522 trial, which showed significantly longer event-free survival (EFS) for patients treated with neoadjuvant pembrolizumab plus chemotherapy, followed by adjuvant pembrolizumab (hereafter referred to as pembrolizumab plus chemotherapy followed by pembrolizumab), than neoadjuvant chemotherapy alone (hereafter referred to as chemotherapy)²⁵. Pembrolizumab plus chemotherapy followed by pembrolizumab was also shown to be cost-effective from the payer perspective in a published analysis²⁶. These benefits may extend beyond the third-party payer to the Swiss government, as benefits in EFS can translate into fiscal benefits by reducing the rates of disease recurrence, as well as the associated fiscal effects of recurrence on employment and social support spending.

The objective of this study is to assess the fiscal burden of disease (BoD), reflecting the net effect on government accounts, of high-risk, early stage TNBC from the perspective of the Swiss government and to capture the fiscal benefits that can be achieved through introduction of pembrolizumab plus chemotherapy followed by pembrolizumab in the early stage as a new therapeutic option. This entails capturing lifetime transactions between citizen and state in terms of taxes paid and benefits received from government from the point of TNBC diagnosis. This modelling framework is similar to cost-of-illness (CoI) studies, however the focus is on estimating the burden from a governmental perspective^22,27. A fiscal analysis model was developed based on a previously published cost-effectiveness analysis²⁶, and includes fiscal effects that fall on government relating to employment, unemployment, disability, retirement, sick leave, as well as corresponding effects on patients’ spouses.

Methodology

The fiscal model framework was assembled using a simulation of TNBC’s natural history based on a previously published Markov state cohort transition cost-effectiveness model, evaluating the economics of adding pembrolizumab to neoadjuvant chemotherapy followed by adjuvant pembrolizumab monotherapy for early-stage, high-risk TNBC in Switzerland²⁶. The modeled population includes TNBC patients with localized tumors that are at high risk of recurrence, and aligns with the population in the KEYNOTE-522 trial^26,28. This backbone simulation model utilizes efficacy, safety, and utility data from the KEYNOTE-522 trial and models the impact of both treatments on outcomes for a cohort of early-stage, high-risk TNBC patients²⁸. Specifically, four health states were included to capture the natural history of this patient cohort: (1) Event-free (EF); (2) Locoregional recurrence (LR); (3) Distant metastasis (DM); and (4) Death. A summary of the key assumptions underlying the cost-effectiveness model are outlined in Table 1, with further details available in the corresponding publication²⁶.

Table 1. Summary of key CE-model assumptions.

Assumption	Value/description	Source
Starting age	49 years of age	KEYNOTE-522^28
Cycle length	1 week	Favre‑Bulle et al.^26
Time horizon	Lifetime horizon (51 years based on starting age)	Favre‑Bulle et al.^26
Intervention	Neoadjuvant pembrolizumab plus chemotherapy followed by adjuvant pembrolizumab	KEYNOTE-522^28
Comparator	Neoadjuvant chemotherapy followed by adjuvant placebo	KEYNOTE-522^28
Data inputs	Clinical efficacy, safety, utility data	KEYNOTE-522^25^,^28

In order to create a fiscal BoD model, the cohort model health states were extended to incorporate fiscal effects. The underlying assumptions and inputs in the published cost-effectiveness model were kept constant, and each health state was assigned different fiscal effects. These effects were then combined with the associated Markov trace within the cost-effectiveness model. This was modeled over a lifetime horizon (i.e. 51 years or until the cohort turned 100 years old) in line with the cost-effectiveness model assumptions²⁶. The comparator arm of the model (chemotherapy) was assumed to approximate the BoD for TNBC, as chemotherapy represents current standard of care for early-stage TNBC patients in Switzerland. The pembrolizumab plus chemotherapy followed by pembrolizumab arm was also analyzed to assess the fiscal benefits that would be realized by the Swiss government if pembrolizumab plus chemotherapy followed by pembrolizumab was available as a treatment option for this patient cohort.

