AMA With Dr Marcus Tan, Co-Founder, and Dr Dumitrel Loghin, Tech Lead of MediLOT
Dear MediLOT Community,
An AMA was hosted by popular crypto-influencer Shin Chan, in his community channel on 5th June 2019. We were glad to see active participation by the community and the quality of questions raised.
Here’s what MediLOT’s Co-Founder, Dr Marcus, and Tech Lead, Dr Loghin, had to say during the AMA.
Note: This AMA has been edited for clarity.
*We will like to express our gratitude to Shin Chan for hosting the AMA session*
Q: Patient data is highly sensitive and privacy is valued. While patients can grant permission to healthcare providers to access the data, will the data also be accessible by other developers or subscribers? How do you ensure that the data in the platform will not be subjected to misuse? Will it be regulated by the Singapore Government?
Dr Marcus: Patients ultimately control access permissions to their own health data — no developers or subscribers can have access to it unless granted by the patient himself. Any data that is shared is also anonymised. The SG government has a national data privacy law in place (Personal Data Protection Act) to safeguard all forms of personal and health data and MediLOT is compliant with that.
Q: In the website and whitepaper, it was mentioned that developers will be able to contribute via the platform. Can you clarify their involvement and how they will earn LOT tokens?
Dr Marcus: Developers can contribute by building apps that utilise data from the platform, and when users use these apps — a portion of the usage fee goes to the developer.
Q: How will MediLOT store the mass data on its blockchain?
Dr Loghin: We are not storing the entire data records (e.g. EHR) on the blockchain, but rather a secure hash of that data (which acts as a pointer in the hospital’s database).
Q: What do you mean by the phase “unlocking health data to its true values for patients?” How does this work ?
Dr Marcus: This would answer the above question as well — we believe that the value of health data is not fully realised for the patient or healthcare organisations. For example, we know that patients’ genomic data, especially those with hereditary diseases — would be very useful. At this point in time there is no convenient way for patients to benefit from sharing that data — sharing of data is often through a voluntary research study or trial. At the same time, healthcare institutions face challenges in trying to access such large amounts of health data. One of our aims is to address this imbalance and increase the value of health data.
Q: How do you utilize both Artificial Intelligence and Blockchain Technology in your project?
Dr Loghin: We are using a dual blockchain to handle data (EHR) and token (LOT) management, while AI is used by researchers and developers to make meaningful use of the medical data. The EHR blockchain ensures that AI models are accessing only the data they are granted access to. The LOT blockchain ensures tokens are sent to all parties involved in the system to incentivise usage (e.g. AI developers get tokens for models).
Q: Do you have any healthcare partners? Currently and in the pipeline?
Dr Marcus: We are working with several local (such as National University Hospital, the largest medical research institution in Singapore) and international partners (China hospitals in Tianjin, Hangzhou and Shenzhen). We are excited to share more in due time!
Q: Is your service already offered by a centralized entity? If so, why does decentralizing it make it better?
Dr Marcus: EHR is already provided by centralised entities all over the world. However, centralisation is prone to single points of failure and does not allow an equitable exchange of assets/data. Hacks to centralised healthcare institutions such as the NHS often lead to large amounts of sensitive data being leaked and reduces public confidence in the security of medical records.
Q: What is your competitive advantage?
Dr Marcus: We have one of the top distributed database/DLT teams in the space and offer advanced AI applications on top of the blockchain architecture. These have been extensively published and are award winning technologies in peer reviewed journals. Please refer to our references in the whitepaper and site for more information.
Q: How do you make sure that the hospital allows data to be sent out? What patient data cannot be restricted from clinicians? Since I can choose whether I would like to allow my personal data to be used for research purposes, do I get anything out of it?
Dr Marcus: For the first question, only established institutions with legacy EHR/security will be able to send top level data (which is anonymised) on a population level. Secondly, infectious diseases and publicly notifiable diseases will not be restricted from all clinicians by law.
For the last question, yes. Since patients have control of their information and can deny research companies access, research companies will have to incentivize patients to share their information through monetary/other means.
Q: Can MediLOT be used to resolve the issue of fraudulent medical insurance claims? Such cases often arise due to fake diagnosis or inflated hospital bills that patients often have to bear due to a lack of medical knowledge.
Dr Marcus: Good question! One of the aims of MediLOT will be to facilitate more seamless insurance claims that can be tracked and audited on the blockchain. We are also engaging insurers to explore their receptiveness towards such a model, and they have been very open so far.
Q: Where can end users purchase the LOT token other than from an exchange? Reason being, it is rather complex for a layman to register an account with an exchange, purchase the required LOT tokens before transferring them to the MediLOT platform. Will there be a more simplified manner to purchase the tokens?
Dr Marcus: Great question, in the future, we do plan to allow easier onboarding by having direct purchases of LOT via the app itself. I agree that only allowing exchanged based buying will hurt the onboarding process.
Q: In your model, patients have to contribute their medical records to participate while earning LOT in return, and the hospitals have to buy LOT to extract the patient’s data. I don’t understand why both parties need to go through such an extent when the patient could just visit the same hospital for treatment and have all his/her records stored in the same hospital’s database.
Dr Loghin: Patients may go (and there are many know cases where this happens in Singapore) to different hospitals for different reasons. MediLOT helps in unifying the medical record history. For this, both patients and hospitals need to participate in the system
Q: I read news about MediLOT since a year ago, what’s your achievement or any development so far?
Dr Loghin: Since last year, we have worked on the PoC and testnet and start more collaborations with hospitals and research partners.
Q: Which exchange are you guys partnership to run the IEO fundraising process? Do you have existing partnerships who has agreed to utilize your platform? Where is the team based and how will the proceeds of the IEO be used?
