Evaluating Payment Models for ATMPs: A Comparative Study of Different Payment Models for Hemgenix® Shalini Padhy Institute of Medicine Master’s in Public Health Science School of Public Health and Community Medicine Sahlgrenska Academy, University of Gothenburg Gothenburg 2025, 15 May ABSTRACT Introduction Advanced Therapy Medicinal Products (ATMPs), like gene therapy Hemgenix® for Hemophilia B, offer curative potential for rare diseases with unmet needs. Despite their clinical promise, high upfront costs and uncertain long-term efficacy pose significant challenges to their uptake through traditional healthcare reimbursement models. Aim This study aims to compare the costs incurred by different payment models for the one-time high-cost ATMP, Hemgenix®, and the budget impact of these payment models on healthcare payers in Sweden. Methods A decision support framework was applied to determine the suitable alternative payment methods for Hemgenix®. Followed by a cost assessment using a Markov model simulated over 5,10, and 15 years for a hypothetical cohort of Hemophilia B patients across three scenarios of clinical response (according to the Hope B trial, full, and partial response). Three payment models: full upfront payment, 30% reduced upfront payment, and outcome-based annuity payment, were compared alongside standard prophylaxis. Results In the short term, Hemgenix® was more expensive than prophylaxis, but over time, the reduced upfront payment and outcome-based models resulted to be less expensive as compared to prophylaxis. At 15 years, outcome-based annuity models proved most affordable under partial response scenarios, while the reduced upfront payment model was cheaper under full response. Sensitivity analyses showed that with increasing clinical uncertainty, the cost incurred by outcome-based models became lower compared to other Hemgenix® payment models. Conclusions: While the reduced upfront payment models with a considerable percentage of concession are a financially viable option, outcome-based annuity models effectively mitigate financial risk and budget impact under clinical uncertainty. Tailored reimbursement strategies can ensure early and sustainable access to ATMPs like Hemgenix® and reduce long-term burden of rare diseases like Hemophilia B. CONTENT 1 INTRODUCTION ................................................................................................................................... 1 1.1 Background ................................................................................................................................... 1 1.2 Hemophilia B: Clinical Background and Treatment ..................................................................... 1 1.3 Gene Therapy for Hemophilia B: Development and Efficacy of Hemgenix® ............................. 2 1.4 Theoretical Framework ................................................................................................................. 4 1.5 The Economic and Reimbursement Challenge: Need for Alternative Payment Models .............. 5 2 AIM ..................................................................................................................................................... 8 2.1 Research Questions ....................................................................................................................... 8 3 METHODS ........................................................................................................................................... 9 3.1 Study Design ................................................................................................................................. 9 3.2 Data Collection.............................................................................................................................. 9 3.3 Decision Support Framework........................................................................................................ 9 3.4 Outcome Scenarios for Comparison of Payment Models ........................................................... 10 3.5 Markov model ............................................................................................................................. 10 3.6 Markov model parameters ........................................................................................................... 12 3.7 Deterministic Sensitivity Analysis .............................................................................................. 14 3.8 Ethical Consideration .................................................................................................................. 14 4 RESULTS ........................................................................................................................................... 15 4.1 Assessment of Cost of Treatment, Hemgenix® .......................................................................... 15 4.2 Effect of Discount Rate on Payment Results .............................................................................. 17 4.3 Deterministic Sensitivity Analysis .............................................................................................. 18 5 DISCUSSION ...................................................................................................................................... 21 5.1 Study strengths ............................................................................................................................ 23 5.2 Study Limitations ........................................................................................................................ 23 6 CONCLUSION .................................................................................................................................... 24 7 PUBLIC HEALTH PERSPECTIVES/IMPLICATIONS ................................................................................ 25 8 DECLARATION OF AI AND AI-ASSISTED TECHNOLOGIES ................................................................. 26 ACKNOWLEDGEMENT ............................................................................................................................ 27 REFERENCES .......................................................................................................................................... 28 APPENDICES ........................................................................................................................................... 33 ABBREVIATIONS AAV Adeno-Associated Virus ABR Annualized Bleed Rate AMP Alternative Payment Models ATMP Advanced Therapy Medicinal Products CED Coverage with Evidence Development CFC Clotting Factor Concentrates EHL Extended Half Life EMA European Medicines Agency FIX Factor IX GC Genomic Copies HTA Health Technology Assessment ICER Institute for Clinical and Economic Review MEA Managed Entry Agreement Nabs Neutralizing Antibodies QALY Quality-Adjusted Life Years SEK Swedish Krona SHL Standard Half Life TLV Tandvårds- och Läkemedelsförmånsverket VAS Visual Analogue Scale Shalini Padhy 1 INTRODUCTION 1.1 BACKGROUND Advanced Therapy Medicinal Products (ATMPs) represent a promising new class of biopharmaceuticals that have the potential to cure severe and otherwise difficult-to-treat diseases and revolutionize medicine.(1) The European Medicines Agency (EMA) broadly classifies ATMPs into gene therapy, somatic cell therapy, and tissue-engineered products.(2) Unlike traditional medicine, these are generally offered as a single administration, resulting in durable, potentially curative outcomes. They offer significant possibilities towards the treatment of various severe, chronic, and previously untreatable diseases such as Alzheimer's, Parkinson's, cancer, muscular dystrophy, and inherited genetic disorders.(1) Globally, over 3,900 gene therapy clinical trials have either been completed, are ongoing, or have received regulatory approval.(3) Several ATMPs have been approved by the EMA, including Luxturna (for Leber congenital amaurosis), Zolgensma (for spinal muscular atrophy), and Casgevy ( for Sickle Cell Anemia).(4) In 2023, EMA granted conditional market approval to Hemgenix® for the treatment of Hemophilia B, which is a rare, chronically debilitating disease. (5) Considering the clinical profile, lifelong economic and societal burden, the therapeutic promise, and recent approval of Hemgenix®, this thesis focuses on Hemophilia B to assess the costs incurred to the healthcare payers using different payment models for ATMPs. 1.2 HEMOPHILIA B: CLINICAL BACKGROUND AND TREATMENT Hemophilia is a rare, congenital bleeding disorder caused by a deficiency in clotting factors, Factor VIII in Hemophilia A and Factor IX (FIX) in Hemophilia B. Hemophilia B follows an X-linked recessive inheritance pattern, predominantly affecting males. Although females are usually asymptomatic carriers, some may experience mild symptoms.(6) According to the World Federation of Hemophilia (WFH) (6), over 1.12 million males are affected by Hemophilia worldwide, approximately 0.418 million of whom are suffering from severe Hemophilia. (7) They estimate that Hemophilia A accounts for 80 – 85% of all Hemophilia cases, with Hemophilia B constituting the remaining 15 – 20%. P R E V A L E N C E * P R E V A L E N C E * P R E V A L E N C E * A T B I R T H ( A L L S E V E R I T I E S ) ( S E V E R E C A S E S ) ( A L L S E V E R I T I E S ) Hemophilia A Hemophilia B Figure 1. Prevalence of Hemophilia A and B by severity and at birth (*per 100,000 males). Data source: (6) Figure 1 shows the estimated global prevalence at birth of 24.6 per 100,000 males for Hemophilia A and 5.0 for Hemophilia B.(6,7) In Sweden, there were approximately 209 cases of Hemophilia B in 2020.(8) Prolonged bleeding episodes are the characteristic clinical feature of Hemophilia. The clinical diagnosis of Hemophilia should be suspected in individuals with a history of the following symptoms: tendency to 1 17,1 3,8 6 1,1 24,6 5 Evaluating Payment Models for ATMPs: A Comparative Study of Different Payment Models and Reimbursement Approaches for Hemgenix® bruise easily, unexpected or unexplained bleeding, especially in joints, muscles, or soft tissues, and/or Prolonged or heavy bleeding after injuries or surgical procedures.(6) The disease severity generally correlates with the level of clotting factor IX. Severe cases (<1% of normal) lead to spontaneous bleeding; moderate cases (1–5%) cause prolonged bleeding after minor trauma or occasional spontaneous bleeding; and mild cases (5–40%) typically result in excessive bleeding only after trauma or surgery.(6,9) Hemophilia B treatment relies on adequate replacement of factor IX (FIX) to ensure blood clotting and prevent complications.(6,10) The standard of care involves regular intravenous infusions of clotting factor concentrates (CFCs), either plasma-derived or recombinant, aimed at maintaining FIX above 1%.(10) While prophylaxis has significantly reduced bleeding episodes, arthropathy, and improved quality of life, it remains burdensome due to its high cost and frequent dosing requirements.(10,11) There are two main types of prophylactic treatment for Hemophilia B.(6) The Standard Half-Life (SHL) CFCs, which require frequent infusions (2–3 times per week) for maintaining recommended FIX levels due to their short half-life. This frequent dosing often leads to adherence challenges.(6) In contrast, Extended Half-Life (EHL) CFCs offer improved pharmacokinetics with 3–5 times longer half-life. This allows infusion intervals of 7–14 days in some patients while still maintaining adequate FIX levels.(6,12) Despite these advancements, Hemophilia B continues to impose a substantial clinical and economic burden due to its high cost and need for long-term adherence. These factors make it particularly challenging from the perspective of healthcare budget allocations and resource management. Therefore, there is an unmet medical need for new therapeutic approaches that are economically sustainable as well as favorable in relieving patients from the burden of frequent infusions. 1.3 GENE THERAPY FOR HEMOPHILIA B: DEVELOPMENT AND EFFICACY OF HEMGENIX® Gene therapy offers a promising alternative to prophylactic factor replacement therapy for Hemophilia B by introducing a functional copy of the deficient gene into the patient’s liver cells.(13) The goal of treatment is to achieve sustained endogenous FIX production at levels high enough to eliminate regular prophylactic infusions.(13) 1.3.1 DEVELOPMENT OF HEMGENIX® Adeno-associated virus (AAV) vectors are specially engineered viruses used as a vehicle to deliver gene therapy into patients' cells. (14) During early research of Hemophilia B, two main types of vectors were used, AAV2 and AAV8. These vectors were chosen due to their ability to enter liver cells easily, as well as not trigger a strong immune response in the human body.(15–17) the viruses are modified to be harmless while maintaining their ability to deliver genetic material efficiently into liver cells.