The stage is set for the incorporation of Blockchain and Artificial Intelligence or AI into healthcare. This has aroused great curiosity in the medical community, and many of its members continue to be puzzled about what they actually are; how they relate to each other; and what does it imply for the future.
Blockchain
Blockchain is a technique of creating and storing records and has become well known because of crypto-currencies such as Bitcoin. It is essentially a digital ledger of transactions that is replicated and distributed across a whole network of computer systems. In IT parlance these are known as nodes. Bitcoin for instance, currently has more than 40,000 nodes. In other words, the records of those participating in Bitcoin transaction are distributed among more than 40,000 computers in its system.
Each block in the chain contains a number of entries (transactions) and every time a participant conducts a fresh transaction (when an entry is made) on the blockchain, a record of that transaction is added to that participant’s ledger. This generates a database of a distributed kind (not centralized) and so the technique is known as DLT or Distributed Ledger Technology. In Blockchain DLT, all entries are recorded with immutable cryptographic signatures called hash.
Centralized databases operate with a single block (node/computer/server). Any person with access (such as administrators or hackers) can make changes to the records with just a few key strokes. In Blockchain, however, change made in the records in a single block would not be reflected in the other blocks. Such a change would be glaringly obvious and generate an alarm signal and immediate counter-measures. Also, any hackers attempting to tamper with a blockchain system would have to make changes in every block in it. In the case of Bitcoin for instance this would require over 40,000 actions almost simultaneously. Thus Blockchain is a system of storing information where it is nearly impossible to change, tamper or hack into data, for mischief or theft.
As a Blockchain grows, more and more nodes are incorporated into the system. As the numbers of nodes increase, it becomes even more difficult to hack. The system becomes increasingly more secure. This is the case with both Bitcoin and Ethereum which are continually growing as further blocks are added to the chain.
Artificial Intelligence
Artificial Intelligence or AI is all about machines that replicate the way humans analyze situations and solve problems in order to complete tasks. And to do so, AI needs a plentiful database.
AI works through a set of algorithms (instructions for step by step activity) embedded in a cognitive system of neural networks and operating through a computer device. This AI “machine” first needs to be adequately trained before it can fulfill its function. This process is called machine learning. Just as humans need repeated practice in order to master a subject, machines have to go through rounds of repeated data processing before it can function properly.
In order to operate driverless cars, recognize stop signals, or to distinguish a pedestrian from a lamppost; machines must go through a process called deep learning. This involves copious supply of data. Any corrupt data supplied at this stage will lead to faulty learning – the machine will be undependable. Being resistant to data corruption, Blockchain technology is the ideal way to store and make available such data enabling the AI to fulfill its purpose.
Thereafter, AI has the potential for automated decision-making, conducting recurring tasks, and reducing human errors required for technologies such as the internet of things (IoT), and robotics. Blockchain supplies the incorruptible data that it needs to function.
Is Blockchain Technology and AI the same?
From all that has been presented above the answer is obviously: No! They are not the same at all. However, once harnessed, they complement each other and so they can be of great benefit to every sector of human activity ranging from finance and commerce to manufacturing, travel, and even healthcare.
NOTE: AI is active; Blockchain is passive.
Features of Blockchain Technology
Figure 2 illustrates the seven features of Blockchain that make it most suitable as the database for applications involving AI.
How AI works
Figure 2B is a simplified illustration of the AI process.
Pros & Cons of Blockchain Databases
Tamper-proof: Blockchain does not allow erasure or replacement of data once recorded – thereby preventing data tampering within the network. Other data-bases are not so.
Transparency: Being decentralized, any network member can verify recorded data in the blockchain. By permitting the public to become network members, content can be placed in public domain e.g. licenses and contracts.
Traditional databases are centralized. Individuals cannot verify recorded data or its subsequent corruption.
Traceability: Blockchain creates an irreversible audit trail, allowing easy tracing of changes to the network. Audit trails do not show up in other databases.
Speed and performance: Blockchain can be slower than other database due to greater numbers of operations that are inherent in it; such as signature verification, signing transactions cryptographically, and waiting for consensus of all members before entries can be validated.
Cost: Blockchains cost more than traditional databases to install and operate.
Data modification: Modification of data once recorded requires rewriting the codes in all of the blocks. This is difficult, time-consuming and expensive.
Accordingly, organizations are advised to conduct due diligence and deep dive analysis to examine its suitability, especially in a migration to Web3 context.
