“You might have a sports car, but if it is not burning the tarmac, it is no good”.
I will come to this analogy later but in today’s trying circumstances, everyone is missing something, be it friends, family, or going out on a retreat. People are making plans for the post-COVID world, but they’re not sure whether travelling or people whom they meet will be safe from COVID and other such peripatetic questions.
We are in an uncertain state because we don’t have enough knowledge about our surroundings. Anyone we are meeting these days in a grocery store might be carrying the infection, but we are not aware, simply because we don’t have enough information about the people. It is true that authorities have come up with apps which help us in this regard, but it is very reactive and dependent on how accurately people declare the information.
Data insights is an answer, but not the complete one
Assume a scenario where health organizations are aware of underlying health conditions of individuals over a long period of time and know which kind of diseases they are susceptible to or are currently being treated for. If there is a health score every individual can carry based on their health condition, social habits, location, climate etc., that might help the authorities make better decisions around lockdowns, self-care, and isolation. This information must be shared over a period of time in such a way that it is easily consumed and the insights are relevant to the concerned authorities. If the information is more dependent on how an individual declares it, its authenticity can be questioned.
In the case of COVID-19, by using Machine Learning (ML), researchers can predict the patients who are likely to develop severe respiratory distress and the risk of mortality even if they are not showing symptoms. Based on these insights, authorities can mark an area unsafe/safe based on the population health classification. All this depends on what level of patient data is available for research and at present, the lack of data is the biggest challenge across the globe.
Machine learning in drug development
In the case of the ebola virus, researchers trained the Bayesian machine learning models with viral pseudo-type entry assay and Ebola replication data to help determine the scoring process. Data science provides a large range of algorithms such as support vector machines, decision trees, random forest classifiers, and many more. For instance, Random forest classifier or boosted regression trees is being used in H7N9 influenza to design scoring functions for binding affinity of specific proteins.
Predicting the outcome of treatment
Using machine learning (ML) algorithms, predictions can be made around the outcome of treatment on different individuals in terms of survival rate based on geography, previous health conditions, social habits and more. Similar predictions were made for cancer immunotherapy by researchers.
Using social networks for prediction
The efforts to cure are directly proportional to how many people are self-declaring their symptoms. Most of the people do not declare their symptoms on official portals but show signs on social media networks by posting status updates. Using ML algorithms, social media interactions can be interpreted, and correlation can be drawn against the COVID-19 symptoms; this will help determine the spread of the disease in real time.
In short, the richer the data, the more insights can be driven from it and there are multiple ML algorithms available to reach a decision.
Real time information dissemination compliments data insights.
Data insights are good, but what if they are lying in a silo where only a handful of people have access. How quickly an insight reaches the concerned stakeholders can make a huge difference. Take vaccine development for instance, it takes time and effort. Hundreds of pharma organizations globally are trying to develop sequences of COVID-19 but the way they share data amongst themselves take days due to various operational hurdles such as slow approval cycle and siloed systems. If we can share data real time to concerned stakeholders, that can really turn the tables for us.
That is where technology like Blockchain comes into play. If the pharma organizations are on a blockchain network, they can share the information in real time with everyone or with a select few, depending on the network design. If approvals are needed to share such information, blockchain consensus mechanisms are available, so that no false or unauthorized information goes through.
All this is possible using distributed ledger technology (DLT) in blockchain. Once a transaction (record) is uploaded on a blockchain network, this goes for approval in real time to all the concerned stakeholders. Once a transaction is approved, it gets recorded onto the ledger in a cryptographically secure manner and shared with all the organizations on the blockchain network, who are concerned about such information. Transactions remain immutable and nobody can alter the transaction, which makes it ideal for audits and progress trail.
Similarly, hospitals can share the critical equipment information amongst themselves on blockchain so that each of them is aware of the inventory pool in real-time and can take help in case these equipment’s needs to be channelized amongst them.
Build insights fast but share them even faster
If we try to relate to the analogy, data insights are like sports cars, but blockchain becomes a driver which can make these insights reach faster to the intended audience. Blockchain not only provides real-time dissemination of insights but also provides data security which is highly important when it comes to patient health information (PHI). As a society, we need to find a balance between data privacy and data transparency so that when our data is required for the greater good like fighting the COVID-19 crisis, it should be available in a safe and secured way and fast enough so that we are ahead in the road to recovery.