By Caitlin Kramer, Research Analyst, LSN
As Life Science Nation’s 10th RESI Conference approaches, we are proud to offer a glimpse into the topic of a new panel which will debut in Boston on September 13th: Big Data in Healthcare. This panel will feature the insight of five investors actively seeking companies that are advancing big data solutions to problems in healthcare and medicine.
The paradigm of digital information is still relatively new, and techniques for dealing with massive quantities of it – Big Data – are still in their early stages. LSN asked investors for their thoughts on the space, and their responses are below.
What is exciting about big data? What role do you see it playing in the future of healthcare?
One of the areas I continue to see active investor interest in is companies working with patient data. There’s a lot of excitement about innovative solutions that help patients and practitioners securely manage and share medical records and data while still maintaining regulatory compliance. This extends into the insurance industry as well.
I’ve also found the application of big data-native concepts like computer vision, machine learning, and natural language processing to the medical field incredibly interesting. Examples include technologies that help doctors diagnose diseases earlier based on a programmatic analysis of medical imagery, or a cross-correlation of 200 million patient records looking for subtle trends, or a mental health patient conversing with a digital psychologist AI to help set some context before scheduling a meeting with a licensed doctor.
Another investor said:
I believe that within my lifetime, it will be considered malpractice for doctors to diagnose without using computational aid. Systems biology is too complex to be held in the brain, and AI will become a routine, standard part of medicine.
With so much excitement about the potential big data has to transform healthcare, the space is prone to hype. Entrepreneurs should keep in mind the unique challenges of working in the space, especially when speaking with investors. This led us to ask:
What are the biggest hurdles you see to using big data in healthcare?
An investor we spoke with countered with another observant question:
Why hasn’t there been a big medical breakthrough or discovery yet, despite billions of dollars already invested in big data AI and machine learning?
He continued on to discuss why:
Most big data projects indiscriminately mine data without considering uncertainty. These can be uncertainties in measurement, biological variability, or patient comorbidity. When I ask CEOs of big data companies simple questions about how they plan to deal with mismatches in ontology, or error propagation, I am often met with a blank stare. Medical science, done well, requires setting up an experiment, making measurements, accounting for uncertainty in measurement, and propagating that uncertainty. To effectively do that at this stage, we need to first focus on developing algorithms that work on small data with constraints, and a scope confined to the clinical question.
Another core component to developing big data algorithms for use in healthcare is recognizing that the needs of the algorithms differ from those applied in other fields. A failure to correctly classify or categorize a piece of information used in a marketing or financial setting can be corrected later, while the failure to distinguish between a benign and malignant tumor image can cost a life when applied in medicine. The mentality of software development needs to understand the consequences of error.
Important insights and opinions such as these will ultimately shape how big data manifests a presence in healthcare in the coming years. We look forward to announcing the panelists for Big Data in Healthcare in the coming weeks, and to hearing more investors speak about their experiences investing in this new and exciting future for medical care at RESI 10.