The fiscal effects associated with each health state were identified through a targeted literature review. The objective of this review was to identify data showing the link between TNBC, or breast cancer more broadly, and employment and social support programs (such as disability, sick leave, and early retirement). The search was conducted in both the PubMed and Embase databases, with full details outlined in Supplementary Appendix S1. Following identification of relevant studies in the literature review, population-level fiscal data was sourced to generate the following economic effects for the “average” Swiss population: (a) Economic and employment activity; (b) Unemployment; (c) Sick leave; and (d) Disability. Modifiers identified in the literature were subsequently applied to this data to generate the relative values for each of the cohort model health states. These values were primarily in the form of relative risks (RRs) or hazard ratios (HRs), which were applied directly to Swiss population data. Where data were reported as odds ratios (ORs), these were converted to RRs using the formula from Gidwani and Russel²⁹.

The targeted literature search identified several studies demonstrating the economic impact of breast cancer, with a reduction in employment observed in breast cancer patients following diagnosis^30,31. A number of studies evaluated economic effects specifically among those patients with and without disease recurrence, with greater economic impact observed following disease recurrence^16,17. Where data were available for breast cancer patients with and without recurrence, values were subsequently applied to the EF state and the LR/DM states, respectively. Where this split was not available, a constant effect was applied to all three living health states.

In addition, given that some studies reported an impact of breast cancer on patients’ spouses^19,20, a scenario was included to model a separate fiscal effect for spouses of all living patients. It was conservatively assumed that only spouses of living patients would have a fiscal effect, however it is likely that all spouses would be impacted in practice. It was assumed that 65% of patients would have a spouse based on the results of a study in Italy and Switzerland³². Previous research indicates that having a spouse with breast cancer can lead to reduced income and increased rates of work absences, therefore these effects were captured within the analysis^19,20. A separate scenario also explored the impact of TNBC on rates of retirement, which is shown to be higher among breast cancer patients¹⁷. Table 2 presents an overview of the assumptions and effects applied to each of the three living TNBC health states.

Table 2. Overview of fiscal effects.

Economic effect	Value (type)	Relevant health states	Source
Employment post-diagnosis	0.82 (RR)	EF/LR/DM	Noeres et al.^30
Economic activity (recurrence)	0.61 (HR)	LR/DM	Sieluk et al.^16
Long-term disability (recurrence)	2.48 (HR)	LR/DM	Sieluk et al.^16
Sick leave (EF)	1.91 (OR)	EF	Lundh et al.^15
Sick leave (recurrence)	2.13 (HR)	LR/DM	Sieluk et al.^16

Abbreviations. DF, fisease free; DM, distant metastasis; EF, event free; HR, hazard ratio; LR, locoregional recurrence; OR, odd ratios; RR, relative risk.

The assumptions and inputs applied in the scenario analysis to generate spousal effects in the model are described in Supplementary Appendix S2, while further information on the scenario exploring a retirement effect is presented in Supplementary Appendix S3.

Equations (1–12) outline the key calculations used to combine the fiscal effects with Swiss population-level data to generate the fiscal BoD for each treatment arm and health state. Employment rates were calculated using Swiss labor force statistics from the Organization for Economic Co-Operation and Development (OECD)³³. These values were adjusted to reflect employment effects of patients in the EF, LR, and DM TNBC health states, as described in Eq. (1)(1) $${\mathit{\text{Emp}}}_{i,j}\left(t\right)={(\mathit{\text{EA}}}_{i,j=1}\left(t\right)\mathrm{*}{\mathit{\text{RR}}}_j^{\mathit{\text{EA}}}\mathrm{*}{\mathit{\text{ER}}}_{i,j=1}\left(t\right)\mathrm{*} {\mathit{\text{RR}}}_j^{\mathit{\text{EA}}}\mathrm{*}(1‐{\mathit{\text{SL}}}_j(t\left)\right)$$(1) . Age-specific gross earnings (2023) were subsequently applied to the employed population using data from the Swiss Federal Statistics Offices (FSO)³⁴, as shown in Eqs. (2)(2) $$Y_j\left(t\right)={\mathit{\text{Emp}}}_{i,j}\left(t\right)\mathrm{*} Y_i\mathrm{*}{\mathit{\text{WG}}}_t$$(2) and (3)(3) $$Y_{\mathit{\text{Lifetime}}}=\sum _{j=2}^4{X}_j\mathrm{*}\sum _{t=0}^{\mathit{\text{TH}}}{\mathit{\text{Emp}}}_{i,j}\left(t\right)\mathrm{*}{Y}_i\left(t\right)\mathrm{*}({1+\mathit{\text{WG}})}^t\mathrm{*}({1+\delta )}^{-t}$$(3) . Over the model horizon, wages were assumed to grow by an average of 0.22% annually, based on data from the Swiss FSO, and discounted at a rate of 1.75%³⁵. Several different social benefit (Sb) payments (2023 prices) were included in the model to capture the effect of TNBC on government expenditure. Rates of disability, sick leave, and unemployment were calculated using Swiss data^33,36,37. The prevalence of recipients of each social benefit was calculated by adjusting the population-level data with the disease-specific modifier obtained from the targeted literature review (Eq. 4(4) $${\mathit{\text{Sb}}}_j\left(t\right)={\mathit{\text{Sb}}}_{j=1}\left(t\right)\mathrm{*}{\mathit{\text{RR}}}_j^{\mathit{\text{Sb}}}$$(4) ). The net cost of these programs was calculated using both the prevalence and associated payments (Eqs. 5(5) $${\mathit{\text{SbC}}}_j\left(t\right)={\mathit{\text{Sb}}}_j\left(t\right)\mathrm{*}{\mathit{\text{SbP}}}_j$$(5) and 6(6) $${\mathit{\text{UnEmpC}}}_j\left(t\right)=(1-{\mathit{\text{Emp}}}_j(t\left)\right)\mathrm{*} {\mathit{\text{UnEmpP}}}_j\left(t\right)$$(6) ).

