Oldenburg University’s Health App – A Decentralized Medical Storage Solution
Remote patient monitoring
In 2020 the corona pandemic changed our lives drastically. One of the major challenges we faced were overloaded hospitals and medical centres. This led to the exhaustion of the medical resources and as a result not all patients could receive the required care. Although most patients did not need intensive care, they occupied the intensive care resources in order to be monitored. Accordingly, some patients with more severe symptoms did not receive the level of care and support they needed.
This problem can be addressed by enabling remote patient monitoring. Through remote monitoring, patients with mild symptoms can remain under hospital monitoring without occupying the resources in the medical centre. Remote monitoring can be realized with sensors that are connected to a patient's body. The sensors transmit data via Bluetooth or a direct Wi-Fi connection between a mobile phone and the medical centre. Sensor devices can automatically collect health metrics like heart rate, blood pressure, temperature, and more from patients who are not physically present in a healthcare facility. This eliminates the need for patients to travel to the hospitals, or for patients to collect data themselves.
Health data storage challenges
The common practice to store the gathered data is to use a centralized storage system. However, centralized storage requires proper maintenance policy which in most cases exceeds the resources and budget of a medical centre; therefore the storage of data is often outsourced to third-parties, based on cloud infrastructures. Healthcare companies were projected to spend $11.4 billion in 2019 on cloud computing for streamlining health data sharing and driving innovation. However, storing data on the cloud has disadvantages: Since clouds are normally run by third-party companies (like Google, Amazon, Microsoft Azur and so on); the stored data remains under their control. In the case of personal medical data; it would violate the privacy of patients. Moreover, the trace of data on the cloud is almost impossible. In the cloud each chunk of data might be stored in a different physical server, which might be in completely different geographical locations with different regulations. The personal data of citizens of one country are subjected to the regulations of that country, and might not coincide with the regulations of other countries.
Distributed Ledger Technology as a solution
Another disadvantage of cloud and self-managed servers in case of lack of proper backup strategy is the single point of failure. Implementing a reliable backup strategy in self-managed servers can be very costly for an organization. Considering the mentioned shortcomings, using distributed ledger technology (DLT) can address the limitations and increase the reliability of the service and decrease costs. Distributed ledger is an umbrella term for a set of technologies that can be used to store data in a distributed manner. Through the rest of the text the term DLT and blockchain are used interchangeably.
Using DLT's is a relatively new practice, and due to the diversity in the domain and lack of best practices, the adoption of proper DLT is not straightforward. There are various DLT's with different applications available. We can categorize existed DLT's considering their application domain into two main categories:
- Public blockchain: a blockchain that is publicly accessible and can be accessed by anyone. The transactions are verified by a public key and can be seen by anyone.
- Private blockchain: is a blockchain that is private and can only be accessed by the limited participants on the chain. Therefore it is not possible for the public to see the transactions or write on the ledger.
Considering the above definitions, we need to define the type of DLT we need for this specific use-case. On the one hand, we need the flexibility of the public blockchain to ease the access; on the other hand we need to enable only the relevant users to be able to write on the ledger.
Eliminating the options
The most popular public blockchain is Ethereum. Ethereum is a blockchain that is public and can be accessed by anyone. However the transactions on the Ethereum are based on proof of work and therefore there is a fee to verify the transactions. On Ethereum every transaction is visible to the public, as such in the current use-case where privacy is one of the main concerns, it’s not an ideal choice.
IOTA is another public DLT which can be accessed by anyone. The IOTA network Tangle is a directed acyclic graph (DAG) that is used to store data. The IOTA protocol verifies two other transactions with every transaction by default, therefore there is no fee involved in the transactions. Although IOTA is a public blockchain, and in its primary form is visible to the public, there is another underlying protocol which is built upon IOTA protocol, called IOTA Streams. IOTA Streams provides a means of channel structure which enables the publishing subscribing capabilities. If the transaction is conducted under the IOTA Streams framework, the content can be hidden to the public and accessed by only those with the pre-shared keys.
IOTA seems to be a good solution for our use case. IOTA provides fee-less transactions, as well as, the openness for the public to join; and through applying IOTA Streams transactions remain private. However, the transactions on IOTA are limited. The latest transaction state on IOTA network is about 12000 tps. This is a high amount of transactions compared to Bitcoin (7-8 transactions) or Ethereum (200 transactions). However, since we are dealing with medical devices the frequency of the transactions is crucial. In this regard, storing data directly on the distributed ledger might cause latency which might turn critical in a health-related use-case.
The ideal combination
The ideal solution in this use-case is to have: complete decentralization with immutability along with the rapid and fee-less transaction. Taken the mentioned objectives, we will need a solution to keep fee-less transactions minimized on the DLT while keeping the data stream immutably persisted. Given this, a combination of IOTA and IPFS network would be ideal. In this architecture the actual data persisted on the IPFS network which does not need transaction validation — therefore it will be faster, however, the key to access IPFS will be stored on the IOTA network.
BLING & the Health Care app
Thanks to the BLING project, the Oldenburg University is able to further explore the possible solutions for the health care app. This project is currently in its final phase of the architectural design. The aim is to launch the app before June 2023. Further developments will be shared in our next newsletters.