(10) However, a key limitation of AAV-mediated gene therapy is the body´s immune response, which forms antibodies against these vectors, anti-AAV neutralizing antibodies (Nabs).(13,18) These Nabs can be both pre-existing and treatment-induced. The studies find that many people may already have these antibodies due to previous exposure to naturally occurring AAVs, which block these vectors, preventing them from delivering gene therapy effectively. Furthermore, the patients receiving gene therapy can develop antibodies over time, thus ruling out the possibility of repeat dosing.(13,18) Etranacogene dezaparvovec, the active ingredient in Hemgenix®, is the first approved gene therapy for adults with moderate to severe Hemophilia B. (19) It uses an AAV5 vector to deliver a modified version of the Factor IX gene (called Factor IX Padua) to the liver cells. (13) AAV5 was chosen because it has 2 Shalini Padhy a low prevalence of pre-existing antibodies in the population and reduced chances of immune reactions, improving the likelihood of durable gene expression and longer therapeutic benefits.(20,21) Hemgenix® is administered as a single intravenous infusion and has shown sustained FIX production in the liver. Clinical trials demonstrated that it significantly reduced annualized bleeding rates and allowed most patients to stop regular prophylactic treatment.(19) It was clinically validated in the HOPE-B trial. (22) 1.3.2 EFFICACY AND VALIDATION OF HEMGENIX® Coppens et al. (22) reported the post-hoc efficacy of the HOPE-B (Health Outcomes with Padua Gene; Evaluation in Hemophilia B) trial – a Phase 3, open-label, single-arm, multicenter study evaluating the efficacy and safety of Hemgenix® in adults with moderate to severe Hemophilia B.(22) It is the first large-scale gene therapy trial conducted in this patient population. In the study, a total of 54 adult males (≥18 years) with severe or moderately severe (FIX 0 – 2%) Hemophilia B were enrolled. All participants had a history of severe bleeding and were receiving stable prophylactic treatment with Factor IX replacement therapy. Each subject received a single intravenous infusion of Hemgenix® at a dose of 2×10¹³ genome copies per kilogram of body weight.(22) The primary goal in the ongoing HOPE B Trial, as reported in April 2022, is to compare the efficacy of annualized bleeding rate (ABR) during the post-treatment period (months 7–24) with the lead-in period (≥ 6 months) of continuous prophylaxis before gene therapy.(22) The clinical trial observed statistically significant reductions in bleeding events following treatment with Hemgenix®.(22) The Mean ABR decreased from 4.18 during the lead-in period to 1.51 post-treatment. Furthermore, the Adjusted ABR specifically for FIX-treated bleeding episodes dropped from 3.65 to 0.99. Additionally, post-treatment, 67% of participants had zero bleeding episodes compared to 26% during prophylaxis. The secondary outcome from the trial also highlighted sustained endogenous production of Factor IX due to Hemgenix® therapy.(22) Mean FIX activity remained stable at 36.7% for up to 24 months, approaching the normal FIX activity threshold (> 40%). It was also reported that approximately 33% of participants reached FIX levels within the non-hemophiliac range (>40%), and 96.29% of patients discontinued routine prophylaxis, significantly reducing disease burden. 1.3.3 IMMUNOGENICITY AND SAFETY The trial included participants with pre-existing AAV5 neutralizing antibodies (NAbs). Those with low titers (≤678) achieved FIX expression comparable to NAb-negative individuals. (22) This suggests that gene therapy may be effective even in patients previously considered ineligible due to their immune status. Safety outcomes were favorable, with minimal and manageable adverse events and no serious treatment-related complications reported over 24 months.(22) This supports a favorable risk–benefit profile for Hemgenix®. However, several challenges persist. The therapy's high upfront cost places considerable strain on healthcare systems. Additionally, the presence of NAbs against AAV capsids can interfere with vector transduction, potentially limiting the treatment’s effectiveness. The humoral immune response also prevents repeat dosing, which may be problematic if therapeutic levels decline over time. Lastly, while early results are promising, long-term data on the safety and durability of Hemgenix® remains limited, which highlights the importance of continued follow-up and further studies.(23,24) 3 Evaluating Payment Models for ATMPs: A Comparative Study of Different Payment Models and Reimbursement Approaches for Hemgenix® 1.4 THEORETICAL FRAMEWORK Despite their potential, ATMPs unique characteristics complicate health technology assessment (HTA) and pose significant challenges to healthcare regulation, pricing, and payment/reimbursement models.(25) Detela et al. (26) summarize the regulatory framework led by the European Commission and implemented by the EMA, which ensures clinical efficacy and safety for all human medicinal products, including ATMPs. Through the Marketing Authorization Application process, the EMA, together with national regulatory agencies, reviews clinical data rigorously, maintaining high standards before granting market approval. Most ATMPs follow a centralized authorization pathway, where the EMA does benefit–risk assessment without comparing alternatives or evaluating cost-effectiveness. Conditional approvals may be granted with limited data if early access is justified, with follow-up data required post-approval. (26) In Sweden, the Tandvårds- och Läkemedelsförmånsverket (TLV) (24) evaluates subsidy eligibility, based on ethical principles and a value-based pricing model that accounts for disease severity and the cost per health gain. TLV also supports the New Therapy Council with health economic assessments, informing regional decisions. While national recommendations influence the access and pricing of ATMPs, individual regions retain the authority to decide local uptake.(24) Clinical trial data alone is often insufficient for informed decision-making, therefore, TLV considers modelling analyses the standard approach to extrapolate costs and effects over time.(24) Decision- analytic models are widely used in economic evaluations, especially when assessing healthcare resource allocation.(27) Among them, the Markov models are crucial for capturing multiple possible outcomes.(28) Drummond et al. (27) describe a Markov model to represent patient health as a set of distinct "states" at specific points in time. Time progresses in cycles, which typically range from one month to one year, depending on the nature of the disease and interventions, during which patients may transition between states or remain in the same one. Selecting an appropriate cycle length is essential to avoid multiple clinical events occurring within a single cycle. Likewise, the time horizon must reflect the period over which meaningful differences in cost and outcomes are expected.(27) Health outcomes are commonly measured in quality-adjusted life years (QALYs), a standardized way to capture both quantitative health improvements (e.g., reductions in morbidity) and qualitative benefits (e.g., mortality reduction) into a single metric for comprehensive health measurement.(24,27) Health- related quality of life is scored on a scale from 0 (death) to 1 (perfect health).(27) In clinical trials, Quality of life is often derived using standardized instruments such as the EuroQoL, Health Utility Index Mark 2, or Quality of Well-Being Scale.(29) EuroQoL consists of two components: EQ-5D-5L and Visual Analogue Scale (VAS).(29) The EQ-5D index value assesses health across five dimensions: mobility, self-care, usual activities, pain/discomfort, and anxiety/depression. Each dimension is rated across five levels, ranging from slight to extreme problems, generating a score between 0 (death) and 1 (perfect health).(30) In contrast, the VAS is a self- rated health scale where respondents score their current health from 0, indicating the worst imaginable health, to 100, representing the best imaginable health.(30) TLV suggests three methods to reflect uncertainties for ATMPs like Hemgenix® with limited long-term clinical data: scenario analysis, deterministic sensitivity analysis, and probabilistic sensitivity analysis. (24) One-way sensitivity analysis, a form of deterministic analysis, examines the effect of varying a single parameter while holding others constant at baseline values.(27) Drummond et al. (27) highlight its utility in model development and validation, but caution that it may underestimate overall uncertainty. Market authorization by EMA and health economic evaluations by national HTA authorities are the first steps towards the commercialization process of ATMPs, as they undergo tedious negotiations with individual national authorities regarding payment and reimbursement. (31) Several ATMPs like ChondroCelect®, Skysona®, and Zynteglo® were withdrawn from the market by the manufacturers due to a lack of reimbursement, highlighting the need for managed entry agreements and 4 Shalini Padhy innovative/alternative payment methods to confirm early patient access. (32,33) In the following section, we will discuss alternative payment methods and their core features. 1.5 THE ECONOMIC AND REIMBURSEMENT CHALLENGE: NEED FOR ALTERNATIVE PAYMENT MODELS While ATMPs like Hemgenix® offer potentially curative, single administration treatments, they pose significant challenges to traditional reimbursement systems, which are structured for periodically administered chronic therapies. Their one-time, high-cost nature, combined with clinical and long-term efficacy uncertainties, makes them challenging to evaluate, fund, and adopt into existing payment frameworks. Two core economic challenges emerge in the reimbursement of ATMPs. (24) The first is the Irreversibility Problem. Payment is typically made in full at the time of treatment, even though the clinical benefits may take years to materialize. If the therapy underperforms, payers bear the full cost with no recourse, creating significant financial risk for healthcare systems and insurance companies.(24) The second is the Budget Impact Problem. While ATMPs may be cost-effective over a patient's lifetime, the large upfront cost can overwhelm annual healthcare budgets. This concentrated spending may limit access and divert funds from other essential services, particularly in resource-constrained systems.(24) Table 1. Dimensions for payment models based on their design Dimension Model Description Reimbursement Finance-based Focus on pricing arrangements without linking payment to treatment outcomes. Outcome-based Tie reimbursement to the observed clinical effectiveness, sharing the risk with manufacturers. Payment Timing Upfront payment Full or partial payment made at the time of treatment. Delayed payment Payments are spread over time or triggered by clinical milestones. Payment Level Patient-level Payment is based on treatment success or outcomes at the individual patient level. (Sub)population-level Payment structures apply to all eligible patients within a healthcare system. Source:(34) To address these challenges, various alternative payment therapies (APMs) have been developed and formalized through Managed Entry Agreements (MEAs) between payers and manufacturers.(35) These models are broadly categorized in Table 1 based on their design. 1.5.1 TYPES OF PAYMENT MODELS Several APMs have been developed to address the clinical uncertainty and financial challenges posed by ATMPs. These models are grouped into three broad categories: financial payment models, outcome- based payment models, and market-based payment models.(24) Financial payment models Financial payment models focus primarily on price-based strategies. One approach is a Percentage discount/ Confidential rebate, where the manufacturer offers an unconditional confidential discount on the list price, either upfront or as a rebate. (23,24) Among positives, the model reduces decision uncertainty and overall treatment cost. Furthermore, it is easy to implement, as discounts apply uniformly across patients, irrespective of the treatment. Among its challenges, it has limited use for ATMPs with uncertain efficacy, unless discounts are substantial. They are commonly used in Sweden for hospital and benefit–covered pharmaceuticals. Another finance-based strategy is Price-Volume Agreements. In this model, the unit price of a treatment decreases based on decided tiers once predefined volume thresholds are met. The model can improve affordability and help prioritize access for patients who benefit most. 5 Evaluating Payment Models for ATMPs: A Comparative Study of Different Payment Models and Reimbursement Approaches for Hemgenix® However, it is challenging to implement for rare diseases and does not directly address clinical uncertainty. Additionally, steep price drops at high volumes can reduce market competition by favoring a single product. This model requires adequate infrastructure and is most cost-effective when applied on a large scale. (23,24) The third model is Fixed Annual Sum, which functions as a subscription. A flat fee is paid annually for treating all eligible patients.