Our team are big fans of rare disease events, and the discussions that happen there! In an attempt to replicate that, and get the gears greased ahead of CRDN’s Rare Summit 2021, we asked you to send us any questions that you think the people at Healx could help answer about the world of rare disease treatment development.
And you delivered. Thank you to everyone who got in touch – a lot of areas were raised, from data sources for artificial intelligence to clinical trials.
Here are the answers to some of the top questions from members of the Healx team!
“Clinical trials have existed for decades. Why have there been so little efficiency gains in that time? It’s like each clinical trial is the first one to ever take place.”
“In a sense, every clinical trial really is the first one to take place. When you are testing a drug in people, you have to demonstrate that this specific drug is safe and effective for these specific patients in this specific indication. Therefore, you need to treat enough patients over a long enough period of time to demonstrate safety and efficacy. Patient safety is of course paramount, and that’s why it takes some time.
Furthermore, it is becoming more and more a requirement to demonstrate the value of new drugs so that payers can decide whether they should be paid for with public money and this adds an additional level of evidence that needs to be produced.
But there have been efficiency gains in the clinical trials process (for example, electronic data capture has completely replaced paper over the past 20 years) and new advances in technology are improving the process all the time.”
Dr Anthony Hall, Chief Medical Officer
“We’re currently identifying collaborative R&D projects with our researchers and clinicians. Which types of data are important when using Artificial Intelligence (AI) in drug discovery?”
“AI requires large volumes of complex data and disease-specific datasets to produce the best drug predictions. Our AI platform integrates knowledge on rare diseases, compounds and targets. Much of the data comes from public biomedical databases which we enrich with information extracted from the scientific literature and other sources using Natural Language Processing (NLP), a process through which computers are taught to quickly read and process huge amounts of digital knowledge. Our team of expert curators validate and ensure our data is high quality and has as broad coverage as possible. Invaluable additional data comes from our patient group and academic partners, and from commissioned datasets. The types of data that we use includes gene expression and other “omics” data types, high-throughput screening data, clinical information, preclinical model information, to name just a few.”
Jane Brennan, Associate Director of Scientific Curation
“I’ve tried to work with you, and other organisations, to find treatments for my rare disease of interest, but we are often told that we’re not ready. Why is that?”
“Different AI companies have different platforms, expertise and interests that align best to a subset of diseases. Before we initiate work on a disease project, our cross-functional team evaluates each disease for key preclinical, clinical, commercial and data feasibility to ensure that they provide the potential for us to have a reasonable opportunity to achieve the preclinical validation and clinical trial success that will lead to patient impact. This includes access to, and understanding of, data and other knowledge related to:
● the disease biology, including the natural history and progression of the disease
● relevant in vitro and in vivo models of the disease, and how they relate to the human disease
● measurable clinical endpoints and/or biomarkers
● patient and clinician availability to participate in, and benefit from, clinical trials
Having all of this knowledge upfront gives us the best possible chance to find novel therapies, validate our AI-predicted drug candidates, and take them towards clinical trials. If some of these key elements are not yet available for your rare disease of interest, the best thing for us both to do is work with your research, clinical community and patient group to expand these resources. This is our approach – other companies might have a different process.”
Clara Tang, Alliance Strategy Manager
“I’m interested in artificial intelligence and how it is helping uncover new treatments, but I don’t really see how me and my patient group can do anything to help.”
“Patient groups are really important when it comes to improving artificial intelligence, because they can generate and contribute some incredibly valuable data. Patient groups can help by encouraging their members – and their families, including those unaffected – to provide tissue and cell samples for genetic sequencing, and then asking researchers to make their data available, either publicly or privately to industry partners. They can also help to gather together small but valuable datasets through surveys and questionnaires to patients, carers and healthcare professionals. Important topics include: symptom types and frequency in the patient population, and which symptoms are considered the most important unmet needs for the patients; treatments tried and used successfully, or unsuccessfully, by individuals and which are used in combination; which symptoms they are affected by each therapy and what is the efficacy. Such real world data and evidence provides us and our AI with a far sharper view of the true nature of each disease or disorder. This helps us better tailor our AI tools, such as our Knowledge Graph, thereby improving drug predictions – and not just for the disorder in question, but also all of those that connect to it in the larger network.”
Richard Gibson, Principal Curator
“Are there ways for Patient Groups to be more proactive in managing the funding and organization of disease research, and the ownership and management of the data, knowledge and intellectual property that is created as research progresses?”
“Yes! Patient groups fund a lot of research, inspire researchers and clinicians to get involved in their disease, and they are the experts in what is most important to patients and their carers and loved ones. When patient groups drive the research agenda and have ownership in the outcomes funded by the patient group, they can make a lot of progress more quickly.
There are patient groups who have already done this. A story that comes to mind is that of the Multiple Myeloma Research Foundation, which many years ago created the Multiple Myeloma Research Consortium to establish coordination, collaboration, control or management over many key research and clinical issues, including patents, publications, best practices, patient record management, etc. Another success story is the Foundation for Angelman Syndrome Therapeutics (FAST) that controls a lot of the models used to validate research in Angelman Syndrome, and actually set up a biotech to create an antisense oligonucleotide that is already in patient trials.”
Bruce Bloom, Chief Collaboration Officer