Accelerating health solutions with discovery science

Anya is the mother of a six-year-old boy named Ethan, who is a devoted fan of baseball and the "Octonauts." She sits in a sterile hospital room, her eyes glued to her son who is struggling with a rare brain tumor. Every minute feels like a lifetime as she waits for some sort of breakthrough that might hold the key to her son's survival. But the breakthrough has already happened—it just doesn’t have an integration into care yet.

Potential discoveries hidden in research papers or labs provide no relief to someone battling a health crisis. Ethan and Anya's story highlights the "17-year gap"—a shorthand for the long wait from a medical discovery to its real-world use. This delay is not just a mere statistic, it is a difference between life and death, between prolonged suffering and a chance at health, especially in urgent health matters such as cancer, child and adolescent suicide, pandemic diseases, environmental and toxic exposures, and opioid abuse. This gap is what we need to confront and minimize, since patients, like Ethan, cannot afford to wait.

How can we accelerate research and link its results to practical use more swiftly? We can leverage the remarkable convergence of vast amounts of data and powerful technology. This enables discovery science.

What is discovery science?

Discovery science uses big data and computational methods to reveal patterns and hunch associations, create hypotheses, and generate innovation and breakthroughs. It creates spaces that are not simply inactive repositories of historic data but are dynamic and enable real-time query and application. It arms both traditionally trained researchers and health professionals with the data and technology they need to build real-time queries and quick applications and health impacts.

Discovery science has many benefits:

• It aids rapid knowledge generation, which is crucial in urgent health crises like COVID-19. It quickly identifies patterns, risk factors, and potential interventions, allowing more immediate responses.
• It uncovers unknowns. Unlike hypothesis-driven research, which can be limited by current knowledge, it reveals unexpected relationships and solves complex issues we do not understand.
• By using modern computational tools and machine learning algorithms, it focuses on analyzing large datasets, identifying subtle patterns and correlations missed by traditional statistical methods. This can lead to the identification of novel disease subtypes, predictive models, and personalized treatment strategies.
• It promotes interdisciplinary collaboration and expands participation in research by integrating datasets and expertise from various fields to encourage innovation and a holistic understanding of illness and health.
• It enables proactive identification of potential health threats, allowing for early intervention and prevention.

While discovery science is not new, it might be entering a golden era. Organizations like the Center for Data Driven Discovery in Biomedicine at Children's Hospital of Philadelphia are leading the charge in speeding up discoveries by connecting clinical and research stakeholders in real-time through innovative research methods, data, and technology. Georgetown University’s Health Informatics and Data Science program gives students a full suite of skills to enable discovery at the intersection of health data and technology. Open Science initiatives, which strive to make research processes and products readily accessible through open access, open data, and open source, are highly compatible with this approach. And the American Association for Cancer Research 2022 Progress Report celebrates discovery science’s key role in supporting precision medicine, driving innovation, and supporting new cancer treatments.

How to apply discovery science

The research community can take some practical steps to leverage the power of discovery science:

Cultivate a culture of openness and interdisciplinary collaboration: Collaborative efforts among researchers from different institutions or disciplines can speed up discovery and translation. This requires a cultural shift towards data sharing and integration, which can foster new insights and discoveries. Organizations can foster this by setting up interdisciplinary teams or centers that bring together researchers, data, and techniques from various fields like genomics, epidemiology, bioinformatics, and public health. Principles like FAIR (Findable, Accessible, Interoperable, Reusable) and federated data frameworks (Amazon Web Services’ Data Mesh is an awesome example) make this sort of collaboration possible while setting expectations and shaping behaviors. The government can also promote discovery science research by advocating for data sharing and open science policies—the National Institutes of Health Data Management and Sharing (DMS) Policy is a good example, as it mandates plans for data sharing in NIH-funded research.

Be tech-savvy: With the rise of big data in biomedical research, robust technology, data management, and analysis infrastructure are crucial. Organizations should invest in developing and maintaining secure cloud-based technology, computational environments, and advanced low-code/no-code platforms. This infrastructure is essential for handling the large and complex datasets generated in discovery-based research and can aid in the collection, analysis, and interpretation of massive amounts of data. It is also a critical component to enabling interoperability and open science.

Prioritize upskilling: The shift towards discovery-based research necessitates new skills in areas such as cloud-based technologies, bioinformatics, data science, and statistics. Organizations can support this by offering training and professional development opportunities in these areas for their researchers and staff, and the government can establish training grants and fellowships aimed at equipping researchers with skills needed for discovery science. This helps build a capable and modern workforce at the leading edge of discovery science.

Develop ethical and regulatory frameworks: The use of large datasets in discovery science brings up significant ethical and regulatory issues like data privacy, consent, and bias. Organizations can help resolve these issues by proactively addressing governance and developing comprehensive ethical and regulatory frameworks that guide data collection, storage, sharing, and use. This ensures the rights and interests of users and participants are protected and that data is FAIR and ethical.

Reconsider research funding: To further help, the government could reassess priorities to better support discovery science. It should ensure that funding is distributed fairly, avoiding preference towards individual researchers or groups, and encouraging broader participation in funded research sparks innovation by incorporating new perspectives. Promoting public-private partnerships encourages industry investment in discovery science through incentives like tax breaks or matching funds. The government can facilitate such partnerships and provide platforms for collaboration between academia and industry. And when soliciting federal contracts for research programs, alternatives to multi-year IT operations and maintenance acquisition strategies that lack incentives for breakthrough advancements and discoveries should be explored.

While discovery science presents a golden opportunity to accelerate new knowledge generation, we must remember the pursuit of novel discovery should not eclipse the validation and expansion of existing knowledge. Complex data analysis or machine learning can present interpretation challenges, which could impede the practical application of findings. Discovery science can also be resource-intensive, demanding substantial computational power and expertise, potentially diverting resources from other essential health research, and ethical, privacy, and security concerns arise when dealing with large datasets, especially those containing personal health data. The future of health research lies not in completely shifting to one approach, but in an optimal combination of both discovery and hypothesis-driven research, leveraging the strengths of each.

In Ethan's story, every moment counts. In biomedical research, every discovery counts. What matters most is ensuring that these discoveries do not languish undiscovered, but instead find their way rapidly and effectively into the hands of those who need them. Our efforts in discovery science should not just be about making breakthroughs, but also about breaking down the 17-year gap experienced by people every day. In the end, it’s not just about better healthcare, it's about time. It's about Ethan, and the many like him who cannot wait.