(24) This model improves access to patients who may have little benefit from the treatment and reduces the risk of exceeding healthcare budgets. That said, if fewer patients than expected use the treatment, the average cost per patient increases. Due to the high manufacturing cost of ATMPs, this model may be unattractive for manufacturers. Maximum cost per patient per year is a financial model that establishes an annual expenditure ceiling per patient. Any additional cost beyond the ceiling is absorbed by the manufacturer.(23,24) It is well-suited for treatments administered by weight or over long periods.(24) The model is less applicable for single-administration ATMPs due to the absence of prolonged dosing or variable costs. The fifth financial model is Indication-based pricing. In this model, the price of a pharmaceutical product varies based on the condition it treats, allowing higher prices for indications with greater health benefits, introducing fairer pricing structures.(24) However, they are not beneficial for specialized treatments like ATMPs. Lastly, coverage with evidence development provides reimbursement only for patients participating in formal studies that evaluate the treatment’s outcomes. It differs from outcome-based payment models as they mitigate the clinical or budget uncertainties of future patients, without affecting current patients.(23,24) For ATMPs, this method may have limited applicability due to high upfront costs and lasting impacts. Outcome-based (Performance-based) Models These models link payments to real-world effectiveness. One such model is Payment by result, where payments are contingent on achieving predefined clinical outcomes (e.g., biomarker levels), either at the individual or population level.(23,24) It aligns the cost of the treatment with its real-world effectiveness, thus encouraging early access to specialized treatments.(24) The model supports early access while managing financial risk, but requires robust infrastructure, outcome tracking systems, and long-term monitoring to be effective.(24) They can be instrumental in alleviating the cost incurred due to clinical uncertainties in ATMPs. Another outcome-based model is Conditional Treatment Continuation.(23,24) This model links payment to patient response, where full payment is made only for patients achieving pre-defined clinical endpoints. In case of partial response or failure, the treatment is either free or discounted. The model encourages close monitoring and early discontinuation of ineffective treatments but needs predefined milestones and diligent follow-up. It may help in cost containment of ATMPs, considering their uncertain clinical efficacy. Market-Based Model This model adjusts pricing based on market dynamics or future alternatives, such as the entry of new or more cost-effective alternatives.(24) It supports access even when better options may become available soon and can prevent overpayment. However, it needs constant market monitoring and complex payment contracts. It is a promising approach for reimbursement of fast-evolving areas like ATMPs, but is limited by practical challenges. (24) 1.5.2 KNOWLEDGE GAP AND POTENTIAL PAYMENT MODELS FOR HEMGENIX® IN SWEDEN Hemgenix® clearly illustrates the importance of Alternate Payment Methods. Its high upfront cost, coupled with limited long-term data due to small clinical trial sizes and reliance on surrogate endpoints, 6 Shalini Padhy introduces substantial uncertainty. Traditional reimbursement models through a one-time payment place payers at considerable financial risk if real-world outcomes fall short of expectations. The reimbursement landscape for ATMPs requires fundamental innovation. APMs offer flexible and adaptable solutions to address the unique risks and rewards posed by therapies like Hemgenix®. Studies by Callenbach et al. (34,36) have highlighted the importance of aligning payment structures with clinical value to facilitate access, ensure economic sustainability, and support continued research investments. An outcomes-based annuity model is especially relevant in this context. It allows costs to be spread over time and is linked to actual health improvements, ensuring broader access to Hemgenix® while minimizing the upfront budget strain, incentivizing the manufacturer to ensure continued benefit. Despite the potential, there is a critical knowledge gap in understanding how different APMs, particularly in the Swedish healthcare setting, can address the clinical and financial uncertainties associated with one-time, high-cost ATMPs like Hemgenix®. Current literature provides limited insight into how these models compare in terms of effectiveness, risk-sharing, and long-term affordability for payers. This thesis aims to address that gap by systematically comparing alternative reimbursement and payment models and evaluating their implications for Swedish healthcare payers. By doing so, it will explore what types of APMs should be considered, how they reduce clinical uncertainty, and their potential to improve affordability and access for advanced therapies like Hemgenix®. 7 Evaluating Payment Models for ATMPs: A Comparative Study of Different Payment Models and Reimbursement Approaches for Hemgenix® 2 AIM To compare the costs incurred by different payment models for one-time high-cost ATMPs like Hemgenix® and the budget impact of these payment models on healthcare payers in Sweden. 2.1 RESEARCH QUESTIONS 1. Which alternative payment models are most financially suitable for healthcare payers while reimbursing the one-time high-cost ATMP, Hemgenix®? 2. How does clinical uncertainty affect the cost incurred by the healthcare payers for Hemgenix® treatment, and how can different payment models mitigate it? 3. How do the alternative payment models impact the affordability of ATMPs? 8 Shalini Padhy 3 METHODS 3.1 STUDY DESIGN An assessment of the costs associated with Hemgenix® treatment was conducted from a healthcare payer perspective, considering different payment models and long-term assumptions to assess their budget impact. The methodology involved identifying suitable APMs using the Decision Support Framework by Callenbach et. al. (34). This framework, designed to guide the selection of suitable models for high- cost, potentially curative one-time therapies, was initially used in this study to subjectively assess the most appropriate APM. Because these therapies are new in the market, factors like clinical uncertainty, the length of follow-up studies, therapeutic efficacy, and financial burden impact the choice of payment models. To capture long-term uncertainty around the clinical efficacy of Hemgenix®, three distinct outcome scenarios were considered: base case according to the HOPE B trial, full response, and partial response, each representing different assumptions regarding the treatment success. These scenarios were subsequently incorporated into a Markov model, specifically developed for Hemophilia B, to compare the long-term costs of Hemgenix® treatment with prophylaxis, under different payment models. Additionally, a one-way deterministic sensitivity analysis was performed to assess the impact of varying failure rates and concession levels on the treatment costs. These analyses determine threshold values at which cost or effectiveness changes would alter the preferred payment model, thereby facilitating optimal reimbursement decisions.(27) 3.2 DATA COLLECTION We begin with data collection, emphasizing the importance of a methodical strategy for identifying and incorporating relevant data.(28) Correspondingly, a systematic approach to collecting data is applied through published scientific articles, clinical trials, HTA assessment reports, and public databases of Sweden. The literature review for relevant data was conducted using PubMed, Google Search, Scopus, and the Cochrane Library. The search strategy included various combinations of keywords, such as hemophilia B, gene therapy, Hemgenix®, ATMP, cost-effectiveness, the cost of Hemgenix®, Etranacogene dezaparvovec, MEA, reimbursement models, APM, and outcome-based agreements, among others. 3.3 DECISION SUPPORT FRAMEWORK To determine suitable payment methods for Hemgenix®, considering the clinical and economic advantages, a decision support framework described by Callenbach et al.(34,36) is partly applied. The framework consists of four parts, and in our study, we adapted the first part for matching disease and treatment characteristics with reimbursement/payment modes. Instead of using the calculation tool developed by Callenbach et al. (34), a Markov model was constructed to compare various payment and reimbursement models. Figure 2 shows the decision support framework adopted in this study. 9 Evaluating Payment Models for ATMPs: A Comparative Study of Different Payment Models and Reimbursement Approaches for Hemgenix® Figure 2. Decision Support Framework adapted from Callenbach et. al. (36) On aligning the characteristics of the disease and the treatment with payment models following the decision framework steps (see Appendix A1), an outcome-based model was subjectively found to be the most relevant payment method for Hemgenix®. Therefore, this study compared upfront, reduced upfront payment, and outcome-based annuity models, further analysed through a Markov model. 3.4 OUTCOME SCENARIOS FOR COMPARISON OF PAYMENT MODELS To account for the uncertainty surrounding the long-term effectiveness of Hemgenix®, the associated costs of the payment models were calculated using the three clinical response scenarios. In the study, in the base case scenario, all patients were assumed to respond according to the outcomes reported in the HOPE-B trial. In the full response scenario, all patients were assumed to respond positively to Hemgenix®, eliminating bleeding episodes and the need for prophylaxis. Finally, in the partial response scenario, a 10% annual failure rate of Hemgenix® was assumed. 3.5 MARKOV MODEL A simplified Markov model was developed to determine the long-term costs and health benefits of Hemgenix® and to compare them using different payment models and their assumed effectiveness. The model was designed for a hypothetical cohort diagnosed with Hemophilia B and includes three health states – Hemgenix®, Prophylaxis, and Death (see Figure 3). All the patients treated with Hemgenix® can transition into 3 states: remain prophylaxis-free, transition to prophylaxis, or death. Patients who transition to Prophylaxis require lifelong treatment and cannot be re-treated with Hemgenix®. Similarly, patients who transitioned to Prophylaxis remain under its treatment or may transition to Death. This structure helps capture key treatment directions while supporting comparisons between different reimbursement models. Figure 3. Markov Model (Hemgenix®, Prophylaxis, Death) 10 Shalini Padhy The comparator consists of two state models – Prophylaxis and Death (see Figure 4). Figure 4. Comparator Markov Model (Prophylaxis, Death) Due to limited data on bleeding-related transition probabilities for the relatively new Hemgenix® treatment, the model was adapted to account for these data limitations. The simplified structure is well- suited for evaluating payment models, as it allows for patient transitions back to prophylactic treatment post-Hemgenix® administration—an important factor influencing both cost and reimbursement outcomes. 3.5.1 STUDY POPULATION FOR THE MODEL A hypothetical cohort consisting of one hundred patients diagnosed with Hemophilia B was considered. The cohort at the time of administration of Hemgenix® was assumed to be 45 years old with an average body weight of 70 kg. Treatment success with Hemgenix® was defined as achieving Factor IX activity levels ≥ 2%. Patients below this level post-treatment are assumed to require weekly prophylaxis.(37) The transition probabilities for the failure of Hemgenix® treatment remained the same for the entire 15-year duration of the model. All patients (both under Hemgenix® and Prophylaxis treatment) experienced the same age-specific mortality risk (calculated from mortality rates in Sweden).