To calculate direct tax revenue from employment earnings, the tax wedge (ratio of taxes paid by the average worker including labor costs for the employer) was applied to gross earnings³⁸ (Eq. 7(7) $${\mathit{\text{DirectTax}}}_j\left(t\right)={(Y}_j\left(t\right))\mathrm{*}\mathit{\text{TaxWedge}}$$(7) ). Indirect tax revenue was also calculated to reflect government income from consumption taxes by applying value added tax (VAT) to both employment income and social benefit payments (as social benefit recipients will spend part of these payments on consumption), as shown in Eq. 8(8) $$\mathit{\text{Indirect}}{\mathit{\text{Tax}}}_j\left(t\right)={(Y}_j\left(t\right)+{\mathit{\text{Dis}}}_j\left(t\right)+{\mathit{\text{SL}}}_j\left(t\right)+{\mathit{\text{UnEmpC}}}_j\left(t\right))\mathrm{*}\mathit{\text{VAT}}$$(8) ³⁹. Once the relevant adjustments were applied to each health state, the total tax revenue and social benefit spending were calculated for the TNBC cohort and general population, using the proportion of patients in each health state from the Markov model (Eqs. 9(9) $${\mathit{\text{TaxRevenue}}}_K=\sum _{j=1}^4{X}_j\mathrm{*}\sum _{t=0}^{\mathit{\text{TH}}}\left({\mathit{\text{DirectTax}}}_j\right(t)+{\mathit{\text{IndirectTax}}}_j\left(t\right))\mathrm{*}{(1+\delta )}^{-t}$$(9) and 10(10) $${\mathit{\text{SbSpending}}}_K=\sum _{j=1}^4{X}_j\mathrm{*}\sum _{t=0}^{\mathit{\text{TH}}}\left({\mathit{\text{RetC}}}_j\right(t)+{D\text{i}s}_j\left(t\right)+{\mathit{\text{SL}}}_j\left(t\right)+{\mathit{\text{UnempC}}}_j\left(t\right))$$(10) ). To calculate the fiscal BoD for TNBC, the fiscal effects for the chemotherapy arm were added to the associated healthcare costs (obtained from the cost-effectiveness analysis) and compared to results for the average population (Eqs. 11(11) $$\Delta \mathit{\text{Fiscal Effect}}=\left({\mathit{\text{TaxRevenue}}}_{j=1}-{\mathit{\text{SbSpending}}}_{j=1}\right)-\left({\mathit{\text{TaxRevenue}}}_{j=2,3,4}-{\mathit{\text{SbSpending}}}_{j=2,3,4}\right)$$(11) and 12). The healthcare costs included in the analysis represent excess healthcare costs associated with TNBC treatment and management, as it is assumed that all other non-TNBC healthcare costs would be equal between the TNBC population and the average population.

The incremental fiscal effect of introducing pembrolizumab plus chemotherapy followed by pembrolizumab was calculated as the difference in the fiscal BoD between both treatment arms.