Application of Blockchain-based AI in Healthcare
The potential for such systems in Healthcare is immense. It should also be noted that AI can be taught to examine the Blockchain for irregularities –
rendering the system even more secure. Some of the benefits are given below.
Record-keeping Needs
Such a system would be a one-stop shop catering to all stakeholders of the medical community with authenticated information readily at hand. It would cover doctors, nurses, technicians, clinics, hospitals, institutions, insurance providers, pharmaceutical companies, R&D set-ups, legislators, regulators, finance pundits and logistic operators and if relevant, even patients. All could be brought into one loop.
Patient-Centric Approach
- Complete medical histories are recorded and protected to be made available as and when required.
- Provide an interface with the Internet of Things (IoT) such as chip monitors enabling parameters such as blood sugar and sugar levels to be recorded
- Improve diagnostic accuracy. For difficult cases for instance, AI can even correlate a variety of factors (pandemic, endemic, symptoms, etc.) and come up with possibilities while also assessing relative probabilities
- Optimizing treatment. AI can, if asked, provide suggestions and options for the supervising doctors. Transcending purely medical data, these could consider factors such as affordability and mobility
- Interoperability. Should a patient move to a different location or care center complete details would be available for the new care team to take over
- Reduce Waiting. Registering and billing (discharge) times can be reduced drastically. Verify insurance status – flag all concerned. Negotiate clearances
- Maintain impeccable “Need-to-Know” Confidentiality
Analysis
AI can effectively mine through huge volumes of data. It can also be trained to imbibe and analyze.
- It can spot patterns and trends quickly which humans may struggle to find in a block-chain containing literally billions of bits of data
- In can generate reports in myriad formats, from summaries to exhaustive ones, depending upon the requirements of end users or decision makers
- This should be of enormous value to local and national level medical bodies, insurance agencies and such like. Such analysis can be extended to projections (example: it can provide accurate estimates for ventilators or support technicians required during a pandemic)
- Comprehensively implemented into establishments across a country, it can generate whole sets of scenario possibilities – likely prevalence of diseases, optimum strategies and so on
At national level, Cloud with its enormous data storage capacities could join the Blockchain-AI mix. AI would safeguard against any data corruption in Cloud.
Logistics
By incorporating IoT in a Blockchain-AI combo great logistical support can be provided from end to end of the supply chain.
- Manufacturers and suppliers can be informed of hospital inventories and stocks and so anticipate requirements. The status of shipments could be monitored
- Meanwhile stockists and hospitals can schedule purchases efficiently to minimize disruption or shortage. Flags would be raised and suggested alternatives provided
Research & Development
Number crunching is essential to much of R&D activity. Linking blockchains across medical communities would generate adequate data for AI to then step in. Hundreds of correlations, possibilities and hypotheses can be examined before a final conclusion.
Clinical Trials are at the core of the release of drugs and medical equipment for public use. Both protocol and statement of results can be checked for veracity. Data on test outcomes, participating numbers, patient records, and every other variable can then be authenticated for Regulatory, Statutory and other purposes. In the process scientists, connected pharmaceutical firms, doctors, policymakers and even the lay public can be confident about results.
It would enhance transparency and accountability.
Health Study Programs
Inter-linked medical blockchains systems would allow conduct of many more health study initiatives and save both money and effort for those concerned. The special parameters that these may require can be inbuilt as part of patient detail.
Health Information Exchange
Subject to the sides that propose exchanges having Blockchain-AI systems in place, such beneficial exchanges on health issues across nations, regions and the globe are now entirely possible.
Fraud & Malpractice
An audit trail is inevitable in Blockchains while inconsistent data is easily spotted by AI. Data manipulation, whether intentional or because of error raises flags in the system and so defeats its purpose.
Even unintentional errors – an incorrect procedure or administrating the wrong medicine – will remain on record. AI will point these out as lessons for the future.
Addressing Administrative Burdens
Doctors and indeed many others engaged in healthcare are burdened with administrative overload. Blockchains with AI will relieve them of much of this.
Will Blockchain and its accessories replace doctors? Driverless cars are negotiating distances without a driver. They are intelligent robots, and with advances in robotics one can expect medical intervention in the future that will be performed by robots. These will bring great precision to the task – it needs to be such to replace a human surgeon. However, just as driverless cars operate on human’s guidance, so too will medical robots require to be directed by humans. Robotics has already arrived in medicine – but human roles have not gone away. Just the modalities have shifted.
