Spinal Muscular Atrophy: The High Costs of Innovative Therapies for Rare Diseases

Background: Despite numbers of prescriptions lled remaining constant, expenditures for drugs are rising. Among others, this is caused by high prices for orphan drugs and advanced therapy medicinal products (ATMPs). Despite attempts by policymakers to intervene, the increasing use of these therapies poses numerous challenges on the health care system. Results: Using data from the University Hospital Heidelberg, we found that the division of pediatric neurology is experiencing a strong increase in drug costs, caused by two pharmaceuticals for the treatment of spinal muscular atrophy (SMA), Nusinersen and Onasemnogene Abeparvovec. To put this nding in a broader context, we conducted a survey of 41 German SMA treatment centers revealing a general lack of human and infrastructure resources for therapy and follow-up care, which demonstrates insucient reimbursement of treatment. To improve structural conditions, we propose that disease-independent registries for medication surveillance that comply with the rules of the European Union are needed as well as the development of a fair funding model. Conclusion: Innovative forms of therapy require a critical discussion and concise regulation of treatment application and follow-up reimbursement as well as international industry-independent registry work. Based on the previously described model of “evidence-based dynamic pricing” by the Techniker Krankenkasse, we propose a revised model which offers an internationally scalable solution to meet these challenges.


Background
The pharmaceutical market has been experiencing rising expenditure on prescription drugs over the past decades although the number of prescriptions lled has remained constant. (1)(2)(3)(4) This is mainly attributed to high costs and increased availability of patented drugs (mainly oncology-related medication), including orphan drugs (OD) and advanced therapy medicinal products (ATMPs), a group which includes gene therapies, somatic cell therapies and tissue engineered products.(5) As these therapy forms often represent the rst causal treatment approach, they are commonly referred to as "innovative therapies". In 2019, expenses for drugs in the statutory health insurance system in Germany exceeded € 45 billion for the rst time, and while OD accounted for only 0.05 % of all prescribed daily doses, they caused 10 % of the total expenditures with a drug market share expected to further increase. (1,6) This will have a strong impact on the health care system, as patients with rare diseases will increasingly be entitled to complex forms of therapy that can only be performed in specialized treatment centers. Thus, university hospitals with longstanding experience in treating rare diseases will have to cope with increased indirect personnel and infrastructure costs.
Since the early 1990s, the European Union has set incentives for the development of orphan drugs. (7,8) Pharmaceutical companies have since begun to target a market that would otherwise be unattractive due to the low number of cases, which has ultimately resulted in high costs for individual preparations.
National legislators are therefore trying to nd ways to improve regulation of drug prices. In Germany, the conditional marketing authorization stipulated in the 2011 German Pharmaceutical Market Restructuring Act (Arzneimittelmarktneuordnungsgesetz, or AMNOG) is the linchpin of the Federal Joint Committee (G-BA, the German main regulatory institution for legally binding decisions in healthcare) for price negotiations of the statutory health insurance companies and the pharmaceutical industry.(9, 10) However, during the rst 12 months after drug approval, prices are freely chosen by the manufacturers, only afterwards negotiations are possible. Nonetheless, years after its introduction, the AMNOG has improved the ratio of negotiated prices and the clinical bene t of evaluated compounds. (11,12) Recently, the G-BA formulated the requirement of an "application-accompanying data collection", whose legal basis was established in Germany in the so-called Law for More Safety in the Supply of Medicines. (13,14) This is intended to help close the existing evidence gap in the approval of ODs and ATMPs, since their e cacy and safety is proven in the context of the centralized approval for the European Economic Area.(7) As a result, clinical trials are performed on small groups of patients with only sparse long-term data on clinical e cacy and safety. (15) A sharp increase in medication expenditure was observed with the approval of two drugs for the treatment of spinal muscular atrophy (SMA). Nusinersen, which was approved in 2017, followed by Onasemnogene Abeparvovec, a gene therapy for the treatment of SMA. Onasemnogene Abeparvovec received a conditional approval based on the results of a phase 1/2 study with as little as 15 SMA-patients as compared to historical control cases and preliminary nonpublished data of an international phase 3 study with 55 individuals suffering from SMA. (16)(17)(18) With costs of € 1.945 million per single application, Onasemnogen Abeparvovec is the world's most expensive drug to date. The high costs associated with such innovative therapies require standards in research and risk assessment that are at least equivalent to those of non-orphan drugs. Regardless of their therapeutic potential, ODs and ATMPs generate major challenges for national healthcare systems, healthcare providers and insurance companies. Many issues regarding their use, associated (organizational and personnel) costs as well as the reimbursement of follow-up care remain to be addressed.
In order to address the challenge of rising drug costs and to help close the evidence gap regarding e cacy and safety of innovative therapies, we have developed a cost model in collaboration with the Techniker Krankenkasse, Germany's largest health insurance company. This model proposes a structural framework for nancing of medication-associated costs and patient registries, as well as follow-up care.