(38) A hypothetical population with characteristics outlined in Table 2 is chosen. Table 2. Hypothetical population characteristics chosen for the model Category Details Sex Male Age 45 years old Weight 70 kg (average) Diagnosis Diagnosed with Hemophilia B and currently receiving prophylactic treatment Disease Severity Moderate to severe Hemophilia B Inhibitor History No present or past factor IX inhibitors Immune Profile No neutralizing anti-AAV5 antibodies above Source:(39) 3.5.2 PAYMENT METHODS IN MODEL In the model, three payment methods for Hemgenix® were compared with each other and against the cost of continued prophylactic treatment. These methods included: (1) a full upfront payment model, (2) a reduced upfront payment model, and (3) an outcome-based annuity payment model. The reduced upfront payment model, representing the percentage discount model, is a type of financial payment model. In our study, we assumed an upfront payment with a 30% concession. As the actual concession rates are confidential, a conservative estimate of 30% was applied in the model. The outcome-based annuity model was motivated by the approach proposed by the Swedish Institute for Health Economics (40) for the one-time treatment of Beta-thalassemia. This model includes a 5-year payment plan, with the first one-fifth of the cost paid upfront, and the remaining four-fifths distributed over four years, conditional on treatment success. (40) In our study, the outcome-based annuity payment model was designed by dividing the cost into equal annual installments with payment conditional on the observed treatment efficacy of Hemgenix. 11 Evaluating Payment Models for ATMPs: A Comparative Study of Different Payment Models and Reimbursement Approaches for Hemgenix® 3.5.3 TIME HORIZON Literature suggests that the duration of a payment agreement should be long enough to enable reliable clinical evaluation and thorough data collection. However, it should not extend so far that the agreements become challenging to implement or enforce. Consequently, a five-year time frame is often recommended as a practical upper limit.(1, 17) However, given that Hemophilia B is an incurable, lifelong condition, and Hemgenix® treatment claims to offer long-term therapeutic benefits, a longer time frame can be considered, especially in the context of outcome-based annuity payment models. In 2024, Denmark approved a 10-year outcome-based annuity payment for Hemgenix®, and the first patient has been treated according to the payment model.(42) Given that Hemgenix® is a relatively new drug requiring long-term evidence generation, three time- horizons: 5, 10, and 15 years were considered in the model, with annual cycle lengths, to compare the different payment models in terms of cost. 3.6 MARKOV MODEL PARAMETERS The model parameters are summarized below in the Table 3, which are described in this section Table 3. The model parameters for the Markov model Model Parameters Model Sub-category Base Value Sub Parameters (years) Transition Probabilities Risk of failure of Hemgenix® 0.019 (0 – 1) Risk of Death 45 – 49 0.00145 50 – 54 0.0022 55 – 59 0.00396 60 – 64 0.00669 Cost Hemgenix®– full cost 29,251,075 (SEK) Prophylaxis 2,209,868 per year Hemgenix® cost – 30% concession 20,475,753 Hemgenix® concession for upfront payment 30% Outcome-based payment – Hemgenix® 5 4,875,179 cost/year 10 2,65189 15 1,828,192 Discount Rate (%) C ost 3 Source: Risk of Hemgenix® failure (22), Risk of death (38), Cost of prophylaxis & Hemgenix® (37), Discount Rate (43) 3.6.1 TRANSITION PROBABILITIES FOR HEMGENIX® In a Markov model, a transition probability represents the likelihood of moving from one health state to another during a given cycle.(28) To simulate the clinical efficacy of Hemgenix®, clinical trial data were used to estimate success rates and transition probabilities. A post hoc analysis by Coppens et al. (22) of the HOPE-B (phase 3) trial after 24 months, showed that 2 patients out of 54 required FIX replacement therapy despite receiving Hemgenix®. The causes included hypersensitivity reactions and limited expression of endogenous FIX due to antibodies. Based on the HOPE-B trial, an efficacy rate of 96.29% (52/54) was considered for 24 months (2 years), thus, the transition probability of shifting from Hemgenix® to Prophylaxis was 3.71% or 0.0371.(22) 𝑝 = 0.0371 (1) One-year transition probability (12 months) is calculated using the formulation by Gidwani and Russell (44), explaining the relation between rate (r), time (t), and probability (p) in the equations: 12 Shalini Padhy −𝑙𝑛 (1 − 𝑝) (2) 𝑟 = 𝑡 𝑝 = 1 − 𝑒−𝑟𝑡 (3) Substituting Eq.(1) in Eq. (2) we evaluate the monthly rate, −𝑙𝑛 (1 − 0.0371) 𝑟 = 24 𝑟 = 0.0016 (4) Substituting Eq. (4) in Eq. (3) we evaluate the transition probability, 𝑝 (12) = 1 − 𝑒(−0.0017∗12) 𝑝 (12) = 0.019 (5) The base transition probability of treatment failure of Hemgenix® (shifting from Hemgenix® to prophylaxis) for an annual cycle is calculated as 0.019. Similarly, for the full response scenario, where no patients experience treatment failure, the transition probability was assumed to be 0.00. For the partial response scenario, with a 10% failure rate, the annual transition probability was calculated as 0.1. These values were used to describe different levels of Hemgenix® efficacy across scenarios in the Markov model. 3.6.2 RISK OF DEATH The all-cause mortality risk was calculated based on the published Swedish Statistics database (SCB) data.(38) Age-specific mortality rate per 1000 men for the year 2023 was utilized to calculate the all- cause death rate for a cohort of 45-year-old men for 15 years in total. To cater for the age progression of the cohort, the statistics shown in Table 4 for different age groups was used to determine the annual probability of death, applied equally to all patients undergoing Hemgenix® and prophylactic treatments. Table 4. Age-related mortality rates calculated from the Swedish Statistics Database Age Group Mortality Rate Transition Probability (Years) (per 1,000 men) (All-Cause Death) 45 – 49 1.45 0.00145 50 – 54 2.20 0.00220 55 – 59 3.96 0.00396 60 – 64 6.69 0.00669 Source:(38) 3.6.3 COST OF HEMGENIX® AND PROPHYLAXIS This study considers only the manufacturer's price of Hemgenix® as reported by TLV (37) in 2023, which is SEK 29,251,075 per patient. Costs related to drug administration, screening tests, and wound management have been excluded from this study. For the comparator arm, Alprolix® is used as the standard prophylactic treatment. It is assumed that the average dose is 52 IU / kg body weight/week. Based on an average patient weight of 70 kg, in the chosen hypothetical cohort, the cost corresponds to SEK 2,209,868 per patient per year.(37) Two alternative payment models for Hemgenix® were considered: the reduced upfront payment model and the outcome-based annuity model. In the first model, a 30% concession resulted in a reduced cost per patient of SEK 20,475,753. The latter model divided the total cost into equal annual installments, payable contingent on sustained treatment efficacy over the selected time horizons. In all cases, a downpayment equivalent to one annual installment is made up-front for the entire cohort. Specifically, in a 5-year time horizon, the total cost was divided into 6 equal installments of SEK 4,875,179. For the 10 years, the cost was divided into eleven installments of SEK 2,659,189 per patient annually. For the 15-year horizon, sixteen installments with an annual payment amounting to SEK 1,828,192 per patient 13 Evaluating Payment Models for ATMPs: A Comparative Study of Different Payment Models and Reimbursement Approaches for Hemgenix® were taken. After the initial payment, subsequent annual installments were made only for patients who remained free from prophylactic treatment. 3.6.4 DISCOUNTING As per the recommendations by the Pharmaceutical Benefits Board (43) and per Drummond et al. (27) a 3% discount rate is required to be applied to the costs to reflect the time preference for immediate benefits over future gains. We consider the costs incurred both with and without the application of discounting. 3.7 DETERMINISTIC SENSITIVITY ANALYSIS Based on the Markov model, it is crucial to explore how uncertainty in key parameters influences cost outcomes across payment models. Correspondingly, one-way sensitivity analyses were conducted with their respective parameters summarized in Table 5. Firstly, to explore the impact of this uncertainty, a one-way sensitivity analysis was performed by varying the annual risk of failure of Hemgenix® from 0% to 10%, in 1% intervals. The total cost of Hemgenix® was recalculated under each of the three payment models (upfront, reduced upfront payment, and outcome-based annuity) for 5-year, 10-year, and 15-year time horizons, while keeping all other parameters constant. Secondly, for the reduced upfront payment model, a one-way sensitivity analysis was conducted on the base scenario by varying the concession rate applied to Hemgenix® from 5% to 30%, in 5% increments. The resulting total cost under each concession level was then compared with the total cost under the outcome-based and prophylactic treatment models over 5, 10, and 15 years. Table 5. Sensitivity Analysis Parameter Base Value Sensitivity Analysis Risk of Failure of Hemgenix® 0.019 0 – 0.1 (0 – 10 %) Concession percentage – Reduced upfront payment model 30 % 5 – 30 % 3.8 ETHICAL CONSIDERATION The study uses a hypothetical cohort; ethical approval was not needed. However, we need to consider the potential ethical issues of conducting a comparison of reimbursement methods for ATMPs and their budgetary impacts. There is limited clinical data on Hemgenix®, and the study will be utilizing that to determine the most feasible payment method. There exists an ethical dilemma in proposing the uptake and utilization of treatments with clinical uncertainty since long-term safety and efficacy are unproven. Finally, outcome-based payment models involve patient monitoring for outcome tracking, which raises the issue of patient confidentiality. If adopted, patient privacy needs to be protected along with transparency regarding the manner of use of data to adjust payments. 14 Shalini Padhy 4 RESULTS 4.1 ASSESSMENT OF COST OF TREATMENT, HEMGENIX® The assessment of the cost associated with Hemgenix® treatment by a combination of different payment methods and long-term assumptions was evaluated and compared with standard prophylactic therapy over long-term horizons of 5, 10, and 15 years. Three scenarios were explored: treatment response as per HOPE-B outcomes, complete response, and a 10% failure rate for Hemgenix®. 4.1.1 SCENARIO 1: ALL PATIENTS RESPOND ACCORDING TO THE HOPE B TRIAL This scenario assumes a 96.29% success rate over two years, consistent with the clinical outcomes reported at 24 months in the HOPE-B trial. Figure 5 summarizes treatment costs over the three timespans for different payment models. As shown in Figure 5, over a 5- and 10-year time horizon, the cost of prophylaxis remains lower than all three Hemgenix® payment models. However, the gap narrows notably at the 10-year mark. Among the Hemgenix® options, the reduced upfront payment model consistently demonstrates the lowest cost across both 5- and 10-year timeframes. In contrast, both full upfront payment and outcome-based annuity payment are more expensive. 40 Upfront payment - no discount 30 Upfront payment - 20 30% discount Outcome based 10 payment Prophylaxis 0 5 10 15 Treatment period (years) Figure 5. Comparative cost trends across payment models for Hemgenix® and prophylaxis over time, based on a 96.29% success rate over two years reported in the HOPE-B trial. All values are in SEK. Values shown in Appendix table A3.1. By the 15-year mark, both outcome-based and reduced upfront payment models of Hemgenix® become more affordable than prophylaxis. The reduced upfront payment model remains the least costly option, saving approximately 4,561,866 SEK (4.56 million SEK) compared to the outcome-based annuity payment model. 4.1.2 SCENARIO 2: FULL RESPONSE TO HEMGENIX® This scenario assumes that 100% of patients respond positively to Hemgenix®, eliminating the need for prophylaxis over the entire study period. Figure 6 summarizes treatment costs over the three timespans for different payment models. As shown in Figure 6, over all three time-horizons (5, 10, and 15 years), the reduced upfront payment model consistently results in the least overall expenditure for Hemgenix® treatment. At 5 years, it is followed by the outcome-based annuity payment and full-upfront payment, both of which are nearly identical in cost. Prophylaxis, however, is still the least expensive option in the short term. 15 Cost of treatment (million SEK) Evaluating Payment Models for ATMPs: A Comparative Study of Different Payment Models and Reimbursement Approaches for Hemgenix® 40 Upfront payment - 30 no discount Upfront payment - 20 30% discount 10 Outcome based payment 0 Prophylaxis 5 yr 10 yr 15 yr Treatment period (years) Figure 6. Comparative cost analysis of Hemgenix® payment models versus prophylaxis under the assumption of full treatment response. All values are in SEK. Values shown in Appendix table A3.2 At the 10-year mark, prophylaxis becomes more expensive than the reduced upfront payment model of Hemgenix®. The ranking of costs (from lowest to highest) becomes reduced upfront payment, prophylaxis, outcome-based annuity payment, and full upfront payment. This shift signals the beginning of Hemgenix®'s long-term economic favorability. By 15 years, the cost trend becomes even more pronounced. Prophylaxis becomes the most expensive option, with all three Hemgenix® payment models proving more economical. Among these, the reduced upfront payment model is the cheapest option. 