2023: The challenge of Rare Diseases

Rare is not so rare. If no more than five in 10,000 people are affected, a disease is considered rare in the EU. However, in Germany alone, around four million people live with a rare disease; in the EU as a whole, it is estimated that around 30 million people are affected. They all have the same right to the quality, safety and effectiveness of drugs and further forms of treatment as people with other diseases.

Generating scientific evidence on the diagnosis and treatment of rare diseases is associated with particular challenges. At IQWiG's 18th Autumn Symposium, experts from various areas of the health care system described these challenges and how they can be overcome.

On 24 and 25 November 2023 in Cologne, almost 200 participants from self-help groups, the health care system, research institutions and industry discussed what patients with rare diseases need and what is currently missing to achieve good care. A further 132 participants took part in the symposium online.

People with rare diseases should not be overlooked

“Common things are common and rare things are rare.” This succinct answer from a senior physician to his question as to whether the patient might be suffering from a certain rare disease is still remembered today by IQWiG Director Thomas Kaiser when he thinks back to his time as a junior doctor. He warns: "Although this approach reflects the pragmatism of everyday medical practice, it is insufficient to make a correct diagnosis and should no longer be used by medical professionals if people with rare diseases are not to be overlooked." The key to this is early and robust evidence on the diagnosis and treatment of rare diseases. However, the (fictitious) added benefit that has so far been formally attributed in Germany to drugs for the treatment of rare diseases (orphan drugs) at the time of approval has the opposite effect: "For 70 percent of orphan drugs, the 'non-quantifiable added benefit' only certifies one thing: we do not know whether these drugs are better, equally good or even worse than the treatment options available to date."

IQWiG's Autumn Symposium 2023 focused on how to improve awareness of rare diseases and evidence for good treatment decisions in seven presentations and included a debate on the right cost balance for drugs.

The odyssey to the correct diagnosis

Searching, fighting, hoping: These are the characteristics of the lives of people affected by rare diseases. Mirjam Mann, Christine Mundlos and Michael Scholz from the German Alliance of Chronic Rare Diseases (ACHSE for short) thus introduced the main topic. As the father of a seriously ill child, Michael Scholz described the odyssey leading up to the correct diagnosis, which suddenly and fundamentally changed his life. Although he works in the social insurance sector, he found himself faced with almost inscrutable requirements: Where and when does which application have to be made by whom - and with which technical terms - in order to actually get the right aids for everyday life?

The three key demands of ACHSE are therefore:

• Accelerate the diagnosis! The technology and linked structures are there, but the planning perspective and awareness are still lacking.
• Improve the evidence! There is a lack of specialized staff, and disease registries as an important basis for more knowledge are uncoordinated and underfunded.
• Expand national networking! Disease-specific outpatient clinics are lacking and the reliable networking of patient organizations with doctors and hospitals needs to be expanded. However, there is hope for projects at the national and EU level, which will start in 2024.

Finally, someone is listening to me!

Unlike human staff in everyday medical practice, artificial intelligence (AI) is under no time pressure: AI can take the time to take a detailed medical history on screen in order to find out more from patients about their rare disease by asking 80 questions. What AI can contribute to the diagnosis of rare diseases was demonstrated by the practical examples given by Martin Hirsch, who works at the University of Marburg on the interfaces between biomedicine and information technology. "But don't stop thinking when you're working with artificial intelligence!" he appealed, because it's all about probability models and correlations. There is still often a lack of data sets to train the digital networks and generate reliable evidence. In future, however, digital symptom checkers could act as small "AI labs" at home to continuously collect important data and thus provide more comprehensive information for health care - for patients and medical professionals alike.

Needs are infinite, but resources are limited

Health economists recommend health economic evaluations (HEEs) as an effective tool for dealing with this dilemma. Julian Witte, an expert in economic modelling and data-based storytelling, took a look at the cost structure in the pharmaceutical market: patented drugs and orphan drugs are the primary cost drivers here - with an upward trend because there will be more products and therefore more prescriptions in the future. Although the privileges of orphan drugs have been reduced in Germany, with the lowering of the arbitrary sales threshold to €30 million, affordability remains a challenge in the long term. "We spend more money every year than in the previous year - and what do we get for it?" The HEEs based on real-world data could provide evidence on this question and for cost-effective treatment decisions. So far, Germany is one of the few countries in Europe that does not use such supplementary instruments.

Can I prescribe it?

In a debate on orphan drugs, Martin Hug, a pharmacist from Freiburg University Hospital (and also a father of a child with a rare disease) and Wolf-Dieter Ludwig from the German Medical Association's Drug Commission discussed whether the current high expenditure on orphan drugs is justifiable and affordable in the long term.

Sales of original drugs have doubled in the last ten years, and those of orphan drugs have even tripled, Hug explained in his introduction. At the same time, there have been drastic price increases, such as for the gene therapy onasemnogene abeparvovec (Zolgensma) for spinal muscular atrophy: although the manufacturer's selling price has fallen in the past three years, it is still very high at €1.65 million per single-use syringe (2020: €2.25 million). However, drug development is only worthwhile for a company if the revenue is greater than the costs in the end. It sometimes takes many years for the operating profit to exceed the high costs for production, marketing and research & development, as Hug explained using an example.

