The USV portfolio network consists of 67 active companies with over 7,000 employees across the US, Canada and Europe. We believe in using the power of networks to help our portfolio companies build better businesses through peer to peer learning, external network relationships, and shared resources. On average we host 60 portfolio events each year and, in 2017 we are on track to host nearly 80 in total.
Within the USV portfolio, we have several companies working in the healthcare space. Based on their feedback and recommendations we organized an intimate session with leaders from the USV portfolio, USV partners and staff, and external industry leaders. We held this discussion at our office and called it “Building the Healthcare Stack”.
This “Hacking Healthcare” event was the brainchild of USV CEOs who wanted to connect with other healthcare organizations and professionals that are spearheading initiatives and development within this industry. They wanted to debate high-level areas such as telemedicine as well as dive into granular challenges facing open source health data and patient care.
In USV fashion we organized this as an “unconference” style conversation and segmented the day into 4 high-level topics.
Our goal for this session was for participants to come away with thoughtful perspectives on a variety of areas. We chose specific, complex topics that would seed provocative and unfiltered discussions and I have chosen to highlight 3 of the top takeaways.
1. Healthcare companies are investing too much money in walled gardens
When thinking about medical facilities (hospitals, doctor’s offices, research centers, etc.) every single one of them controls their own data — what they collect, how it is stored, cleaned, and analyzed. There is no universal system that allows patient data to connect and talk to each other.
One reason why walled, healthcare gardens exist and in fact flourish is that there are no financial incentives to open them up. As a result, healthcare organizations only see the world through their own institutional myopia. This nearsighted mentality can have lasting effects on the staff, patient experience and culture. An example of a set standard are the Press Ganey scores which are calculated from post-appointment patient surveys. Some facilities use these scores to award bonuses which can cause tension and wrongfully placed incentives for the physician. This can also shift their priorities from treating someone as effectively as possible to focusing on personal, soft skills and excellent bedside manner. As we see more digital health companies emerging in the space, we may need to find ways to blend and ultimately open up some of these datasets.
2. How can we open source healthcare data?
In software, open source data is anything made publicly accessible for others to use or modify. But what does this mean when talking about personal data in healthcare, such as patient records and diagnoses?
One view in favor of open sourcing health data is that asynchronous medical care is absolutely required to improve the efficiency of the medical industry. In theory, this could allow a patient to visit any clinic in the world, see a physician, and get a diagnosis utilizing their existing data. This diagnosis could be based on the patient’s medical history as well as aggregated data from other doctors and facilities.
One point made against open source health data is that since clinical data is not being captured in structured data formats, there would need to be major, additional regulation around the input and “cleanliness” of the data points. It would also be extremely time consuming and costly to implement this open source database. Finally, if this data is open to anyone, then patients would of course have access as well. Is this in fact useful to you as an individual or rather is it too much clutter and detail to comprehend in a relevant way? If it is accessible to anyone, it could be incredibly susceptible to abuse and malpractice.
3. More data ≠ better data.
Some of the participants challenged the group on whether healthcare data should be considered “big data.” One opinion was that big data is dependent on the quality of the data, not quantity. In other words, if you feed bad data into the system, the AI and data output will produce bad results. Deep learning has struggled to work in healthcare as, contrary to other industries, there is not a lot of big data to be analyzed yet. Let’s take Facebook for example — if you think about facial recognition technology, Facebook services millions of people through their tagging feature. There is not that kind of system built yet for health systems.
Another crucial issue worth mentioning is that there is a tremendous amount of bias in datasets. When aggregating data in a machine learning context it is difficult to deduce evidence based conclusions. The methods used to analyze and collect data vary from one provider to another and every facility wants to use their data to help control their own standard of care. A suggested solution to this issue is to build a business model focused on aggregating data on a patient level. This would allow facilities to recognize the behaviors of individuals (habits, socializations, family complexity, etc.) and collect rich, unbiased data.
Staying in line with the idea that less data is better, what is the least amount of data that could be collected to achieve optimal results? For example, if a pregnant woman is asked if she previously had a premature birth (Yes or No response), based on her answer she could receive more targeted treatment and precautions to reduce complications and medical bills. Rather than attempting to tackle big healthcare data, one could focus on “small data for small outcomes”. This would result in more precise patient data through light-touch interactions which would lead to better facility / patient habits.