4.1.3 SCENARIO 3: THE FAILURE RATE OF HEMGENIX® IS 10% This scenario assumes a 10% annual failure rate of Hemgenix®, requiring those patients to transition back to prophylaxis treatment. The results reflect the additional cost burden incurred when treatment outcomes are uncertain. Figure 7 summarizes treatment costs over the three timespans for different payment models. 50 Upfront payment 40 - no discount 30 Upfront payment - 30% discount 20 Outcome based 10 payment 0 Prophylaxis 5 yr 10 yr 15 yr Treatment period (years) Figure 7. Cost comparison of Hemgenix® payment models and prophylaxis under a 10% treatment failure rate. All values are in SEK. Values shown in Appendix table A3.3 As shown in Figure 7, the 5-year cost trend mirrors that of Scenarios 1 and 2. Prophylaxis remains the least costly option, followed by the 30% reduced upfront payment and outcome-based annuity payment. Notably, the cost difference between the reduced upfront payment and outcome-based models is much smaller than in the previous scenarios, indicating that the economic advantage of a fixed price reduction is less significant when a significant portion of Hemgenix®-treated patients require ongoing prophylaxis. By the 10-year mark, treatment with Hemgenix® remains more expensive than prophylaxis under all payment models. However, by 15 years, the outcome-based model has the least financial burden among the three Hemgenix® strategies and prophylaxis, highlighting its ability to adjust payments in response to real-world performance and uncertainties. 16 Cost of treatment Cost of treatment (million SEK) (million SEK) Shalini Padhy 4.2 EFFECT OF DISCOUNT RATE ON PAYMENT RESULTS If the recommended 3% discount rate by the Pharmaceutical Benefits Board (43) is included in the analysis, then we observe a change in the cost dynamics of the payment models, as shown in Figure 8. The bar chart compares the total per-patient cost of Hemgenix® under three payment models—Upfront, reduced upfront payment, and Outcome-Based Annuity—over 5, 10, and 15-year time horizons, assuming a treatment success rate of 96.29% based on HOPE-B trial data. The two upfront payment models exhibit relatively similar cost behavior, with only a modest cost advantage for the reduced upfront payment model appearing at the 15-year horizon. However, the impact of discounting is most pronounced in the Outcome-Based Annuity model, where costs decrease noticeably as the time horizon extends. This model becomes increasingly inexpensive, with the greatest savings observed at 15 years, even surpassing the reduced upfront payment model and prophylaxis in terms of total cost. Figure 8. The effect of adding the Pharmaceutical Benefits Board (43) recommended discounting is explored on the payment model outcomes of Scenario 1. Each bar cluster displays both costs with and without 3% discounting, with black-outlined bars indicating the costs with discounting. This trend continues in the partial response scenario, where the cost-saving margin of the Outcome- Based Annuity model further widens compared to the other payment models, over 15 years. In one-way sensitivity analysis comparing the impact of clinical uncertainty on the cost of Hemgenix®, like the non-discounting results, outcome-based annuity payment becomes more affordable over a longer period. However, it is seen to start with a lower percentage of failure of Hemgenix®, as compared to the non-3% discounted costs. Comparing the outcome-based model with different discount levels of the upfront discounted model, the outcome-based model was cheaper than the reduced upfront payment model up to significantly higher levels of discounts, in contrast to non-3% discounted costs. 17 Evaluating Payment Models for ATMPs: A Comparative Study of Different Payment Models and Reimbursement Approaches for Hemgenix® 4.3 DETERMINISTIC SENSITIVITY ANALYSIS Three cases are explored in this section. 4.3.1 TREATMENT COST SENSITIVITY TO HEMGENIX® FAILURE RATES This analysis explores how changes in the annual failure rate of Hemgenix®—defined as the percentage of patients shifting from Hemgenix® to prophylaxis—impact the total treatment costs under each payment model across 5-, 10-, and 15-year time horizons. Upfront Upfront with discount Outcome based 50 45 40 35 30 25 20 15 0 1 2 3 4 5 6 7 8 9 10 % of Patients shifting from Hemgenix® to Prophylaxis Figure 9. Sensitivity analysis of total cost over 5 years as the percentage of patients transitioning from Hemgenix® to prophylaxis increases. Values shown in Appendix table A4.1. 5-Year Horizon Figure 9 illustrates a small consistent increase in the total cost for both upfront and reduced upfront payment models as failure rates rise from 0% to 10%. In contrast, the outcome-based payment model shows a slight but steady decline in total cost across the same range. This behavior reflects the design of the outcome-based model, where payment is tied to actual treatment success. At 5 years, 30% reduced upfront payment remains the most economical model, although the gap between it and the outcome- based model narrows considerably at higher failure rates. 10-Year Horizon In Figure 10, at the 10-year mark, differences in cost trajectories across the models become more pronounced. The upfront models (both full and reduced cost) show increasing cost curves as failure rates rise. In contrast, the outcome-based model decreases, albeit less steeply, due to its built-in cost adjustment based on real-world outcomes. This trend reflects the growing inefficiency of fixed-cost models considering treatment uncertainty. Outcome-based annuity payment becomes the least costly strategy above a 7% failure rate, better than the reduced upfront payment model. Upfront Upfront with discount Outcome based 50 45 40 35 30 25 20 15 0 1 2 3 4 5 6 7 8 9 10 % of Patients shifting from Hemgenix® to Prophylaxis Figure 10. Ten-year cost sensitivity analysis for increasing Hemgenix® failure rate. Values shown in Appendix table A4.2. 18 Cost Cost (million SEK) (million SEK) Shalini Padhy 15-Year Horizon In Figure 11, by 15 years, the outcome-based annuity model demonstrates good risk mitigation. While the cost of upfront payment (with or without reduction of cost) rises sharply with increasing failure rates, the cost of outcome-based annuity payment remains largely flat and stable. This results in outcome-based annuity payment being the most economic model across all failure rates beyond 5%. The data reinforces the benefit of coupling reimbursement to patient outcomes, especially in long-term treatment settings where uncertainty gets magnified. Upfront Upfront with discount Outcome based 50 45 40 35 30 25 20 15 0 1 2 3 4 5 6 7 8 9 10 % of Patients shifting from Hemgenix® to Prophylaxis Figure 11. Fifteen-year cost sensitivity analysis for increasing Hemgenix® failure rate. Values shown in Appendix table A4.3. 4.3.2 IMPACT OF CONCESSION LEVEL ON THE COST OF HEMGENIX® BY UPFRONT VERSUS OUTCOME-BASED ANNUITY PAYMENT MODELS This section compares the total treatment costs under two payment models—upfront payment with varying levels of concession (5–30%) and outcome-based annuity payment—over 5-, 10-, and 15-year horizons. Prophylaxis is the baseline reference, and the treatment success rate reflects the HOPE-B trial. Upfront payment with discount Outcome based Prophylaxis 35 30 25 20 15 10 5 5% 10% 15% 20% 25% 30% % cost reduced from full price Figure 12. Total cost comparison over 5 years between outcome-based annuity payment and upfront payment at various concession levels. Values shown in Appendix table A4.4. 5-Year Horizon As shown in Figure 12, when the upfront payment concession is low (5%), its total cost is comparable to the outcome-based model. However, as the concession increases beyond 5%, the reduced upfront 19 Cost Cost (million SEK) (million SEK) Evaluating Payment Models for ATMPs: A Comparative Study of Different Payment Models and Reimbursement Approaches for Hemgenix® payment becomes more economical. Prophylaxis remains significantly cheaper than gene therapy currently. Upfront payment with discount Outcome based Prophylaxis 35 30 25 20 15 10 5 5% 10% 15% 20% 25% 30% % cost reduced from full price Figure 13. Ten-year analysis of cost under increasing concession levels. Values shown in Appendix table A4.5. 10-Year Horizon Figure 13 shows that in 10 years, the comparative cost dynamics begin to shift. For concessions below 10%, the outcome-based model incurs lower expenditure than the reduced upfront payment. However, as the concession reaches 10% or more, the reduced upfront payment model overtakes outcome-based annuity payment as the cheaper alternative. Meanwhile, prophylaxis remains inexpensive as compared to Hemgenix®, irrespective of the payment model. 15-Year Horizon By 15 years as seen in Figure 14, the cost advantage of Hemgenix® becomes more pronounced. Prophylaxis has become the most expensive option overall, regardless of the concession level. The outcome-based model maintains a relatively stable cost, while the upfront payment with a concession becomes increasingly advantageous as the concession exceeds 15%. Upfront payment with discount Outcome based Prophylaxis 35 30 25 20 15 10 5 5% 10% 15% 20% 25% 30% % cost lowered from full price Figure 14. Fifteen-year analysis of increasing concession levels shows that upfront payment with a concession becomes the most economical strategy beyond a 15% concession. Values shown in Appendix table A4.6. The results clearly show that the affordability of upfront payment improves significantly with larger concessions, especially over longer time horizons. 20 Cost Cost (million SEK) (million SEK) Shalini Padhy 5 DISCUSSION This study compared and demonstrated the financial impact of adopting different payment models for one-time high-cost therapies like Hemgenix®, particularly considering long-term clinical uncertainty and associated financial risks for healthcare payers. The analysis was structured in two main parts: a methodological matching of APMs with treatment characteristics using a decision support framework; and a cost comparison using Markov Modelling. The framework helped identify the most suitable APMs for Hemgenix® (an ATMP). Two APMs – simple concession (30%) and outcome-based annuity – were evaluated under various clinical scenarios and time horizons. Their costs were compared to standard prophylactic treatment and a full upfront payment model. The Decision Support Framework by Callenbach et al.(34,36) provides a structured approach to match the ATMPs with preferred alternative payment models based on their evidence of clinical effectiveness and financial impact on payers. Applying the framework, the study found that outcome-based annuity payment methods were most suitable for ATMPs like Hemgenix. The results of our study using the same framework were consistent with those of Callenbach et al.(34) Cost estimates based on the long-term assumptions of the HOPE-B trial showed that, over a 5-year horizon, Hemgenix® treatment was more expensive than prophylaxis, irrespective of the payment model or assumed failure rate. The difference in cost of Hemgenix® treatment by full upfront payment and upfront payment with concession remained similar between 5- and 10-year duration. Notably, the cost difference between outcome-based annuity and the reduced upfront payment model declines from approximately 7,316,651 SEK (7.36 million SEK) at 5 years to 5,920,762 SEK (5.92 million SEK) at 10 years, indicating that outcome-based models become comparatively more favorable over time. However, by 15 years, both outcome-based and simple reduced upfront payment models became more economical than prophylaxis when we assume clinical uncertainty as per the HOPE-B trial. Similarly, in the partial response scenario, over 15 years, the outcome-based model becomes the most economical option, even outperforming standard prophylaxis. This illustrates how risk-sharing models such as outcome-based annuity payments can reduce long-term financial exposure in cases of uncertain treatment success. The full upfront payment model’s one-time nature and fixed cost structure posed a high risk in all the scenarios, as it doesn't adjust for patient outcomes. When all patients are full responders to Hemgenix®, outcome certainty improves the viability of all payment models over time. Upfront payments with considerable concessions can be most lucrative, but only when long-term efficacy is assured. Sensitivity analyses further support the study’s findings and can inform reimbursement negotiations. As the failure rate of Hemgenix® increases, the outcome-based model becomes increasingly affordable over the long term. Although higher concession rates favor upfront payment, outcome-based models remain superior at lower concession levels (5–10%), especially over extended time horizons. While large concessions can make upfront payment feasible, their benefit is tempered by clinical uncertainties and the immediate budgetary strain caused by a single large payment. The cost comparison with a 3% discount rate further emphasized the superiority of outcome-based payment, as they have a substantial amount of money paid yearly, which undergoes discounting. This contrasts with upfront payment methods, where the discounting is only applicable to cases of Hemgenix® failure on weekly prophylaxis. Even though discounting is an important feature to value costs and benefits in a cost-effectiveness analysis, it is rarely used in budget impact analysis to better understand the impact of an expense incurred in the budget. (45–47) This is because budget impact analysis aims to reflect the actual annual expenditure faced by healthcare payers. In our study, discounted costs were not utilized for the primary comparison, as our primary objective was to compare the payment models according to the actual budget outflow every year as faced by the healthcare payers. 