Is Blockchain the key to a breakthrough in cancer research?
What’s being called ‘Precision medicine’ is a revolutionary approach powered by the accumulation and analysis of massive amounts of health and biologic data, and is being touted as the equivalent a ‘cancer moonshot’. However, notwithstanding initial advances, the practice is currently not in use for a variety of reasons, most of which can be traced back to the problems of large scale, expensive and centralized projects that are subject to a changing political climate.
There are critical failure points that remain unaddressed and are responsible for the current lack of progress:
Arguably, the problems start with siloed data that inhibit discoveries and exist for multiple reasons:
- Data normalization: most hospitals and leading centers struggle to meaningfully organize the genetic and phenotypic data of their own patients in a fashion for informing clinical decision making.
- Legal: medical centers and practitioners are reluctant to share medical data, with liability issues a prime concern in case private medical records become public
- Competitive advantage: medical data is often perceived as providing a competitive advantage for the institutions and researchers who generated them, presenting a dilemma of choosing between maintaining an advantage or enriching a common data set that advances cancer research
The progress at the All of Us NIH project illustrate difficulties and the time and effort to achieve simple things such as registering people and collecting data:
“The hard part is scheduling future appointments and managing logistics. At most partner sites and at most of the mobile engagement unit’s exhibit stops, enrollees haven’t been able to simply fill out paperwork and walk into a room across the street to get their blood drawn.[i]
Placing medical records and patient data onto a centralized database automatically implies vulnerability, a prime consideration for any large scale project. As Eric Dishman reports, the more data is collected from people, the more it becomes a target for hackers:
“Certainly privacy and trust is huge for us and we developed a set of principles with industry, with the website, with cybersecurity groups at the White House… we will be a huge target, and it’s one of the things I lose sleep about…[ii]
According to the Ponemon Institute, 91 percent of health care organizations have experienced at least one data breach, costing more than $2 million on average per organization[iii]. The American Action Forum estimated that medical breaches have cost the U.S. health care system more than $50 billion since 2009.[iv] Medical records are extremely valuable to thieves, with such data sold for an average of $363 per record,[v] which is much higher than for credit card data.
Lack of tools and data science
In addition to data sharing and exchange, the analysis of petabyte level data volumes with hundreds of attributes working in conjunction is impossible without deep neural network (AI) technologies. As data scientists know, AI loves consuming data and it can provide a continually self-learning infrastructure with real-time knowledge output. AI enables a system that is preventative and predictive, a key factor that is missing in current research methodologies which largely rely on human interpretation
How is Blockchain an answer?
A new way to store and share personal data
Blockchain does not just store data in a distributed and encrypted form, but it does it on a sequential chain in which each block contains a cryptographic hash of the block before it in the chain. This links the blocks and creates a decentralised transaction ledger.
The Bockchain ledger can be made visible or public by the entity (person or organization) who has uploaded their data to a Blockchain address, thereby providing a transparent view into the historical sequence of facts and events, or it can be permissioned for privacy reasons.
Whether public or private, the ledger allows the stored data to be verified as consistent, and because it is decentralised it provides resistance to external attacks and malicious actors within the system.
Blockhain and AI
There is a ‘divide’ between genotype (genome sequence of an individual) and phenotype (actual physical characteristics such as age, race, weight) which prevents to associate genetic data and disease.
Deep Learning, commonly known as a form of AI, is able to bridge the genotype-phenotype divide by aggregating an exponentially growing amount of data for analysis of the complex relationships and biological processes between the two.
Since most phenotypes are influenced by both a person’s genotype and by the unique circumstances in which a person has lived, including everything that happens, the nature of how Deep Learning works enables a continuous learning and discovery of variables that influence health and the onset of disease.
The need for Deep Learning vs. human analysis is further highlighted by the amounts and types of personal data that are made available in our evolving world. Electronic health records, live streamed data from IoT (Internet of Things) devices (wearables, health apps, tracking sensors), social media, cell phone GPS signals and more are generating a deluge of data – all of which when combined and analyzed provide the means for discovery.
How would it work?
Personal Medical Lockboxes (PMXs) on the Bockchain can provide a solution for collecting, analyzing and sharing medical data and further gives people ownership of their health data. Since sharing data is in the hands of the data owner, cross-border and privacy regulations no longer apply. People can decide to share part or all of their data with an external party based on their interests and circumstances.
Medical R&D and health care industry participants could offer crypto tokens in exchange for limited views into the Lockbox, thereby incentivizing people to place as much of their health data on their PMXs. The more people expose their data to a global market, the more value they receive and the more AI agents are able to uncover unseen relationships that contribute to disease.
With insights obtained to guide research and develop treatments, researchers could access a global registry with a new data model for discoveries and new prevention methods. Data can be shared across different ecosystems (e.g. medical practitioners + fitness centers + nutrionists), or with individual participants on a global scale ecosystem. The more data collected from more sources all over the world the more predictive the models become as a result of machine learning.
This is part 1 of how Blockchain can revolutionize how we manage our health. On part 2, I will talk more on the role of AI agents to enable discoveries.