_ _ _
I wish I could say that we unraveled answers to some of these complex questions. Instead, the conversation between 30 industry leaders was rich, unfiltered and provocative — in our eyes, a success. Everyone was willing to share critical developments, milestones and roadblocks. Industry giants heard the voices of mighty, lean startups and vice versa. Arguments and compromise ensued, relationships were built and partnerships were seeded.
With some of the brightest minds in medicine, technology, insurance, non-profit, and academia, some leaders around the room had been confronting these issues for decades — others for less than a year. Out of all of those brilliant minds, not one person could pinpoint the solution to one of these healthcare challenges. Personally, I left with hope that big strides are being made in this industry. On behalf of USV, our hope is that we can continue to facilitate open and transparent conversations like this across the country and world.”
Everyone knows the big Web 2.0 companies use hundreds of data points to determine which ad we might prefer. And yet — in the deathmatch against disease, we reduce human health to single variables.
Granted, this has partially been due to immature technology and infrastructure; after all, an assembly line of PhDs can only annotate the genome so quickly. There is also a hard limit on a human’s ability to find patterns within the noise.
In the last couple of years however, a few trends have reshaped the landscape for startups working at the intersection of computer science and biology:
1) the hardware layer of the genomics stack has been commoditized,
2) the cost of genomic sequencing has fallen below the threshold required for routine reads,
3) data storage is effectively free, and
4) sophisticated computational tools, including deep learning, have matured, allowing us to apply strategies that were not possible before
Once in a while, there is an inflection point that completely changes the rules of the game. We saw this in the early 2000s, for example, when suddenly you didn’t need a big check to build your own servers and infrastructure, just to get a website up and running.
What this shift enables, is a new generation of biotechnology companies very distinct from its predecessors, with characteristics not unlike the software and machine learning companies we are familiar with.
The characteristics that make software startups so appealing — that you can test your idea cheaply, that you can de-risk early, that you can scale quickly, etc — will be found in this new generation of biology companies also. In fact, many of these startups should really be thought of as machine learning/software companies with domain knowledge in biology. Just as we saw an explosion of web startups running many experiments at a low cost in the mid 2000s, we expect to see a similar phenomenon in the biology space.
And clearly genomics is a big-data problem — arguably the biggest today. The thing is, most people think of the genome as a static tell-all dataset. In reality, even your somatic dna changes at an astonishing rate; in fact, we can predict your age, within around a 5 year confidence interval, from your genome. That would not be possible if your genome was static. So we need to reframe the genome as a dynamic real-time data stream of what is happening in the body. Then of course, we also need to couple longitudinal genomic datasets with time series biomarker data before we can use our new tools to understand human health a little better.
We have been excited to meet teams that are fully leveraging the promise of this new era. A couple weeks ago, for example, we met cancer diagnostic startup Freenome, which uses cell-free dna from liquid biopsies to detect cancer at an early stage. If that sounds scary, at a very high level, it is just a machine learning categorization algorithm. What is exciting is that they have essentially taken an agnostic approach to the problem. Healthcare is a notoriously slow-moving industry, but imagine that in the future, new findings will simply be incorporated through a software update.
Beyond disease diagnosis, we have seen startups working in agriculture genomics, drug response, and even designing a new genomic programming language, that have all captivated our imagination.
It will be tempting at times to dismiss these startups as naive; after all, many of them are tackling highly complex problems that generations of scientists have given blood sweat and tears to, only to make tiny contributions. And indeed, there are many technical and commercial bottlenecks we have yet to overcome (next post). However, we have seen impressive real-world results, and we are excited about what is to come.”
I'm Jennifer, one of the new analysts on the Investment Team at Union Square Ventures!
Before joining USV, I studied statistics and computational biology at Harvard, and then worked at a hedge fund after graduation. I'm particularly interested in new financial/banking paradigms, blockchain + decentralized networks, and genomics.
Growing up, many people around me were entrepreneurs. I was completely fascinated by the idea that you could make your own path, especially in a culture where adherence to the rules was glorified. And whether they were successful or tremendously unsuccessful, they all tried to leverage technology to make the world a better place — and I really admired that.
I'm very excited to get to know all of you — please say hello at @jml_campbell or email me at jennifer AT usv DOT com”