Dr Marcus: The exchange that MediLOT will IEO on will be announced by the exchange itself soon. Yes, MediLOT has partnered with the National University Hospital (one of the largest medical research institutions in Singapore). More contracts with other Singapore hospitals will be announced. The team has also secured a foothold in the fast-growing China healthcare market with hospitals from Tianjin, Hangzhou and Shenzhen.
The core team is based in Singapore. 45% raised will be used for systems development, 30% for Legal/ Admin/ Tax, 10% operations and lastly, 15% for business development, marketing and reserves.
Q: What is the consensus algorithm used by the project and how was it picked?
Dr Loghin: The EHR blockchain is based on Hyperledger 0.6 which uses PBFT (Practical Byzantine Fault Tolerance). PBFT achieves much better performance in (smaller) network setups compared to PoW (proof-of-work) used by Ethereum and Bitcoin.
Q: Who is your biggest competitor in the market now?
Dr Marcus: Medibloc — however they are mainly Korea-focused and aim for domestic growth. There is also not much information on the AI utilities and applications from what we know so far. Rather than seeing them as a competitor we do want to engage them further down the road to see if there are any potential ways to collaborate and enhance the patient experience for all.
Q: Who do you think will purchase LOT tokens? Which hospital in Singapore are currently utilizing your platform?
Dr Marcus: Research organisations, insurance companies and healthcare organisations who want access to aggregate amounts of large, clean datasets will transact through our token to have access to such datasets — part of this will then go to the people who have contributed such data. Currently, some of our technologies are being used in NUHS with clinical data. We aim to have large scale adoption in the next 2–3 years.
Q: How do you use Hyperledger to improve scalability?
Dr Loghin: We recently published a research paper where we showed how to scale Hyperledger using sharding and trusted execution environments (e.g. Intel SGX). On more than 1000 cloud nodes, we show that our Hyperledger system reaches >3000 tps.
Q: Do you have any existing relationships with Pharmaceutical companies?
Dr Marcus: Good point — we are currently reaching out to Pharma companies: as a side note — pharma companies are one of the biggest investors and spenders in healthcare data. Billions are spent on research to acquire valuable data that might lead to the next blockbuster drug.
Q: How will existing medical records be integrated with the MediLOT platform? Will the authorities have access to the platform?
Dr Marcus: A significant amount of work will need to be spent customising MediLOT to older specc’ed EHR systems. By law, the authorities (meaning any entity who is sanctioned by a Court) — can have access to data that is being requested.
Q: What makes MediLOT special?
Dr Marcus: We have one of the top distributed database/DLT teams in the space and offer advanced AI applications on top of the blockchain architecture. These have been extensively published and are award winning technologies in peer reviewed journals.
Q: It seems that MediLOT’s mainnet will take some time to launch, why is this so?
Dr Marcus: We need to ensure integration is complete and because health data is so important — there is quite a bit of redundancy when transiting from one system to another: we have thus set what we feel is a more realistic timeline rather than rushing for one that we may not meet satisfactorily.
Q: What is your Fiscal Policy for MediLOT tokens?
Dr Marcus:
The following mechanisms increase the value of the LOT Token:
The MediLOT Platform acts as an absorber of LOT tokens by retaining 30% of all LOT Tokens in its Fee Structure during Data Sharing. LOT Tokens held by the Platform will be held in reserve to insure against Black Swan events; they may also be burnt at the Managements discretion. In this way, the velocity of LOT tokens can also be controlled and adjusted. Healthcare Institutions and Research Organisations that stake tokens also act as reservoirs of LOT Tokens, removing them from circulation.
Q: Which coding or programming system are you guys using and any reason for choosing that particular system?
Dr Loghin: On the blockchain side, we are using Go which is employed by many other blockchain systems, including Ethereum and Hyperledger. Go is safer than traditional imperative languages such as C/C++ due to its better memory management, while being faster than higher-level languages such as Node.js, Java, etc.
Q: MediLOT will require a good amount of patient data to build its platform capabilities in terms of research purposes etc. What is your marketing strategy to achieve that?
Dr Marcus: Yes, that is a good point — to reach out to potential users we have to offer them value. One of our onboarding approaches is to offer patients a customised health analysis/report that they can get for free if they wish to use the platform — this will not include advanced analytics, which can be separately accessed via tokens. A very simple example of this would be individualised meal advice for patients who want to upload their meals onto the platform — they can currently test this out using our app http://www.foodlg.com/
Q: Why is the GitHub activity progress slowing down?
Dr Loghin: Apache Singa has reached a mature, stable version, hence there is lower activity. We are working on newer versions of Blockbench to target Hyperledger 1.0+. However, most of the benchmarking code is there. Recently, we were working on Hyperledger++, you can see the project on MediLOT’s GitHub.
Q: Can you elaborate more on the GEMINI system that you are using?
Dr Loghin: GEMINI is a system that enables AI-driven predictive analytics. It covers the entire pipeline of data analytics and AI modelling, including data cleansing, data preprocessing, model development and training, cohort analysis and data visualisation. GEMINI uses Apache Singa for AI models, and ForkBase for data storage
About MediLOT
MediLOT is a dual blockchain-based Electronic Health Records platform that allows individual patients to grant healthcare providers, research institutions and pharmaceuticals access to their encrypted medical data by indicating the storage location. This is a highly secure solution which adheres to GDPR and other stringent data privacy regulations. Patients will be incentivized and empowered to permit the usage of their data to facilitate research. For more information, pls visit:
Whitepaper: https://medilot.com/wp/wp-content/uploads/2018/06/white-paper-en.pdf
Summary: https://medilot.com/wp/wp-content/uploads/2018/06/summary-en.pdf
Intro Video: https://youtu.be/DbyOi53OhSU
POC Video: https://youtu.be/tW7iS9rp_ac