Results
From 2010 to 2020, inpatient departments of the Heidelberg University Hospital incurred average annual costs of € 48.7 million (SD = € ± 14.4 million) for drugs. In 2020, the highest cost ever of € 88.1 million was recorded. Table 1 shows the annual expenditure on drugs of the ve departments with the highest annual costs for drugs at Heidelberg University Hospital. A detailed investigation identi ed in particular two drugs for the treatment of SMA as the main cost drivers: Nusinersen, responsible for an increase of drug costs from 2017 on and Onasemnogene Abeparvovec with a single dose applied in 2019 and a total of 10 doses applied in 2020. In 2019, 92 doses of Nusinersen were applied at the Centre for Children and Adolescent Medicine Heidelberg with a total expenditure of € 8.44 million. The average drug cost of a single application was € 91,739.13. In the same year, a rst patient was treated with Onasemnogene Abeparvovec, the net drug cost amounted to € 1.945 million. (19) The conditional marketing authorization of Onasemnogene Abeparvovec by the EMA was granted in May 2020 which has led to a rapid increase in demand for the drug. During 2020, a total of 10 patients were treated with Onasemnogene Abeparvovec, resulting in total drug spending worth € 19.45 million plus import VAT.
To put these ndings in a broader context, we concluded a survey of 41 German SMA treatment centers caring for a total of 1,097 patients. Within the patient cohort, 646 (59 %) were younger than 18 years of age at the time of the survey. Based on our survey, in Germany, 92 patients are expected to qualify for start of treatment with Onasemnogen Abeparvovec during the following 12 months, confronting insurance companies with total drug costs of € 178.94 million plus VAT. Compared to other drug groups, such as TNF-alpha inhibitors (Adalimumab: € 468.1 million) or modern anticoagulants (Apixaban: € 475.3 million) (20), this amount appears to be small, but it must be emphasized that the number of patients treated with Onasemnogen Abeparvovec is much lower and it remains unclear whether recurrent drug administration will be necessary for individual patients. According to the survey, the number of patients with Nusinersen therapy will decrease from 867 during the most recent 12 months to 761 in the coming 12 months (Supplementary Table 1).
All 41 centers state that the application of innovative SMA therapies in standard care creates new challenges, 28 (68.3%) centers expect challenges of medical and non-medical character. A total of 82.9 % (n = 34) report an additional workload of more than at least 5 hours per week. New tasks will primarily arise in the areas of case management, pre-and post-clinical care, and administration of reimbursement or invoicing. The need for additional human resources (even more so than infrastructure or nancial resources) is thus perceived as the biggest challenge in the overall process of applying innovative therapies ( Figure 2).
Those responsible for the SMA treatment centers see the need for standardization of processes. While both, the indication process as well as the pharmacy process were reported to be well regulated, a strong need for improvement and more precise standardization of follow-up care was reported by 39 (95.1 %) respondents. At the time of investigation, only 25 (61 %) of the centers reported to have standard operating procedures (SOPs) related to SMA treatment. While reporting an increase of workload, 80.5 % (n = 33) of the respondents do not consider the reimbursement for the healthcare providers of innovative therapies to be appropriate. With regard to follow-up care, 87.8 % (n = 36) stated that compensation was insu cient, and 95.2 % (n = 39) saw that the reason for this is in the existing outpatient reimbursement system. When respondents were given the opportunity to prioritize different areas of optimization, the compensation for aftercare was ranked highest.
All respondents indicated that structured documentation of the outcome of innovative therapies should be carried out in international registries. A majority of 94.4% (n = 39) stated that this should happen independently of the pharmaceutical industry. There was also broad consensus concerning nancing of clinical trial registries, where 80.5 % (n = 33) of the respondents were in favor of a fund to be administered in trust by the Federal Joint Committee, into which the pharmaceutical company would have to pay an amount based on the price of the drug after approval. Only a small proportion of those surveyed saw responsibility for structured documentation with the insurance companies (12.2 %, n = 5) or with the health-care providers (7.3 %, n = 3). With few exceptions, (21) international recommendations for the implementation of registry work are missing so far.