21 Evaluating Payment Models for ATMPs: A Comparative Study of Different Payment Models and Reimbursement Approaches for Hemgenix® The value of the cost incurred in the future may change over time, but the actual cost paid for the treatment would remain the same. Using undiscounted costs ensures greater simplicity and transparency, allowing the healthcare payers to interpret the actual budget impact and effectively utilize resources. Hemgenix® is a recently introduced ATMP, and to the best of our knowledge, there is limited literature evaluating the suitability of APMs for its reimbursement. Among its literature, we refer to a study by Mairof (48) who compared direct and outcome-based payment methods for Hemgenix® against the standard prophylaxis of Alprolix. The author conducted a cost-effectiveness analysis and found the outcome-based annuity payment method to have a higher cost saving as compared to direct payments, especially with increasing loss of ATMP efficacy.(48) Previous cost-effectiveness studies by the Institute for Clinical and Economic Review (ICER) (49) and TLV (37) found Hemgenix® to be dominant over prophylaxis, but highlighted considerable uncertainty regarding its long-term clinical outcomes. The clinical efficacy can be limited by a lack of long-term data from the HOPE-B trial and potential real-world deviations, such as the development of comorbidities (e.g., Hepatitis C), thus reducing the efficacy of gene therapy.(50) Cost-effectiveness analysis compares the health gained in relation to the cost incurred by alternative interventions.(51) Our study predominantly focused on comparing the costs incurred due to different payment models for the same intervention, as compared to different interventions, thus differing from cost-effectiveness analysis. ICER (52) states that a fair price for therapy will entail covering the cost of research, development, production, and manufacturing, along with a reasonable profit. While deciding the price with a concession, a fair price must be considered, while aiming to reduce the expenditure for the payers. TLV (22) also substantiates the superiority of highly reduced pricing over other models, while highlighting the limitation of the inability of payers and companies to agree on an extremely high concession. Therefore, pricing negotiations must balance affordability with fair compensation to manufacturers. Given the challenge of fair yet sustainable pricing, APMs like outcome-based payment models enforced by MEAs are increasingly gaining relevance to address clinical uncertainty and affordability. To support this, structured tracking of real-world effectiveness needs to be established by the health authorities for the successful adoption of ATMPs like Hemgenix®. Sweden has predominantly implemented risk- sharing MEAs for 56 therapeutic products, but all are non-outcome-based.(52) Access of Sweden to ATMPs is low compared to countries like Germany, highlighting the need for innovative reimbursement methods.(53) Linking payment to real-world patient outcomes would help manage clinical uncertainty by allowing adjustments in reimbursement if treatment effectiveness deviates from expected levels.(54) Result monitoring for Hemgenix® would involve regular data collection on key indicators, such as bleeding episodes, FIX activity levels, and health-related quality of life measures. In addition to promoting financial risk-sharing between payers and manufacturers, these agreements can also create incentives for continuous monitoring and evidence generation.(54) The MEAs could improve payer confidence, leading to early access for patients, but outcome-based annuity payment models are not without their limitations. A few of the challenges faced by MEA with outcome- based annuity payments are the additional monetary burden and administrative complications for healthcare authorities with data collection and management, difficulty in stopping payment for an ineffective treatment, and the manufacturers may decide on a high initial price to ensure their financial security in case of treatment failure.(54) A report by the Alliance of Regenerative Medicine (55) states that successful risk-sharing agreements between manufacturers and the payers would require an early dialogue highlighting the expectations of the payers from the MEA. They mention that manufacturers are more open to outcome-based payment methods in contrast to refunds on non-efficacy, but are limited by a lack of flexibility, openness of payers, and a standardized framework for adoption and implementation of the innovative MEAs. There is a need to acknowledge the uniqueness of ATMPs and the adoption of a sustainable payment method that benefits both healthcare payers and ATMP manufacturers. (55) 22 Shalini Padhy 5.1 STUDY STRENGTHS Payment models adopted for high-cost, one-time ATMPs, like Hemgenix®, should be able to overcome two main obstacles to access: clinical uncertainty and short-term budgetary impact for the payers. (56) The main strength of this study was the analysis of different scenarios simulating the clinical uncertainties and time horizons (5, 10, and 15 years), capturing the chronic and lifelong nature of Hemophilia B. Moreover, the study conducts comprehensive deterministic sensitivity and scenario analysis to explore the effects of transition probabilities and concession rates, strengthening the model's robustness under uncertainty, an important consideration for ATMPs. Several European countries have adopted alternative payment methods for ATMPs; for example, simple concessions have been applied to therapies like Luxturna® and Kymriah® in Italy and Luxturna® in the UK, although the exact concession rates are confidential.(55) In another example, a 10-year outcome-based payment has been adopted for Hemgenix® in Denmark, but the figures and details remain classified. (38) The three payment models - upfront payment, reduced upfront payment, and outcome-based payment reflect those commonly implemented across Europe, increasing the policy relevance of the study. 5.2 STUDY LIMITATIONS Even though the study provides valuable insight into the feasibility of different payment models for Hemgenix®, the study has several limitations. First, the lack of real-world data led to the simplification of the Markov model, including only three health states. The design does not fully capture the clinical events and outcome endpoints, such as reduction in bleed rates, hospitalization, or joint damage, which impact both costs and quality of life. Secondly, a fixed transition probability was assumed for the entire duration of the model, which may not reflect the real-world conditions. The patients may experience a fluctuating failure rate due to personal patient-specific factors like antibody formation or comorbidities (e.g., liver disease). Thirdly, the study is conducted assuming similar health outcomes for both Hemgenix® and prophylaxis treatments as reflected in the HOPE-B trial. They are based on short-term clinical trial data. This potentially underestimates the true long-term quality-of-life benefit of one-time gene therapies like Hemgenix®, which eliminates the need for frequent infusions and hospital visits. Lastly, indirect costs for patients and caregivers, such as productivity losses, which are particularly relevant for chronic diseases like Hemophilia B, were also not incorporated. 23 Evaluating Payment Models for ATMPs: A Comparative Study of Different Payment Models and Reimbursement Approaches for Hemgenix® 6 CONCLUSION The study examined the impact on healthcare payers of reimbursing Hemgenix®, a high-cost, one-time gene therapy for Hemophilia B, through three payment models: full upfront payment, reduced upfront payment, and an outcome-based annuity model. A Markov model was used to study the long-term costs over 5-, 10-, and 15-year horizons. The results indicate that while in the short term, Hemgenix® incurs higher costs as compared to prophylactic treatment, its affordability improves over time. While the outcome-based model proved to be the most adaptable to the clinical uncertainty of the treatment, the reduced upfront payment emerged as the most affordable option under ideal treatment scenarios. 24 Shalini Padhy 7 PUBLIC HEALTH PERSPECTIVES/IMPLICATIONS ATMPs are a diverse group of therapeutic products, promising long-term, potentially curative outcomes for rare diseases and diseases of unmet needs, which can significantly reduce the lifetime disease burden on patients, caregivers, and the healthcare system.(55) However, the main barrier to early and widespread uptake is their high one-time cost and lack of evidence of clinical efficacy.(55) Diseases such as Hemophilia B have a considerable socioeconomic burden on their patients and the caregivers. Early retirement of the patients (60%) and work absenteeism of caregivers (24%) were observed in a study by Burke et al. (57) in 5 European countries, highlighting the public health burden of prophylactically treated Hemophilia B. The study showcases the critical role of alternative payment methods in enabling early access to life- changing therapies without compromising the financial sustainability of healthcare systems. Outcome– based annuity payment methods link the payer investment with real-world clinical outcomes, thus safeguarding the scarce resources. They mitigate the financial risk and spread the costs over time, thus softening the budget impact and encouraging early adoption of ATMPs in the rare disease population, reducing health inequalities. Studying the financial impact on healthcare payers by comparing different payment methods for ATMPs like Hemgenix® provides relevant information to policymakers and HTA agencies for informed decision-making. It encourages greater stakeholder collaboration between the regulators, manufacturers, payers, clinicians, and patients for accelerated adoption of ATMPs, boosting pharmaceutical innovation and reducing public health burden. However, more research is needed to confirm and build on these findings. Future research using real- world data is needed to better understand treatment efficacy and quality of life after gene therapy. A more comprehensive Markov model, including more health states, such as spontaneous bleeds, joint surgery, could give a more complete picture of the true economic impact. Finally, it would also be important to account for differences between patients, such as age, comorbidities, and physical activities, to improve the accuracy of cost predictions. 25 Evaluating Payment Models for ATMPs: A Comparative Study of Different Payment Models and Reimbursement Approaches for Hemgenix® 8 DECLARATION OF AI AND AI-ASSISTED TECHNOLOGIES During the preparation of this work, I used ChatGPT for the refinement of language, grammar, and sentence structures. It was used to improve clarity and coherence in my writing. After using the tool, I reviewed and edited the content as needed and take full responsibility for the content of the work. 26 ACKNOWLEDGEMENT I would like to thank my supervisor, Pia Johansson, for her guidance, insightful feedback, and academic expertise, which were invaluable throughout this research process. I am grateful to my institution, the University of Gothenburg, for providing the academic resources, support, and infrastructure needed to pursue this research. On a personal note, I would like to thank my husband for his patience and support. I would also like to thank my parents, in-laws, and the rest of the family for their encouragement and support throughout the journey. 27 Evaluating Payment Models for ATMPs: A Comparative Study of Different Payment Models and Reimbursement Approaches for Hemgenix® REFERENCES 1. Hanna E, Rémuzat C, Auquier P, Toumi M. Advanced therapy medicinal products: current and future perspectives. J Mark Access Health Policy. 2016 Jan;4(1):31036. 2. Commission Directive 2009/120/EC of 14 September 2009 amending Directive 2001/83/EC of the European Parliament and of the Council on the Community code relating to medicinal products for human use as regards advanced therapy medicinal productsText with EEA relevance. 3. Ginn SL, Mandwie M, Alexander IE, Edelstein M, Abedi MR. Gene therapy clinical trials worldwide to 2023—an update. J Gene Med. 2024;26(8):e3721. 