Improving data collection and modeling to accelerate predictive medicine efforts
Analytics and AI drive more accurate predictive analyses to advance clinical development.
Faster, more effective drug development through innovative data collection and analysis
Dompé farmaceutici uses SAS® for predictive analytics and quantitative disease modeling
Which patients and populations will respond best to which future drugs? And how can pharmaceutical companies accelerate the processes they use to predict positive patient outcomes?
It all starts with data. As a science-driven biopharmaceutical company, Dompé works to develop innovative drugs for unmet medical needs, and SAS technology is transforming the way Dompé collects and models data.
We spoke with Andrea Beccari, Head of R&D Platforms and Services, and Anna Fava, Senior Software Engineer, at Dompé to get their thoughts on how the company is navigating this evolving landscape.
By doing predictive analysis earlier in the cycle, we can speed our drug development process and deliver solutions to patients faster. Andrea Beccari Head of R&D Platforms and Services Dompé farmaceutici
These are exciting times for the pharmaceutical industry, but they can also be tricky to navigate. What are the primary challenges you are facing today?
Right now, the pharma sector is going through a transition phase. One reason relates to the fact that several new players are entering the digital health services market – think of things like smartwatches and health monitoring apps. Another factor is the pressing need to accelerate innovation in our field.
The industry has accomplished as much as possible with current knowledge and technology. The challenge now is around complex and systemic pathologies, such as those related to oncology, cardiology, immunology and metabolic diseases. For these pathologies, conventional approaches are beginning to show signs of ineffectiveness. For these diseases, there is no single effector (molecule that binds to a protein or enzyme by modifying and regulating its activity) – so we need to act systemically to achieve great results.
To date, we have not had the knowledge or modeling capabilities to handle this level of complexity. An underlying reason is that data has never been collected appropriately. Therefore, our first step involves collecting the right data at the right time.
Traditional medicine is based on trial and error, with 97% of molecules submitted to clinical trials failing. It is expensive to develop drugs based on the 3% of molecule trials that succeed. Further, tracking the population’s responses to those drugs has typically been random.
Today, medicine is moving toward a predictive approach to help us understand in advance which population responds best to which drug. The challenge, therefore, is to collect data with this predictive perspective in mind and use it for quantitative disease modeling.
We’ve heard a lot lately about Exscalate4Cov, the landmark supercomputing project designed to fight the coronavirus. In coordinating this ambitious project, Dompé pulled together 18 institutions from seven European countries. This seems to highlight the value of industry consortia and collaboration. How are these new ecosystems changing Dompé’s approach today, and how might they shape your future?
The future is linked to open data, and where there is an open data approach, there is also an ecosystem concept.
For Dompé, a medium-sized company in the pharmaceutical market, it is crucial for us to be part of such a strategic and collaborative environment. Participating in a broader group gives us access to organizations with many different capacities. Such endeavors help us achieve our full potential, as we did with the Exscalate4Cov project.
Today, we are involved in several European projects. For example, a new initiative that starts soon will seek to repurpose (evaluate an alternative use of drugs already in use) all drugs at the European level. We plan to participate in about 30 projects across Europe with organizations that rely on our public research system to develop solutions. Often, these organizations have clinical evidence but lack the resources to achieve clinical validation and deliver solutions to patients. An industry consortium helps overcome these hurdles.
Relying on a broad ecosystem has significantly influenced our business model. We believe that growth in pharmaceuticals depends on the creation of strategic partnerships around the world. This entails sharing and integrating technologies and development models.
With this in mind, we recently partnered with a British biopharmaceutical startup, Engitix. We gave them access to our Exscalate platform to help them identify new treatments for fibrosis and certain types of cancer. Engitix provided anonymized patient data to initiate the research. This alliance will hopefully lead to a faster development and commercialization of new drugs. By joining forces, we aim to provide new solutions for patients whose needs are presently unmet.
Dompé farmaceutici – Facts & Figures
units manufactured annually
In this scenario, how is your use of the SAS data analytics platform evolving?
SAS Analytics solutions are tightly integrated with the evolution of our drug design and development processes. These processes must account for safety – so we now use predictive analytics to predict toxicity and side effects.
By doing predictive analysis earlier in the cycle, we can speed our drug development process and deliver solutions to patients faster.
By incorporating analytics, including artificial intelligence techniques, we can predict the toxicity and side effects of proposed therapies to avoid adverse events. One of the most significant benefits we’ve achieved with SAS is being able to rapidly put new analytical models into production. This accelerates traditionally laborious processes and enables us to deliver new drugs to market faster.
Validating models is one of the fundamental tasks SAS helps us achieve. When defining models for predictive analytics, there is always a risk of inconsistency between model development and application. In other words, there’s a chance that the model may not “make sense” in terms of how it will be used in the real world.
During the COVID-19 pandemic, there was a lot of confusion because predictions were generated from models that were not validated. There were no controls. Yet, multiple predictions were shared with the scientific community, physicians and the public.
The ability to validate models and use controls for simulated data analysis are indispensable building blocks of the pharmaceutical products we bring to market and the analytics ecosystem that supports their development. SAS provides these capabilities, accompanied by excellent support from technical staff. This is exactly what we need to pursue our mission of strengthening and accelerating research to deliver more effective therapies.
Die in diesem Artikel dargestellten Ergebnisse sind auf die hier beschriebenen besonderen Situationen, Geschäftsmodelle, Dateneingaben und Computerumgebungen zugeschnitten. Die Erfahrung jedes SAS-Kunden ist aufgrund geschäftlicher und technischer Variablen einzigartig, und alle Aussagen sind als untypisch anzusehen. Die tatsächlichen Einsparungen, Ergebnisse und Leistungsmerkmale hängen von den individuellen Kundenkonfigurationen und -bedingungen ab. SAS übernimmt keine Garantie oder Zusicherung, dass jeder Kunde ähnliche Ergebnisse erzielt. Die einzigen Garantien für SAS-Produkte und -Dienstleistungen sind diejenigen, die in den ausdrücklichen Garantieerklärungen in der schriftlichen Vereinbarung für solche Produkte und Dienstleistungen aufgeführt sind. Nichts von dem hier Veröffentlichten ist als zusätzliche Garantie auszulegen. Kunden haben ihre Erfolgsgeschichten mit SAS im Rahmen eines vertraglich vereinbarten Austauschs oder einer Zusammenfassung zum Projekterfolg im Anschluss an einen erfolgreichen Abschluss einer Implementierung von SAS-Software kommuniziert. Marken- und Produktnamen sind Markenzeichen der jeweiligen Unternehmen.