Discussion
The present analysis uses the example of Heidelberg University Hospital to demonstrate a considerable annual increase over the last years in drug costs in the eld of pediatric and adolescent neurologylargely attributable to orphan drugs and ATMPs. As more complex therapies, such as ATMPs will soon reach market maturity, adequate standards of organization and structured pre-and post-treatment pathways for patients and caregivers are needed. The current reimbursement model exposes treatment centers to a potentially threatening liquidity problem if drug prices keep rising. Costs for medication are pre-funded by the treatment center and can only be claimed from the insurance company weeks to months later. Such interest-free loans to a pharmaceutical company by means of public resources is highly questionable. This is especially critical due to the increasing numbers of patients being treated with innovative therapies with high direct and indirect costs.
To overcome challenges of direct costs, direct billing between the insurance companies and the pharmaceutical industry has the potential to allocate processes not directly related to patient care to the primarily responsible actors, as seen with Zynteglo ® , a gene therapy drug for the therapy of Beta-Thalassemia. (22) The incidence of SMA is estimated at 1:7,500 for Germany. It represents the most common genetic cause of infant and childhood mortality.(23, 24) Most patients suffer from subtype 1 (Werdnig-Hoffmann) and are diagnosed in infancy. Our survey of 41 SMA treatment centers revealed that personnel resources for case management and administration are lacking. These and the associated indirect costs arising through the application of innovative forms of therapy must be re nanced. In 2019, the Techniker Krankenkasse introduced a model of dynamic evidence pricing (DEP),(25) which serves to ensure the sustainable nancing of innovative forms of therapy, in particular gene therapies. Most gene therapies are used in a single-stage setting, so that the total therapy costs do not accrue over a longer period of time, meaning that the associated cost-risk lies entirely with the insurance companies. Considering that data availability at the time of approval is low for many ATMPs and their use places high demands on staff and clinicians, we feel a need to address this issue. Under the DEP model, the Federal Joint Committee (G-BA) decides 6 months prior to approval whether a drug will go through the traditional AMNOG process or be assessed using the DEP. The DEP requires mandatory inclusion of patients in industry-independent registries to close the evidence gaps that exist at the time of approval. In this regard, the introduction of so-called "application-accompanying data collection" was recently initiated for the German pharmaceutical market. (13) The G-BA can request the pharmaceutical manufacturer to conduct an extended data collection after (conditional) marketing authorization. In doing so, the G-BA determines the requirements for the type, duration and scope of data collection and evaluations, including the patient-relevant endpoints to be recorded, the methodology of the data collection, as well as the possibility of the evaluation of said data by the pharmaceutical company. (13,26) However, numerous legal and economic questions remain unanswered.
Since neither the initial DEP model nor the "application-accompanying data collection" clari es the funding for registry structures or nancing of costly outpatient follow-up, we propose a revised DEP (rDEP) to meet this challenge. Similar to the original DEP, the G-BA decides 6 months before market approval whether a pharmaceutical product will be assessed via the traditional route or with the rDEP method. At this stage, the relevant medical societies are mandated to develop a treatment pathway and criteria for quality assurance based on the clinical trial data available to date (compare (27)). The G-BA also determines which registry, study modality, and monitoring period will be used for subsequent generation of evidence. In order to clearly regulate nancing of such registry structures, we propose to initiate a trust fund via means of the pharmaceutical companies managed by the G-BA. The collection of the agreed amount of data to monitor diagnostic and therapeutic processes in combination with valid patient outcomes enables the strati cation of treatment success. (28) The legal framework of the European Union has de ned clear requirements for such registries. (7,29) In the rDEP model, cross-disease registries are to be maintained by academic centers and independent of the pharmaceutical industry based on the FAIR principles. Data must be Findable, Accessible, Interoperable and Reusable to ensure the e ciency of the registry structures. (30,31) Figure 3 provides an overview of the rDEP and depicts the timeline from approval submission to the iterative process of pricing.
A maximum price for the therapeutic product in the rst year of use should be based on European average prices for comparable therapies and determined by means of health technology assessment methods (described in detail elsewhere (32,33)). Thus, the adjustment of the cost of medication based on registrybased data analysis of clinical endpoints leads to a dynamic process of price regulation that takes into account the actual positive (or lack of, or even negative) effect, drug safety, and the availability of alternative treatment options. This will ultimately ensure long-term availability of high-quality drugs for improved patient care with innovative therapies and ensure the cost-effectiveness of the health care system.