4. CAT quarterly highlights and approved ATMPs. Amsterdam: European Medicines Agency; 2025 Feb p. 5. Report No.: EMA/CAT/52354/2025. 5. Hemgenix | European Medicines Agency (EMA) [Internet]. 2023 [cited 2025 Feb 3]. Available from: https://www.ema.europa.eu/en/medicines/human/EPAR/hemgenix 6. Srivastava A, Santagostino E, Dougall A, Kitchen S, Sutherland M, Pipe SW, et al. WFH Guidelines for the Management of Hemophilia, 3rd edition. Haemophilia. 2020;26(S6):1–158. 7. Iorio A, Stonebraker JS, Chambost H, Makris M, Coffin D, Herr C, et al. Establishing the Prevalence and Prevalence at Birth of Hemophilia in Males. Ann Intern Med. 2019 Oct 15;171(8):540–6. 8. Orphan Maintenance Assessment Report - Hemgenix [Internet]. European Medical Agency; 2023 Feb p. 16. Report No.: EMA/OD/0000087180. Available from: https://www.ema.europa.eu/en/documents/orphan-maintenance-report/hemgenix-epar-orphan- maintenance-assessment-report_en.pdf 9. White G. C. II, Frits Rosendaal, Jeanne M. Lusher, Chantal Rothschild, & Factor VIII and Factor IX Subcommittee (2001). Definitions in hemophilia: recommendation of the Scientific Subcommittee on Factor VIII and Factor IX of the Scientific and Standardization Committee of the International Society on Thrombosis and Haemostasis. Thromb Haemost. 2001 Mar;85(3):560. 10. Nathwani AC. Gene therapy for hemophilia. Hematol Am Soc Hematol Educ Program. 2022 Dec 9;2022(1):569–78. 11. Sheh-Li Chen P. Economic Costs of Hemophilia and the Impact of Prophylactic Treatment on Patient Management. Am J Manag Care. 2016 Apr 18;22(1;22(5 Suppl)):s126-33. 12. Powell JS, Pasi KJ, Ragni MV, Ozelo MC, Valentino LA, Mahlangu JN, et al. Phase 3 Study of Recombinant Factor IX Fc Fusion Protein in Hemophilia B. N Engl J Med. 2013 Dec 12;369(24):2313–23. 13. Thornburg CD. Etranacogene dezaparvovec for hemophilia B gene therapy. Ther Adv Rare Dis. 2021 Nov 24;2:26330040211058896. 14. Naso MF, Tomkowicz B, Perry WL, Strohl WR. Adeno-Associated Virus (AAV) as a Vector for Gene Therapy. Biodrugs. 2017;31(4):317–34. 15. Nathwani AC, Gray JT, McIntosh J, Ng CYC, Zhou J, Spence Y, et al. Safe and efficient transduction of the liver after peripheral vein infusion of self-complementary AAV vector results in stable therapeutic expression of human FIX in nonhuman primates. Blood. 2007 Feb 15;109(4):1414–21. 28 16. Nathwani AC, Tuddenham EGD, Rangarajan S, Rosales C, McIntosh J, Linch DC, et al. Adenovirus-Associated Virus Vector–Mediated Gene Transfer in Hemophilia B. N Engl J Med. 2011 Dec 22;365(25):2357–65. 17. Manno CS, Pierce GF, Arruda VR, Glader B, Ragni M, Rasko JJE, et al. Successful transduction of liver in hemophilia by AAV-Factor IX and limitations imposed by the host immune response. Nat Med. 2006 Mar 1;12(3):342–7. 18. Pipe SW, Leebeek FWG, Recht M, Key NS, Castaman G, Miesbach W, et al. Gene Therapy with Etranacogene Dezaparvovec for Hemophilia B. N Engl J Med. 2023 Feb 22;388(8):706–18. 19. First gene therapy to treat haemophilia B | European Medicines Agency (EMA) [Internet]. 2022 [cited 2025 Apr 17]. Available from: https://www.ema.europa.eu/en/news/first-gene-therapy-treat- haemophilia-b 20. D’Avola D, López-Franco E, Sangro B, Pañeda A, Grossios N, Gil-Farina I, et al. Phase I open label liver-directed gene therapy clinical trial for acute intermittent porphyria. J Hepatol. 2016 Oct 1;65(4):776–83. 21. Boutin S, Monteilhet V, Veron P, Leborgne C, Benveniste O, Montus MF, et al. Prevalence of Serum IgG and Neutralizing Factors Against Adeno-Associated Virus (AAV) Types 1, 2, 5, 6, 8, and 9 in the Healthy Population: Implications for Gene Therapy Using AAV Vectors. Hum Gene Ther. 2010 Jun;21(6):704–12. 22. Coppens M, Pipe SW, Miesbach W, Astermark J, Recht M, Valk P van der, et al. Etranacogene dezaparvovec gene therapy for haemophilia B (HOPE-B): 24-month post-hoc efficacy and safety data from a single-arm, multicentre, phase 3 trial. Lancet Haematol. 2024 Apr 1;11(4):e265–75. 23. Performance-based managed entry agreements for new medicines in OECD countries and EU member states: How they work and possible improvements going forward [Internet]. 2019 Dec [cited 2025 Jan 29]. (OECD Health Working Papers; vol. 115). Report No.: 115. Available from: https://www.oecd.org/en/publications/performance-based-managed-entry-agreements-for-new- medicines-in-oecd-countries-and-eu-member-states_6e5e4c0f-en.html 24. How should we assess and pay? Health-economic assessments and payment models for precision medicines and ATMPs -2021, The Dental and Pharmaceutical Benefits Agency [Internet]. [cited 2025 Apr 17]. Available from: https://www.tlv.se/download/18.2f080b7e182629be22dd4e3c/1660743327830/rapport_rupm_atm p_eng.pdf 25. Pearce KF, Hildebrandt M, Greinix H, Scheding S, Koehl U, Worel N, et al. Regulation of advanced therapy medicinal products in Europe and the role of academia. Cytotherapy. 2014 Mar 1;16(3):289– 97. 26. Detela G, Lodge A. EU Regulatory Pathways for ATMPs: Standard, Accelerated and Adaptive Pathways to Marketing Authorisation. Mol Ther Methods Clin Dev. 2019 Jan 29;13:205–32. 27. Drummond MF, Sculpher MJ, Claxton K, Stoddart GL, Torrance GW. Methods for the Economic Evaluation of Health Care Programmes [Internet]. Oxford, UNITED KINGDOM: Oxford University Press; 2015 [cited 2025 Mar 28]. Available from: http://ebookcentral.proquest.com/lib/gu/detail.action?docID=4605509 28. Briggs AH. Decision modelling for health economic evaluation. Claxton K, Sculpher M, editors. Oxford: Oxford University Press; 2011. 237 p. (Handbooks in health economic evaluation series). 29. Glick H, Doshi JA, Sonnad SS, Polsky D. Economic evaluation in clinical trials. Second edition. Oxford: Oxford University Press; 2015. 252 p. (Handbooks in health economic evaluation series). 29 Evaluating Payment Models for ATMPs: A Comparative Study of Different Payment Models and Reimbursement Approaches for Hemgenix® 30. EQ-5D-3L User Guide [Internet]. EuroQol Research Foundation; 2018. Available from: https://euroqol.org/publications/user-guides 31. Izeta A, Cuende N. Regulation of advanced therapies in Europe: Are we on the right track? Cell Stem Cell. 2023 Aug 3;30(8):1013–6. 32. Graham Foxon. Remap Consulting. 2022. Why is the number of advanced therapy medicinal product (ATMP) withdrawals from the European market rising? Available from: https://remapconsulting.com/hta/why-is-the-number-of-advanced-therapy-medicinal-product- atmp-withdrawals-from-the-european-market- rising/#:~:text=The%20majority%20of%20ATMP%20product,uncertainty%20and%20hefty%20p rice%20tags. 33. Lucion D, Ghinelli F, Buxton J, Wright J, Steeds C, Hall S, et al. HTA256 Identifying Current ATMP Market Access Challenges and Strategies in France, Germany, Italy, and Spain. Value Health. 2023 Dec 1;26(12.):S369-70. 34. Callenbach MHE, Vreman RA, Leopold C, Mantel-Teeuwisse AK, Goettsch WG. Managed Entry Agreements for High-Cost, One-Off Potentially Curative Therapies: A Framework and Calculation Tool to Determine Their Suitability. PharmacoEconomics. 2025 Jan 1;43(1):53–66. 35. Towse A, Fenwick E. Uncertainty and Cures: Discontinuation, Irreversibility, and Outcomes-Based Payments: What Is Different About a One-Off Treatment? Value Health. 2019 Jun 1;22(6):677–83. 36. Callenbach MHE, Schoenmakers D, Vreman RA, Vijgen S, Timmers L, Hollak CEM, et al. Illustrating the Financial Consequences of Outcome-Based Payment Models From a Payers Perspective: The Case of Autologous Gene Therapy Atidarsagene Autotemcel (Libmeldy®). Value Health. 2024 Aug 1;27(8):1046–57. 37. TLV (Swedish Pharmaceutical Benefits Agency). Hälsoekonomisk bedömning av Hemgenix (etranako-gen-dezaparvovek)(Health Economic Assessment of Hemgenix). 2023 Oct. 38. statistikdatabasen. Mortality rate, per 1,000 of the mean population by age, sex and year. PxWeb [Internet]. 2023 [cited 2025 Apr 7]. Available from: https://www.statistikdatabasen.scb.se/pxweb/en/ssd/START__BE__BE0101__BE0101I/Dodstal/ta ble/tableViewLayout1/ 39. CSL Behring. Phase III, Open-label, Single-dose, Multi-center, Multinational Trial Investigating a Serotype 5 Adeno-associated Viral Vector Containing the Padua Variant of a Codon-optimized Human Factor IX Gene (AAV5-hFIXco-Padua, AMT-061) Administered to Adult Subjects With Severe or Moderately Severe Hemophilia B [Internet]. clinicaltrials.gov; 2025 Feb [cited 2025 Feb 24]. Report No.: NCT03569891. Available from: https://clinicaltrials.gov/study/NCT03569891 40. Persson U, Norlin J. Opportunities and challenges of an innovative payment model for Advanced Therapy Medicinal Products in Sweden – A hypothetical example in the treatment of β- thalassemia. IHE: Lund, Sweden: IHE Report 2020:3; 41. Wills A, Mitha A. Financial Characteristics of Outcomes-Based Agreements: What Do Canadian Public Payers and Pharmaceutical Manufacturers Prefer? Value Health. 2024 Mar;27(3):340–6. 42. CSL Behring Proposes 10-Year Pay-for-Performance Agreement for Hemgenix in Netherlands [Internet]. [cited 2025 Mar 31]. Available from: https://www.navlindaily.com/article/22474/csl- behring-proposes-10-year-pay-for-performance-agreement-for-hemgenix-in-netherlands 43. Axel Edling, Anna Märta Stenberg. General guidelines for economic evaluations from the Pharmaceutical Benefits Board (LFNAR 2003:2). 2003 Apr p. 3. 30 44. Gidwani R, Russell LB. Estimating Transition Probabilities from Published Evidence: A Tutorial for Decision Modelers. PharmacoEconomics. 2020 Nov 1;38(11):1153–64. 45. Leelahavarong P. Budget impact analysis. J Med Assoc Thail Chotmaihet Thangphaet. 2014 May 1;97 Suppl 5:S65-71. 46. Garattini L, van de Vooren K. Budget impact analysis in economic evaluation: a proposal for a clearer definition. Eur J Health Econ. 2011 Dec 1;12(6):499–502. 47. Sullivan SD, Mauskopf JA, Augustovski F, Jaime Caro J, Lee KM, Minchin M, et al. Budget Impact Analysis—Principles of Good Practice: Report of the ISPOR 2012 Budget Impact Analysis Good Practice II Task Force. Value Health. 2014 Jan 1;17(1):5–14. 48. Mairof D. Exploring the Impact of Two Payment Models in Health Economics: [Internet] [thesis]. [Uppsala]: Uppsala University; 2024. Available from: https://uu.diva- portal.org/smash/record.jsf?pid=diva2%3A1869627&dswid=-445 49. Tice JA, Walton S, Herce-Hagiwara B, Fahim SM, Moradi A, Sarker J, et al. Gene Therapy for Hemophilia B and An Update on Gene Therapy for Hemophilia A: Effectiveness and Value; Evidence Report. [Internet]. Institute for Clinical and Economic Review; 2022 Dec p. 136. Available from: https://icer.org/assessment/hemophilia-a-and-b-2022/ 50. Shah J, Kim H, Sivamurthy K, Monahan PE, Fries M. Comprehensive analysis and prediction of long-term durability of factor IX activity following etranacogene dezaparvovec gene therapy in the treatment of hemophilia B. Curr Med Res Opin [Internet]. 2023 Feb 1 [cited 2025 Apr 24]; Available from: https://www.tandfonline.com/doi/abs/10.1080/03007995.2022.2133492 51. Jamison DT, Breman JG, Measham AR, Alleyne G, Claeson M, Evans DB, et al. Cost-Effectiveness Analysis. In: Priorities in Health [Internet]. The International Bank for Reconstruction and Development / The World Bank; 2006 [cited 2025 May 7]. Available from: https://www.ncbi.nlm.nih.gov/books/NBK10253/ 52. Andersson E, Svensson J, Persson U, Lindgren P. Risk sharing in managed entry agreements—A review of the Swedish experience. Health Policy. 2020 Apr 1;124(4):404–10. 53. Fischer P, Reiss T, Mahlich J, Gicquel E, Aichinger H, Pullmann L, et al. Unlocking the value of innovative medicines: Insights from the advanced therapy medicinal products (ATMP) innovation systems in Germany and Sweden. Health Policy Technol. 2023 Jun 1;12(2):100744. 54. Wollberg AR. Final report for Swelife-ATMP system development project 3 (SDP3). p. 57. (Business models and health economics). 55. Prepared by Dolon. INNOVATIVE CONTRACTING FOR ATMPS IN EUROPE: Recent learnings from the manufacturer experience [Internet]. Alliance for Regenerative Medicine; 2023 Aug p. 54. Available from: https://dolon.com/wp-content/uploads/2023/08/Innovative-contracting- for-ATMPs-in-Europe-1.pdf?x23572 56. Phares S, Trusheim M, Emond SK, Pearson SD. Managing the Challenges of Paying for Gene Therapy: Strategies for Market Action and Policy Reform. 57. Burke T, Asghar S, O’Hara J, Chuang M, Sawyer EK, Li N. Clinical, humanistic, and economic burden of severe haemophilia B in adults receiving factor IX prophylaxis: findings from the CHESS II real-world burden of illness study in Europe. Orphanet J Rare Dis. 2021 Dec 20;16:521. 58. Orphan Maintenance Assessment Report - Hemgenix [Internet]. European Medicines Agency; 2022 p. 16. Report No.: EMA/OD/0000087180. Available from: https://www.ema.europa.eu/en/documents/orphan-maintenance-report/hemgenix-epar-orphan- maintenance-assessment-report_en.pdf 31 Evaluating Payment Models for ATMPs: A Comparative Study of Different Payment Models and Reimbursement Approaches for Hemgenix® 32 APPENDICES A1: DECISION SUPPORT FRAMEWORK (CALLENBACH ET AL.) (34,36) Step 1: Characteristics of the disease The disease characteristics are defined by PICO (Patient, Intervention, Comparator, and Outcomes).