Equations (1) $${\mathit{\text{Emp}}}_{i,j}\left(t\right)={(\mathit{\text{EA}}}_{i,j=1}\left(t\right)\mathrm{*}{\mathit{\text{RR}}}_j^{\mathit{\text{EA}}}\mathrm{*}{\mathit{\text{ER}}}_{i,j=1}\left(t\right)\mathrm{*} {\mathit{\text{RR}}}_j^{\mathit{\text{EA}}}\mathrm{*}(1‐{\mathit{\text{SL}}}_j(t\left)\right)$$(1) (2) $$Y_j\left(t\right)={\mathit{\text{Emp}}}_{i,j}\left(t\right)\mathrm{*} Y_i\mathrm{*}{\mathit{\text{WG}}}_t$$(2) (3) $$Y_{\mathit{\text{Lifetime}}}=\sum _{j=2}^4{X}_j\mathrm{*}\sum _{t=0}^{\mathit{\text{TH}}}{\mathit{\text{Emp}}}_{i,j}\left(t\right)\mathrm{*}{Y}_i\left(t\right)\mathrm{*}({1+\mathit{\text{WG}})}^t\mathrm{*}({1+\delta )}^{-t}$$(3) (4) $${\mathit{\text{Sb}}}_j\left(t\right)={\mathit{\text{Sb}}}_{j=1}\left(t\right)\mathrm{*}{\mathit{\text{RR}}}_j^{\mathit{\text{Sb}}}$$(4) (5) $${\mathit{\text{SbC}}}_j\left(t\right)={\mathit{\text{Sb}}}_j\left(t\right)\mathrm{*}{\mathit{\text{SbP}}}_j$$(5) (6) $${\mathit{\text{UnEmpC}}}_j\left(t\right)=(1-{\mathit{\text{Emp}}}_j(t\left)\right)\mathrm{*} {\mathit{\text{UnEmpP}}}_j\left(t\right)$$(6) (7) $${\mathit{\text{DirectTax}}}_j\left(t\right)={(Y}_j\left(t\right))\mathrm{*}\mathit{\text{TaxWedge}}$$(7) (8) $$\mathit{\text{Indirect}}{\mathit{\text{Tax}}}_j\left(t\right)={(Y}_j\left(t\right)+{\mathit{\text{Dis}}}_j\left(t\right)+{\mathit{\text{SL}}}_j\left(t\right)+{\mathit{\text{UnEmpC}}}_j\left(t\right))\mathrm{*}\mathit{\text{VAT}}$$(8) (9) $${\mathit{\text{TaxRevenue}}}_K=\sum _{j=1}^4{X}_j\mathrm{*}\sum _{t=0}^{\mathit{\text{TH}}}\left({\mathit{\text{DirectTax}}}_j\right(t)+{\mathit{\text{IndirectTax}}}_j\left(t\right))\mathrm{*}{(1+\delta )}^{-t}$$(9) (10) $${\mathit{\text{SbSpending}}}_K=\sum _{j=1}^4{X}_j\mathrm{*}\sum _{t=0}^{\mathit{\text{TH}}}\left({\mathit{\text{RetC}}}_j\right(t)+{D\text{i}s}_j\left(t\right)+{\mathit{\text{SL}}}_j\left(t\right)+{\mathit{\text{UnempC}}}_j\left(t\right))$$(10) (11) $$\Delta \mathit{\text{Fiscal Effect}}=\left({\mathit{\text{TaxRevenue}}}_{j=1}-{\mathit{\text{SbSpending}}}_{j=1}\right)-\left({\mathit{\text{TaxRevenue}}}_{j=2,3,4}-{\mathit{\text{SbSpending}}}_{j=2,3,4}\right)$$(11) (12) $$\mathrm{\text{Fiscal BoD}}=\Delta \mathit{\text{Effect}}+\mathit{\text{Health care Costs}}$$(12) where, t is the year of the analysis; j is the population group, with j = 1 as the general population, j = 2–4 are the TNBC non-death health states, and j = 5 is the death health state; i is the age group; Emp refers to employment rate; EA is the economic activity rate; ER refers to employment among the economically active (i.e. economically active and not unemployed); Sb is the proportion of population receiving social benefits (i.e. pension, disability payments, sick leave payments, unemployment benefits); SbC is government costs/expenditure for social benefits; SbP is annual government payments per recipient; UnEmpC is the cost of unemployment benefit for the government; UnEmpP is the unemployment benefit per recipient; Y is employment income; W_g is annual wage growth; δ is discount rate; UnEmp is unemployment; Ret is retirement; Dis is disability; SL is sick leave; VAT is value added tax, proxy of indirect tax; TH is the modeled time horizon; K is a specified population (general population or TNBC); X_j is the proportion of the population in health state j (with ${\sum }_{j=2}^5\mathit{\text{Xj}}=1$); and Δ is difference.