Conclusions
The example of SMA is suitable as a blueprint for the rapidly increasing number of ATMPs and innovative orphan drugs expected in the coming years. First experience shows that existing organizational structures are not su ciently prepared for the associated challenges. However, the health care system with its multiple stakeholders (clinicians, insurance companies, pharmaceutical companies) must prepare for upcoming (cost-)relevant therapies by creating appropriate structures and regulations. The revised dynamic evidence price model presented here serves as a basis for discussion and offers a possible way to address the rising direct and indirect costs of innovative drugs. To put this cost model in simple terms, the higher the documented e cacy of a drug, the higher the cost of that drug may be. In case of low evidence or missing data, the maximum price for the drug can be set lower. This will secure high quality care in the long term, bene ting treatment centers but ultimately the patients as well.

Data sources
Total quantities and costs for prescriptions and ready-to-use drugs of the University Pharmacy of the Heidelberg University Hospital were analyzed for the time period of 2009 -2020. Data from the Center for Pediatric and Adolescent Medicine were assigned to individual departments and the main driving forces behind drug costs were identi ed. A structured questionnaire survey distributed to all German SMA treatment centers was used to investigate the following dimensions: 1) workload before and after the introduction of innovative forms of therapies, 2) standardization of procedures, and 3) availability of registries. Questionnaire items were designed as 5 point Likert scales, implemented into the Limesurvey tool as an anonymous survey and then sent via e-mail to the participating centers of the SMArtCARE registry (34) (URL https://www.smartcare.de/, n = 49) and the list of "Neuromuscular Centers" of the German Society for Muscular Diseases (DGM). A total of 41 questionnaires were completed, covering 1,097 SMA patients (see Supplementary Table 1) corresponding approximately to the SMA prevalence of 1:100,000 expected for Germany. (35) The complete questionnaire can be found in the supplementary material.

Limitations
The National Center for Tumor Diseases (NCT) is located at Heidelberg University Hospital, where patients receive highly speci c and individualized oncological therapy. These treatments generate extremely high costs totaling € 36.46 million per year (corresponding to 37.6 % of total expenditure on drugs at the University Hospital Heidelberg). Due to the lack of comparability to other sites, NCT data were excluded from the analysis. The true cost of medicines in Germany is partly subject to individually negotiated discount rates between the health care provider and the health insurance company, which are highly con dential. In our assessment we only consider the direct costs caused by drugs. Costs for the implementation of the therapy, such as personnel costs and costs for the maintenance of the infrastructure, could not be considered due to their complexity. Our cost analysis can therefore only provide an approximation of the true costs, but it nevertheless expresses a clearly discernible trend in the application of innovative therapies.