(34) The following section outlines the corresponding information accordingly. Patient population Table A1.1 and Table A1.2 define the population characteristics for reimbursement (for an exhaustive list, refer to Appendix A2) and the total expected prevalence of Hemophilia B in Sweden. Table A1.1. Population selection criteria for reimbursement Criteria Category Details (a) Inclusion Sex Males Diagnosis Diagnosed with Hemophilia B and is currently on prophylactic treatment Age Above 18 years Disease Severity Moderate to severe Hemophilia B (≤ 2% of normal circulating factor IX) Treatment stability On stable prophylaxis for at least 2 months Exclusion Inhibitor history Absence of present or past factor IX inhibitors Allergy history Known allergic reaction or anaphylaxis to factor IX products Liver function Abnormal liver function test Source - (18) Table A1.2. Expected Prevalence of Hemophilia B Item Description (b) Expected Incidence / Prevalence Rate 5 per 100,000 males (c) Total Expected Prevalence in Sweden 209 patients in 2020 Source - (7,58) Intervention One-off treatment of Etranacogene dezaparvovec (Hemgenix®).(39) Comparator The current standard of care is ongoing Prophylactic treatment with factor IX concentrate.(6) Alprolix (eftrenonacog alfa), an Extended half-life recombinant FIX concentrate with a dosage every 7 days.(37) Step 2: Characteristics of treatment Outcomes Clinically relevant outcomes that are frequently measured and meaningful to patients include (39): • Reduction in bleeding episodes. • Elimination of the need for ongoing prophylaxis. • Administration of treatment 33 Evaluating Payment Models for ATMPs: A Comparative Study of Different Payment Models and Reimbursement Approaches for Hemgenix® A single intravenous infusion was administered at a dose of 2 × 10¹³ genome copies (gc) / kg body weight.(39) Step 3: Matching disease and treatment characteristics with reimbursement and payment models Are there uncertainties in clinical effectiveness? Yes. There is limited data available from the HOPE-B trial, which provides follow-up for only 24 months. Treatment outcome beyond 24 months is uncertain and needs long-term monitoring.(22) Additionally, the sample size is small, and the bleeding episodes were self-reported.(22,37) Are these uncertainties significant enough to influence the reimbursement decision? Yes. The primary benefit of Hemgenix® is the potential to eliminate the need for ongoing prophylactic treatment. If this outcome proves uncertain, a full upfront reimbursement poses a risk and may not be justified in the event of treatment failure. Is the upfront payment of the therapy financially manageable? No. Hemgenix® has a high upfront cost, making full immediate reimbursement challenging. Is there a financial risk related to the clinical uncertainties identified in Step 2 that can be mitigated by using a delayed payment model? Yes. The short follow-up duration of the HOPE-B trial (24 months) contributes to long-term clinical uncertainty on sustained effectiveness. A delayed payment model, such as an annuity-based reimbursement, can reduce the financial risk arising from these uncertainties.(37) Which reimbursement models could address these uncertainties or challenges? An outcome-based payment model is preferred to address the clinical uncertainties associated with the treatment.(37) This would include payment linked to demonstrated outcomes or coverage with evidence development, ensuring reimbursement is tied to the actual effectiveness of the therapy over time. Step 4: Selection of payment models to be studied for the disease and treatment Comparison of Upfront, reduced upfront, and outcome-based annuity models. 34 A2: LIST OF INCLUSION AND EXCLUSION CRITERIA FOR GENE THERAPY, HEMGENIX®, Source: Supplementary document Pipe et al. (18) Inclusion Criteria 1. Male gender 2. At least 18 years old 3. Diagnosed with Hemophilia B, with a documented severe or moderately severe factor IX (FIX) deficiency (≤2% of normal FIX levels) 4. Receiving ongoing routine FIX prophylactic treatment 5. History of over 150 prior exposure days to FIX therapy 6. Maintained a stable prophylaxis regimen for a minimum of 2 months before screening 7. Demonstrated the ability to independently and accurately complete the study diary on time during the lead-in phase, as assessed by the investigator 8. Committed to using a condom during sexual activity from the time of IMP administration until AAV5 was confirmed to be cleared from semen, defined by at least three consecutive negative semen samples analyzed by a central lab (this applied even to participants who had undergone surgical sterilization) 9. Capable of providing informed consent after being thoroughly informed verbally and in writing about the study. Exclusion Criteria Participants were not eligible for the trial if any of the following applied: 1. Previous history of factor IX (FIX) inhibitors 2. Tested positive for FIX inhibitors at both screening and the final lead-in visit (per local lab results) 3. Laboratory values at screening and final lead-in visit (from central lab) showed: 4. Alanine aminotransferase (ALT) > 2× upper limit of normal (ULN) 5. Aspartate aminotransferase (AST) > 2× upper limit of normal (ULN) 6. Total bilirubin > 2× upper limit of normal (ULN) (unless due to Gilbert’s syndrome) 7. Alkaline phosphatase > 2× upper limit of normal (ULN) 8. Creatinine > 2× upper limit of normal (ULN) 9. Positive HIV test at both screening and final lead-in visit, with CD4+ count ≤ 200/μL and not adequately managed with antiviral therapy 10. Evidence of active Hepatitis B or C infection at screening: 11. Currently receiving antiviral therapy 12. And/or positive for: 13. Hepatitis B surface antigen (unless attributed to prior vaccination) 14. Hepatitis B virus DNA 15. Hepatitis C virus RNA 16. Diagnosis of a coagulation disorder other than Hemophilia B 17. Low platelet count (<50 × 10⁹/L) at both screening and final lead-in visit 18. Presence of severe infection or any serious, uncontrolled medical condition (e.g., renal, liver, heart, blood, digestive, hormonal, lung, nerve, brain, or psychiatric issues), including alcohol or drug dependence, that could impair compliance or tolerance to the investigational treatment, as determined by the investigator 19. Medical conditions that could significantly impair vector transduction or protein activity/expression, including: 20. Disseminated intravascular coagulation 35 Evaluating Payment Models for ATMPs: A Comparative Study of Different Payment Models and Reimbursement Approaches for Hemgenix® 21. Accelerated fibrinolysis 22. Advanced liver fibrosis (e.g., equivalent to METAVIR Stage 3, or FibroScan ≥ 9 kPa) 23. History of allergic reaction or anaphylaxis to FIX treatments 24. Known allergy to corticosteroids 25. Known uncontrolled allergies or hypersensitivity to components in the investigational product 26. Any condition that requires long-term corticosteroid use 27. Prior participation in gene therapy 28. Use of any experimental treatment within 60 days before screening 29. Concurrent or planned participation in another interventional clinical trial involving drugs or medical devices within one year of receiving the investigational product. 36 A3: TREATMENT COSTS Table A3.1: Total cost of treatment per patient under the HOPE-B response scenario, comparing three Hemgenix® payment models with standard prophylaxis over 5, 10, and 15 years. Data derived from Markov model simulations. All values are in SEK Time Upfront payment – no Upfront payment - 30% Outcome-based span concession reduction payment Prophylaxis (Years) (SEK) (SEK) (SEK) (SEK) 5 29,862,656 21,087,334 28,403,984 10,999,720 10 31,409,612 22,634,289 28,555,051 21,884,989 15 33,769,507 24,994,184 29,556,050 32,573,342 Table A3.2: Total cost of treatment per patient across three Hemgenix® payment models and prophylaxis, assuming 100% treatment success and no need for prophylaxis. Costs are presented over 5-, 10-, and 15-year time horizons. All values are in SEK. Time Upfront payment – no Upfront payment - 30% Outcome-based Prophylaxis span concession reduction payment (SEK) (years) (SEK) (SEK) (SEK) 5 29,251,075 20,475,753 29,141,609 10,999,720 10 29, 251,075 20, 475,753 28, 993,935 21,884,989 15 29,251,075 20, 475,753 28,775,652 32,573,342 Table A3.3: Total cost of treatment per patient across Hemgenix® payment models and prophylaxis, assuming a 10% annual failure rate and transition to prophylaxis. Costs are presented over 5-, 10-, and 15-year horizons. All values are in SEK. Time Upfront payment – no Upfront payment - 30% Outcome-based Prophylaxis span concession reduction payment (SEK) (years) (SEK) (SEK) (SEK) 5 32,143,749 23,368,426 25,652,769 10,999,720 10 38,296,164 29,520,842 27,154,846 21,884,989 15 46,244,696 37,469,373 31,710,694 32,573,342 37 Evaluating Payment Models for ATMPs: A Comparative Study of Different Payment Models and Reimbursement Approaches for Hemgenix® A4: SENSITIVITY ANALYSIS TREATMENT COST SENSITIVITY TO HEMGENIX® FAILURE RATES Table A4.1: 5 years horizon - Total cost of treatment per patient using different payment models depending on the increasing risk of failure of Hemgenix® from 0 % to 10%. All values are in SEK. Risk Of Treatment Upfront payment Upfront payment with a Outcome-based annuity Failure (%) (in SEK) 30% reduction (in SEK) payment (in SEK) 0 29,251,075 20,475,753 29,141,609 1 29,576,838 20,801,516 28,748,708 2 29,893,989 21,118,667 28,366,194 3 30,202,722 21,427,399 27,993,833 4 30,503,226 21,727,904 27,631,396 5 30,795,690 22,020,368 27,278,656 6 31,080,298 22,304,976 26,935,392 7 31,357,231 22,581,909 26,601,385 8 31,626,668 22,851,345 26,276,419 9 31,888,783 23,113,460 25,960,283 10 3,214,3749 23,368,426 25,652,769 Table A4.2: 10 years horizon - Total cost of treatment per patient using different payment models depending on the increasing risk of failure of Hemgenix® from 0 % to 10%. All values are in SEK. Risk Of Treatment Upfront payment Upfront payment with a 30% Outcome-based annuity Failure (%) (in SEK) reduction (in SEK) payment (in SEK) 0 29,251,075 20,475,753 28,993,935 1 30,417,766 21,642,444 28,756,718 2 31,516,549 22,741,227 28,533,308 3 32,551,403 23,776,080 28,322,897 4 33,526,086 24,750,763 28,124,720 5 34,444,151 25,668,828 27,938,055 6 35,308,950 26,533,627 27,762,220 7 36,123,648 27,348,326 27,596,572 8 36,891,232 28,115,910 27,440,503 9 37,614,519 28,839,196 27,293,441 10 38,296,164 29,520,842 27,154,846 38 Table A4.3: 15 years horizon - Total cost of treatment per patient using different payment models depending on the increasing risk of failure of Hemgenix® from 0 % to 10%. All values are in SEK. Risk Of Treatment Upfront payment Upfront payment with a 30% Outcome-based annuity Failure (%) (in SEK) reduction payment (in SEK) (in SEK) 0 29,251,075 20,475,753 28,775,652 1 31728,208 22,952,886 29,203,489 2 33,986,014 25,210,692 29,593,444 3 36,044,894 27,269,572 29,949,042 4 37,923,364 29,148,042 30,273,481 5 39,638,228 30,862,905 30,569,662 6 41,204,724 32,429,402 30,840,219 7 42,636,672 33,861,349 31,087,536 8 43,946,595 35,171,272 31,313,779 9 45,145,843 36,370,520 31,520,906 10 46,244,696 37,469,373 31,710,694 IMPACT OF CONCESSION LEVEL ON THE COST OF HEMGENIX® BY UPFRONT VERSUS OUTCOME-BASED ANNUITY PAYMENT MODELS Table A4.4: Cost of Hemgenix® treatment per patient by different concession percentages as compared to the outcome-based payment model and prophylaxis over 5 years. All values are in SEK Percentage Reduced upfront payment Outcome-based annuity Prophylaxis Concession Model (in SEK) model (in SEK) (in SEK) 5% 28,400,102 28,403,984 10,999,720 10% 26,937,549 28,403,984 10,999,720 15% 25,474,995 28,403,984 10,999,720 20% 24,012,441 28,403,984 10,999,720 25% 22,549,887 28,403,984 10,999,720 30% 21,087,334 28,403,984 10,999,720 Table A4.5: Cost of Hemgenix® treatment per patient by different concession percentages as compared to the outcome-based payment model and prophylaxis over 10 years. All values are in SEK Percentage Reduced upfront payment Outcome-based annuity Prophylaxis Concession Model (in SEK) model (in SEK) (in SEK) 5% 29,947,058 28,555,051 21,884,989 10% 28,484,504 28,555,051 21,884,989 15% 27,021,950 28,555,051 21,884,989 20% 25,559,397 28,555,051 21,884,989 25% 24,096,843 28,555,051 21,884,989 30% 22,634,289 28,555,051 21,884,989 39 Evaluating Payment Models for ATMPs: A Comparative Study of Different Payment Models and Reimbursement Approaches for Hemgenix® Table A4.6: Cost of Hemgenix® treatment per patient by different percentages of concession as compared to the outcome-based payment model and prophylaxis over 15 years. All values are in SEK Percentage Reduced upfront payment Outcome-based annuity Prophylaxis Concession Model (in SEK) model (in SEK) (in SEK) 5% 32,306,953 29,556,050 32,573,342 10% 30,844,399 29,556,050 32,573,342 15% 29,381,846 29,556,050 32,573,342 20% 27,919,292 29,556,050 32,573,342 25% 26,456,738 29,556,050 32,573,342 30% 24,994,184 29,556,050 32,573,342 40