Results

Over a 51-year time horizon, an average TNBC patient receiving chemotherapy is expected to generate CHF128,999 less tax revenue for the Swiss government, comprising a reduction of CHF132,491 in tax from employment (both direct and indirect tax) and an increase of CHF3,492 in indirect taxes from social benefits when compared with a typical Swiss individual. An average TNBC patient receiving pembrolizumab plus chemotherapy followed by pembrolizumab in the early stage generated CHF97,008 less tax revenue, consisting of CHF102,758 reduced tax on employment income and an increase of CHF5,750 from taxes paid on income originating from the receipt of social benefits. Based on the results of the cohort modeled, pembrolizumab plus chemotherapy followed by pembrolizumab is associated with healthcare costs of CHF135,143 compared to CHF92,285 for chemotherapy, however 75% of these healthcare costs are offset by the higher level of tax revenue compared to patients receiving chemotherapy alone (CHF31,991). This tax revenue gain is caused by the prolonged EFS and improved survival from pembrolizumab plus chemotherapy followed by pembrolizumab, which delays and reduces the number of patients progressing to advanced disease stages, where there are worse employment outcomes and subsequently reduced tax revenue.

Table 3 outlines the tax revenue for each population in the analysis as well as incremental results.

Table 3. Present value of expected tax revenue and healthcare costs with incremental results (CHF).

Fiscal effect	Average population	P + C→P	Chemotherapy
Tax on income (direct and indirect)	296,570	193,812	164,079
Tax on social benefits	7,745	13,495	11,237
Total tax revenue	304,315	207,306	175,315
Excess healthcare costs	–	135,143	92,285
Incremental Fiscal Effects	P + C→P vs. Average population	Chemotherapy vs. Average population	P + C→P vs. Chemotherapy
Tax on income (direct and indirect)	−102,758	−132,491	29,733
Tax on social benefits	5,750	3,492	2,258
Total tax revenue	−97,008	−128,999	31,991
Excess healthcare costs	135,143	92,285	42,857

Abbreviations. P + C→P, Neoadjuvant pembrolizumab plus chemotherapy followed by adjuvant pembrolizumab.

Government expenditure on social benefits (disability, sick leave, and unemployment) was estimated for both treatment arms and the average Swiss population. These social benefit payments were also calculated based on life years lived to account for the effect of increased survival on social benefit expenditure. Table 4 presents the average social benefit payments per life year, with chemotherapy patients requiring higher payments than those receiving pembrolizumab plus chemotherapy followed by pembrolizumab (i.e. CHF5,765 and CHF5,594, respectively). Both TNBC cohorts require higher payments per life year than the average Swiss individual (CHF1,973).

Table 4. Present value of expected social benefit expenditure (CHF) and expected life years.

Social expenditure type	Average population	P + C→P	Chemotherapy
Disability payments	68,168	41,414	29,483
Sick leave payments	8,154	9,222	7,446
Unemployment payments	24,185	118,987	96,071
Total social benefit payments	100,508	169,622	133,000
Undiscounted life years	50.93	30.32	23.07
Average social benefit payments per life year	1,973	5,594	5,765

Abbreviations. P + C→P, Neoadjuvant pembrolizumab plus chemotherapy followed by adjuvant pembrolizumab.

Scenario analyses were conducted to explore the effects of including additional fiscal effects in the model. The first scenario explored the impact of modeling increased rates of retirement among TNBC patients, while a second scenario considered the economic effect of TNBC on patients’ spouses by modeling income loss and an increase in the likelihood of being absent from work following a spouse’s breast cancer diagnosis. A probabilistic sensitivity analysis was also conducted with 1,000 iterations. This involved varying the key modifiers used to generate relative fiscal effects in each of the TNBC health states, as well as the relevant variables in the underlying Markov model (i.e. healthcare cost and efficacy parameters).

Results of the scenario analyses and probabilistic analysis are outlined in Table 5.

Table 5. Results of sensitivity analyses (CHF).

