The single biggest bottleneck in the drug development process both in terms of time and cost, is to successfully take a drug through clinical trials. Most clinical trials suffer from a multitude of problems such as under recruitment, patient incompliance and dropout, lack of diversity, geographical inaccessibility and failure to meet adequate clinical endpoints.
Thankfully, major advances in Big Data, A.I. and recent Congressional legislation has allowed for radical innovation in the way clinical trials are run, and the data we use to evaluate both safety and efficacy. While the gold standard for evaluating new drugs and therapies has always been Randomized Controlled Trials, we now have a wide range of options to choose from, that allow us to accelerate this process. The FDA now allows for use of Real World Evidence which are insights mined from data outside of conventional clinical trials. This includes patient health records, self-reported outcomes, lab reports, wearable data and data from other sensors in the clinic and at home.
Companies like Science37 are now building “siteless trials” that decentralize the process of conducting trials which were limited to major medical centers. Others like Deep6 AI are optimizing the patient recruitment process by matching patients to the best suited clinical trials and fast tracking enrollment, a process that would otherwise takes months to complete. Lastly, companies like Unlearn actually reduces the need for patient enrollment by creating synthetic control arms for clinical trials. Computational models simulate diseases based on historical data and create a digital twin in place of an actual person. This ensures all trial participants get treatment and some benefit, versus some participants being subject to placebo as is the current norm. As the biotech industry also moves its focus towards developing cell and gene therapies, all of the above innovations also make it easier to conduct trials for rare or orphan diseases, where the patient population is sparse to begin with.
How does patient and data privacy play a role with advancements in healthcare and is there a risk that privacy suffers in the age of digitization?
A lot of the seminal advances in Artificial Intelligence are built on the backbone of machine learning techniques which mine large amounts of training data. This data is often sensitive, confidential and vulnerable for being exploited. This obviously becomes problematic in a healthcare setting where data privacy is of utmost importance, both ethically and legally. If data is the new oil, we had more than a few incidents of “oil spills”. Just in the past few months, there have been huge data breaches exposing millions of medical images, along with incidents where Big Tech now has access to protected health information (PHI) without full disclosure or guardrails in place.
This constant tension between utility and privacy is now being addressed by startups which enable privacy preserving Machine Learning through novel approaches. These include techniques such as Differential Privacy which adds noise to the metadata making it difficult to identify individuals, or Federated Learning which decentralizes the process of model training. A.I. models are brought into the hospital environment for training, allowing for faster deployment, lower latency while ensuring privacy. Another hot area of research is in the field of Fully Homomorphic Encryption which allows third party users to access encrypted data and run computations for gathering insights without having the decryption key or more importantly, even needing it for data analysis. Startups like Duality, Enveil and Cryptonumerics are all building sophisticated encryption techniques with applications in various industries including healthcare.
Which startups do you think have the most potential to unlock the future of healthcare?
From a thesis driven investment approach, two broad trends to think about are precision medicine and reducing clinical inefficiencies. The first trend is largely driven by startups leveraging genomic data, piggybacking on drastic reductions in sequencing costs and using advances in CRISPR, single cell genomics etc. Recent FDA approvals for cell and gene therapy products has kept this area front and center for most biotech investors. I am personally interested in companies which leverage genomic data and generate computationally derived insights from it. A good example for this is Variant Bio. They are taking a contrarian approach to therapy development focusing on cases who are medical outliers instead of discarding them as rare anomalies in the dataset. This allows for locating genetic traits, understanding unexpected evolutionary adaptations and then build new drugs for a large disease population or repurpose old ones based on those insights.
For the second trend of reducing costs and inefficiencies, AI driven products are proving to be useful tools. Multiple startups like Suki are building “digital scribes” which can assist physicians in note taking and charting. This has been reported to be the single biggest reason for physician burnout in recent times. Automating this tedious process allows physicians to spend more time speaking with patients versus typing into their Electronic Health Records systems. Other startups like Ellipsis Health are taking AI for speech one step further and developing machine language models which can smartly comprehend not just what is being spoken but also how it’s being spoken. This ability to understand tone, context and sentiment has huge implications in behavioral health and will help assess people with anxiety, depression or other conditions.
