Why we founded Dandelion Health
The radical transformation of healthcare is underway. Artificial intelligence (AI) is already showing its potential to augment and optimize workflows across clinical, administrative, and operational use cases.
But to truly achieve the ultimate, shared goals of improved outcomes, reduced costs, and better experiences, our healthcare system needs to make some major changes - and we need to make them now.
Anyone who’s touched the health system, as an employee, patient, or caregiver, knows that fragmentation is the default state. Even finding a patient’s data in the first place isn’t easy. It’s maddeningly scattered throughout the labyrinth of legacy digital systems and old-fashioned filing cabinets that still hold sway in many settings, and it takes a huge amount of time, effort, and knowhow to bring it all together.
The risks of continuing down this path of misalignment could be catastrophic, especially when the solutions are sitting right in front of us.
If we could equip life science companies and AI developers with the right kinds of accurate and unbiased data from a wide variety of high-value sources - and do so with a level of granularity that isn’t fully captured by traditional claims and ICD-10 codes alone - we could eliminate so many barriers to better care, starting at the very beginning with drug development and clinical trials, and filtering all the way down to AI tools that power faster and more accurate decision making at the point of care.
In 2020, I joined up with my co-founders, Niyum Gandhi, Sendhil Mullainathan, and Ziad Obermeyer, to start Dandelion Health with the goal of launching this entirely new paradigm of research and clinical care: a seismic shift away from legacy strategies of piecing together bits and pieces of data that happen to be available at the time.
We would do this by empowering life science companies and AI developers with a full spectrum of data to accelerate scientific discovery and the development of novel treatments. That meant putting in the work to create an intuitive, comprehensive platform to unlock some of the industry's most valuable assets: the images, waveforms, and other unstructured data which make up over 80% of total medical data, and contain critical, untapped insights into the delivery and effectiveness of care.
We quickly came to the following conclusion: we would partner with a consortium of large, regional health systems that provide a majority of care to the communities that most accurately reflect the demographics of our national patient population. We would work with these systems to centralize the longitudinal data from all the different modalities within their IT systems so that it would be easier to access and aggregate.
This is not just about being “plumbers” laying down interoperability pipework, or middlemen negotiating for usage rights. Our mission is to take these unstructured data assets and integrate them into a rich, deep, regulatory-grade real-world data (RWD) platform that offers longitudinal context and a truly holistic portrait of the patient journey.
Only with a rich set of modalities across the entire patient journey can AI algorithms be appropriately trained, thoroughly vetted and validated, and sent out into the world with the confidence that they will improve patient care in a measurable manner, whether they’re designed to support drug development research, diagnostic activities, clinical risk prediction, or other high-value use cases.
The Dandelion Health team is passionate about this mission of empowering developers and researchers across the life sciences, medical device, and clinical care communities with the fuel they need to accelerate the entire product development lifecycle.
Pursuing our vision for equitable, trustworthy artificial intelligence
Unsurprisingly, the easiest data to access isn’t always the most representative and inclusive. We know from our own research, and the long history of unintentionally biased clinical research centered on older, wealthier, whiter, male populations, that we need to get much, much better at this foundational part of the process.
Our strategy of engaging with multiple health systems across different geographic and socioeconomic regions in the United States ensures that the algorithms built using Dandelion data are more likely to accurately represent targeted populations.
With this guiding goal at the forefront of our minds, we now have a variety of key health system partners (Sharp in California, Sanford in South Dakota, and Texas Health in Texas) - and more exciting partnerships to announce in the near future.
Already, these relationships have enabled us to collect millions of patient records that cover primary, secondary, and tertiary care delivered to individuals of varied racial and ethnic backgrounds living in urban or rural communities, across the entire socioeconomic spectrum.
Over the past four years, we have become the first company in the world to launch a free validation service for healthcare AI algorithms to ensure the safe and equitable development of clinical AI, helped AI developers find completely novel relationships from unstructured data that will aid in early patient diagnoses, and are about to launch a completely novel data library for medications.
We’re extremely proud of this progress so far, and we’re only at the beginning of our journey.
Architecting the future of clinical research with accessible, actionable data
The unparalleled depth of our real-world data is designed to bring a new degree of insight and clarity to the clinical research and AI environments.
There is so much we can achieve by using AI to enhance our limited human abilities to synthesize massive, multi-source datasets and recognize patterns within thousands or millions of data points, especially in clinical trials where speed, cost, safety, and quality are paramount.
With the right tools, we can identify promising molecules to treat complex diseases, fill clinical trials faster with more diverse and representative patients, collect more detailed and useful information during these trials, and even discover new biological relationships that might completely alter our understanding of how the body does what it does.
The end result will be an environment where precision, personalization, and trust are the foundations of innovation, enabling patients to get faster, more accurate diagnoses and access to highly targeted therapies to improve their outcomes.
Getting there will require a degree of collaboration and shared investment never before seen in the healthcare industry - and it will start by bringing together forward-thinking health systems with future-minded data companies and visionary developers to create an ecosystem where inclusive, comprehensive, longitudinal RWD is fully accessible to those who need it most.
It’s an exciting challenge and a great honor to be part of the exceptional, hardworking Dandelion Health team as we spearhead our industry’s efforts to liberate the data that will lead us into the next evolution of healthcare. As we take our next steps as a company, we invite you all to join us on the journey to more equitable, accessible, and affordable positive health outcomes for all.