Scenario analysis	Average population	P + C→P	Chemotherapy
Scenario #1 – inclusion of longevity and retirement effect
Total tax revenue	335,994	225,647	187,951
Effect on transfer costs
Retirement payments	−410,748	−237,043	−159,525
Total social benefit payments	−511,255	−406,665	−292,524
Average social benefit payments per life year	−10,038	−13,411	−12,679
Scenario #2 – inclusion of spouse effect^*
Total tax revenue	304,315	204,972	172,977
Results of Probabilistic Analysis (incremental)	P + C→P vs. Average population	Chemotherapy vs. Average population	P + C→P vs. Chemotherapy
Tax on income (direct and indirect)	−104,552	−132,491	28,267
Tax on social benefits	6,075	3,492	1,529
Total tax revenue	−98,477	−128,999	29,796
Excess healthcare costs	134,935	92,419	42,516

Abbreviations. P + C→P: Neoadjuvant pembrolizumab plus chemotherapy followed by adjuvant pembrolizumab.

*Spouse effect includes income loss and increased rates of work absences, which only affects tax revenue.

In the probabilistic analysis, pembrolizumab plus chemotherapy followed by pembrolizumab was shown to offset 70% of the incremental healthcare costs through increased tax revenue, which is lower than the 75% calculated in the deterministic analysis.

Discussion

This analysis assessed the fiscal burden of TNBC in Switzerland through application of a government analytic framework. This perspective moves beyond the traditional healthcare system view and estimates the broader economic impact from treating TNBC in early stages on tax revenue and government expenditure. Using the chemotherapy arm of the KEYNOTE-522 trial as a proxy for the TNBC BoD, the analysis shows that an average TNBC patient will provide CHF128,999 less tax than the average Swiss individual, and also require CHF3,621 more social benefit payments for every year lived. The results of this analysis show that healthcare costs are not the largest component of the burden of TNBC. The lost tax revenue of CHF128,999 is approximately 40% higher than the excess healthcare costs of CHF92,285, illustrating the importance of considering these broader economic effects. A probabilistic analysis of 1,000 iterations led to results that were broadly aligned with deterministic results but with some variation in the overall fiscal effects. This is likely due to the use of data from multiple observational studies, which may be associated with greater uncertainty than, for example, randomized controlled trials.

As well as establishing the fiscal burden of TNBC, this analysis estimates the broader economic benefits of novel treatment options. Neoadjuvant pembrolizumab plus chemotherapy followed by adjuvant pembrolizumab monotherapy has demonstrated efficacy in high-risk, early stage TNBC patients through prolonging EFS compared to neoadjuvant chemotherapy alone^25,28. A published cost-effectiveness analysis has also demonstrated that pembrolizumab plus chemotherapy followed by pembrolizumab is cost-effective versus chemotherapy alone in Switzerland²⁶. This analysis further illustrates the benefits of introducing pembrolizumab plus chemotherapy followed by pembrolizumab in Switzerland, as the prolonged EFS prevents patients from experiencing disease recurrence, which is associated with significantly worse outcomes related to employment and higher rates of social benefit payments¹⁶. The increased tax revenue resulting from these improved clinical outcomes for patients receiving pembrolizumab plus chemotherapy followed by pembrolizumab offsets approximately 75% of the excess healthcare costs compared to chemotherapy alone.

This analysis focused on permanent employment and economic activity transitions for TNBC patients. However, there are other economic losses which are more transient in nature, such as presenteeism and work absences, that are not captured in our analysis. It is well documented that breast cancer patients experience greater rates of productivity loss, with this increasing as the disease progresses^40–42. While these costs are likely to be substantial, they are mostly borne by employers and were omitted from this study. However, these losses could still lead to costs for the government from lost worker productivity, which could impact firm profitability and subsequently taxes on profit. This would suggest our results likely underestimate the fiscal consequences of TNBC.

There are several limitations associated with this analysis. Firstly, a range of different sources have been used to estimate the relative fiscal effects of TNBC health states. Due to a paucity of data in Switzerland, this data has been sourced from other countries. While there have been many studies, including systematic reviews, which have demonstrated the direct and indirect costs of TNBC and, more broadly, breast cancer, the exact magnitude of these effects varies⁹. Where multiple values for a fiscal effect were available from different publications, median values were included to avoid the impact of extreme results. However, as data from multiple studies has been combined, there is always a possibility that some fiscal effects have been overestimated. Where possible, conservative assumptions were made to mitigate against this issue. For example, fiscal effects relating to disease recurrence were applied equally to both the locoregional and distant recurrence health states, even though it is likely that the effect would be increased in the distant recurrence state. While the studies informing this model are primarily focused on the TNBC population, some data were used from studies within the broader breast cancer population, when TNBC-specific information was unavailable. Additionally, there are some differences across Europe and internationally regarding the criteria used to define TNBC. Therefore, the TNBC cohorts within the included studies may not fully align with the definition of TNBC applied in clinical practice in Switzerland. However, these population differences are not expected to significantly impact results and the uncertainty is addressed within the sensitivity analyses.