You’ve focused on the genomic landscape and where the investment opportunities lie as the industry moves beyond just a focus on reading, or sequencing, the DNA. Can you elaborate on where you think healthcare-focused VCs and investors should be looking, with regard to genomics, in the next five years?
Broadly speaking, the first wave of genomic innovation was to read or sequence. This led to the creation of the first big genomic powerhouse which is Illumina. The second big advance after read was to write and edit the DNA. We have companies like Twist Biosciences, Molecular Assembly which can synthetically write small to medium DNA strands for various research applications. Similarly, the development of CRISPR enabled gene editing techniques created a whole new set of startups such as Mammoth Biosciences and Editas Medicine developing new diagnostic and therapeutic applications.
The next logical evolution to read, write and edit is to now combine all these techniques to record. Various academic groups are using novel techniques to enable cells to record either changes in their intracellular state or their response to environmental challenges within their own DNA. Other groups are working on creating DNA barcodes that can tag specific intracellular proteins, receptors etc. Such functionality allows to convert your own cells into molecular recording devices which map both temporal and spatial changes. The possibilities for using such techniques in biomedical applications are endless. You could also use this for tracking entry of pathogens in cell population of interest, activation of various downstream cascades (coagulation pathway), assess cyclical or episodical events (menstrual health) or assess cell response of drugs or toxins.
What is the investor sentiment currently and how is the investment strategy changing in the pandemic?
The COVID pandemic took everyone by surprise and even the venture ecosystem has been impacted by the sudden shift in gears. In the short term two trends are very obvious. The investing mood is somber and there are fewer deals happening. The other short term fallout is that investments that are still happening are for later stage deals. Early stage companies will have to be patient and stoic in their fundraising efforts.
Beyond the obvious rise in telemedicine and efforts towards drug and vaccine development, a couple of trends are emerging for the longer term. On the Provider side, most small to medium size hospitals are already in the red and will not have the capacity to withstand the pandemic pain beyond a few months. There will be increased acquisition activity and consolidation with larger hospital networks taking over distressed units. The pandemic also woefully exposed the under preparedness of most Western countries in making testing kits, PPE, and essential medical equipment available. Political pressures will reorganize supply chain and capacity building away from foreign nations. Both of these longer term trends will shape the way investments are made over the next few months.
Is there anything else we should discuss?
Another upcoming area I find fascinating is the digitization of smell, or creation of a “digital nose”. This is enabled by the convergence of software meets biology. Our achievements in computer science over the past few decades have been in successfully training computers to identify objects, scenarios and execute tasks based on training via sight and sound. As humans we also heavily rely on smell to comprehend and navigate our environment.
A slew of new companies are now developing technologies where we can for the first time, objectively measure things in our environment and build applications on top of “smell as a service.” These are poised to have far reaching impact in a wide variety of industries ranging from defense, food and beverage, agriculture and obviously healthcare. Companies like Owlstone Medical and Foodmarble are developing consumer grade “breath biopsy” device which measures chemicals in your breath, to analyze digestion and your gut health and even other conditions like cancer. Other startups like Koniko and Aromyx are building processors with biosensors that actually mimic olfactory receptors. Building such solutions opens up a whole new dimension beyond what computers can currently do. This will create entirely new market opportunities for both companies and investors.
The views of the interviewee of this article do not necessarily represent the views of First Republic Bank.
This document is for information purposes only and is not intended as an offer or solicitation, or as the basis for any contract to purchase or sell any security, or other instrument, or to enter into or arrange any type of transaction as a consequence of any information contained herein.
Investors should seek financial advice regarding the appropriateness of investing in any securities, other investment or investment strategies discussed or recommended in this report and should understand that statements regarding future prospects may not be realized.
This information is governed by our Terms and Conditions of Use.