Wolf-Dieter Ludwig criticized the high prices charged by manufacturers for certain drugs. The cytogenetic markers in oncology in particular are leading to a tendency to "fillet" patient groups and indications in order to achieve higher profits with the orphan drug status due to the privileges associated with it. Martin Hug did not disagree with this. However, he also sees a high medical need, especially for rare diseases, where the disease burden is very high.

The debaters agreed on one point: we urgently need to clarify how we can generate reliable evidence for orphan drugs on a regular basis, ideally in parallel with drug approval. Thomas Kaiser concluded: "Artificial control arms shouldn’t be demonized, but in which cases can I rely on them - or not? This also needs to be clarified, as does the question of why direct comparative randomized controlled trials (RCTs) are not carried out instead. Far fewer study participants are needed for this type of study than for non-RCTs."

Conditional approval comes with requirements

Martina Schüßler-Lenz from the Paul Ehrlich Institute answered questions on drug approval in the field of rare diseases from a regulatory perspective and with a particular focus on gene therapeutics such as Zolgensma. These so-called advanced therapy medicinal products (ATMPs), which target the underlying genetic cause, offer new treatment options for patients with rare diseases. The clinical development of ATMPs approved in the EU mainly takes place in the USA and follows international and EU guidelines.

"The mode of action of gene therapeutics and the causal attribution of the therapeutic effect are properties that enable positive benefit-risk assessments on the basis of single-arm clinical trials," said Schüler-Lenz. However, in the case of conditional approval - commonly used for ATMPs - further data collection is usually necessary after approval to obtain more information on side effects and, in some cases, on the effectiveness of the new drugs. A "normal" approval status is only granted once these additional data are available.

"Drug manufacturers must therefore consider how they will generate the evidence to fulfil any approval requirements even before orphan drugs are approved," the approval expert noted - "because there are specific requirements associated with conditional approval." She predicts that we will not see any platform studies for this purpose in the next five years: "This is because the registry landscape is currently still too poorly networked." However, she believes that EU-wide patient access to ATMPs as well as high-quality EU-wide registries would be very suitable for generating knowledge for health care and supporting research.

It then became clear from the subsequent discussion that there is fundamental methodological agreement between the regulatory authorities and HTA agencies such as IQWiG in important areas. There is also a preference for RCTs in orphan drug approval. Thomas Kaiser explained: "However, since the comparison with the current treatment standard is always relevant for the benefit assessment, a relevant health care question often remains unanswered despite sufficient data being available for drug approval: Which treatment is better - the new one or the one already there?"

The problem is not just prevalence. It is also incidence!

"If no new people are included in the study, we cannot effectively apply specially developed methods for this study," said Ralf-Dieter Hilgers, an expert in medical statistics at RWTH Aachen University who investigates integrated design and the analysis of small populations in clinical studies. He provided methodological solutions for generating robust evidence for rare diseases.

"Without control arms, the effect of a drug is artificially assumed, but if there is no evidence, decisions are more likely to be wrong than right on this basis." Hilger's conclusion: randomization is also possible in studies with rare diseases with small sample sizes, but should be carried out as early as possible. The respective randomization model is decisive. In the planning phase of the study, it should be clarified how bias can best be avoided. The extent of bias must be clarified during the data analysis phase before the results can be interpreted. When asked, Hilgers added that the various published methods were unfortunately hardly used by study designers or regulatory authorities - a situation that should be changed urgently, as otherwise the generation of better evidence for rare diseases would be hindered.

Use health care data, supported by AI

Thomas Klopstock from the LMU in Munich, where he works on translational research approaches for rare neurodegenerative diseases, showed how patient registries and natural history studies can be used to generate evidence for rare diseases: "Even for mitochondrial diseases with great variability, numerous research goals can be clarified with registries: Defining the phenotypic and genotypic spectrum and capturing disease progression, evaluating appropriate clinical and patient-reported outcomes, identifying the appropriate patient population for clinical trials and establishing an external control group for single-arm clinical trials.” His plea: "I think we need to use even more data from health care for research, in the sense of knowledge-generating research. And ultimately, we will also have to make greater use of AI support. With thousands of patients and thousands of parameters, new analysis methods are needed.”

RCTs are possible, even in registries

Lutz Nährlich, a scientist and paediatrician at the University of Giessen, presented what patient registries can contribute to (randomized) controlled trials. "We have to think globally when it comes to rare diseases," he demanded and asked: "Why don't we create the evidence for approval decisions and benefit assessments and pharmacovigilance for long-term monitoring together before approval?" Centralized registries for all parties involved would be a good solution - and are already emerging in the European medicines agencies network strategy to 2025. However, the potential of registries can only be fully realized if the data sets are coordinated and the parties involved are networked: authorities, drug manufacturers, registry operators and study centres. Nährlich: "Then it will also be possible to conduct approval studies in such platforms in the future."

At the end of the two-day symposium, Thomas Kaiser concluded: "In my opinion, it was only through the path to evidence shown yesterday and today - from the real individual case to study methodology - that it became understandable for whom and why we are doing all this! We should expand our two-phase model 'before approval / after approval': There is the important period parallel to the approval process, i.e. while the approval studies are being evaluated by the regulatory authority. We must use this period to initiate high-quality studies before market access. We can then ensure that evidence for good and appropriate treatment decisions is available in a timely manner, without delaying the approval itself."

His closing words: "If we can get better evidence for just two or three rare diseases through this symposium, we will have achieved a lot."