The analysis described here is endogenous in that it focuses wholly on the cohort of individuals with TNBC and their financial transactions with government. What is not captured is the exogenous effect that spending on health and survival of TNBC can have. Due to the presence of multiplier effects of spending and human capital gains from improving health outcomes, the results described here are likely an underestimate of fiscal gains. Taking into consideration multipliers, CHF1 in spending will likely translate into > CHF1 in fiscal gain due to the economic interactions brought about through spending and increased survival of those with TNBC. This is not unique to TNBC alone as spending on health by governments offers one of the greatest economic stimulants^43,44.

This study has shown the potential for extending traditional cost-effectiveness analyses into a broader fiscal perspective, however there are important points to highlight when generating fiscal model output. Notably, interpreting results for social benefit payments can appear initially unintuitive. It is important to note that survival will inherently lead to greater social benefit payments, as the population ages and requires pension payments and possibly increased disability benefits. As shown in Table 4, while the TNBC population has decreased social benefit payments over their lifetime compared to the average population, this is driven by the significantly shorter life expectancy (50.93, 30.32, and 23.07 life years for the average population, pembrolizumab plus chemotherapy followed by pembrolizumab arm and chemotherapy arm, respectively). Unemployment payments are higher for both TNBC groups, which is expected given this primarily affects those patients who have not yet retired, which is within the first 16 years of the model (i.e. starting age of 49 years and retirement age of 65 years). Disability payments will increase alongside age and therefore affect the average population to a greater extent given the increased survival. This illustrates the need for awareness regarding the bidirectional relationship of social benefit payments.

Many countries are planning and developing healthcare policies that have a wider focus than the healthcare system alone, including Switzerland⁴⁵. The Health2030 health policy strategy for Switzerland emphasizes the need for financially sustainable healthcare, and that “methods and organisational solutions need to be developed in order to achieve economical and ethically justifiable prices”⁴⁵. Assessing new medicines using a broader fiscal perspective may represent a useful tool to assist decision-making that reflects the full impact of new interventions beyond the health system alone.

Conclusions

This analysis demonstrates the fiscal burden associated with treating TNBC in early stages in Switzerland, using a government analytic perspective. This perspective moves beyond the traditional healthcare system view and demonstrates the effect that TNBC can have on public accounts, through reduced tax revenue and increased government spending. In addition, this analysis explored the fiscal impact from introducing a novel treatment option for early-stage, high-risk TNBC patients. By extending EFS and reducing the proportion of patients experiencing disease recurrence, treatment with pembrolizumab plus chemotherapy followed by pembrolizumab reduces the economic effects associated with later disease stages and can partially offset treatment costs through gains in tax revenue. Therefore, assessing new medicines using a fiscal perspective provides a broader overview of potential cross-sectoral benefits and may facilitate more holistic policy decisions.

Transparency

Declaration of funding

Funding for this article was received from MSD.

Declaration of financial/other relationships

This work has been sponsored by MSD. The authors CC and NK have received funding from the sponsoring organization for their contributions. Authors AFB, GB, DS, and SS are employees of the sponsoring organization.

Author contributions

Conceptualization of the analysis and model design was discussed among all authors. The authors CC and NK conducted a literature review and developed the fiscal calculations. CC and NK developed the first draft of the manuscript which was then reviewed by all authors. All authors also reviewed the final manuscript.

Reviewer disclosures

Peer reviewers on this manuscript have received an honorarium from JME for their review work but have no other relevant financial relationships to disclose.

Ethical approval

As this study utilizes previously published data and does not include use of individual patient-level data, no ethics approval was required.

Consort statement

The study reported here is not a clinical trial or investigational study. The economic modeling work described here uses no individual patient data. The study is based on secondary analysis of previously published sources. No ethics approval was required.

Data availability statement

This study utilizes data from publicly available sources. No proprietary data has been used in modeling work, and all sources have been referenced throughout this manuscript as appropriate. There is no individual patient-level data used in this work.