Historically, Blockchain is known to be the technology powering Bitcoin, as an open, distributed public ledger recording all the Bitcoin transactions in a secure and verifiable way, without the need for a third party to process payments. In this context, Blockchain can be considered a full history of banking transactions.

More generically, Blockchain is a huge, public, secure and decentralised datastore [8, 9] of ordered records, or events, called blocks. Each block contains a timestamp and is linked to a previous block [10]. Events can be updated by only a majority of users. Information cannot be erased. The datastore is owned by no one, is controlled by users and is not ruled by any trusted third party or central regulatory instance. In fact, trust is encoded in the protocol and maintained by the community of users.

In practice, the Blockchain architecture allows for storing proofs of existence of data. As the only proof of data is the data of proof, we believe that this is a paradigm shift for medical research methodology.

Inviolability and historicity of data are two major features of data at the functional level, “the data level”. Regarding inviolability and historicity of data, it follows that Blockchain ensures that events are tracked in their correct chronological order, which largely prevents a posteriori reconstruction analysis.

First, data integrity is ensured by the cryptographic validation of each transaction [11]. This is key to ensuring the sincerity of data — limiting data falsification, data “beautification” and in some sense data invention. Second, traceability and historicity of the data are among the core functionalities of the technology: each transaction with Blockchain is timestamped [12]. This information is publicly transparent; any user owns a copy of the proof of the time-stamped data. Figure 1 shows the complex flows of heterogeneous data and metadata that circulate in a clinical trial, implying numerous healthcare stakeholders, and all documents whose proof of existence can be stored in Blockchain. Thus, the existence of data becomes provable while the data remain confidential.

At the experience level, “the community level”, Blockchain is sometimes described as “trustless”, which can offer the right conditions for data sharing. In fact, trust is built inside the protocol. Blockchain can be considered as a “privacy-by-design” peer-to-peer infrastructure. With the level of trust it can enforce, it should be considered a path through the age of community-driven methodologies. Polls consistently show that about 80% of consumers are eager to share their medical information [27], provided privacy and security can be ensured. With the transparency of the Blockchain database — owned by no one, publicly writable by anyone and with strong crypto-oriented consistency of the database transaction — users do not need any third party to trust the system. Thus, the database opens a wide path to the data user’s control or differential privacy, data sharing and community-driven clinical study [23]: in a trusted environment, clinical research teams can “crowd-recruit” people to be enrolled in protocols with the help of community management techniques, and people can also volunteer to participate in such studies. Indeed, the Estonian e-Health authority has just implemented a Blockchain solution enabling storage of a million health records, letting patients control data access through a “Keyless Signature Infrastructure” [28, 29].

On the researcher side, data sharing is a subject of great interest and can provide many benefits. Indeed, sharing anonymised raw data, analysable datasets or a statistical analysis plan is a strengthening factor for reproducibility in science, opening clinical trials to secondary analysis or meta-analysis [24, 30–32]. Blockchain implementations can enable distributed, secure cloud data sharing. The advanced Massachusetts Institute of Technology (MIT) project Enigma, still under testing and not officially released, is most promising. Enigma’s Blockchain approach enables secure data sharing on a large scale and on a perimeter, finely controllable by the user who is sharing the data. With this kind of implementation, the data can be shared among any users or group of users, whether investigators, publishers or patients. The idea behind the technique is differential privacy: the user can fine-tune the equilibrium dose between publicly transparent data and control of the shared part between approved entities. Blockchain enables differential privacy in a secure way.

Besides archiving clinical trial-phase-compilable metadata on the Blockchain, we can also chain together different clinical trial steps so that each step depends on its predecessor. Blockchain technologies bring tools to achieve these “slicing” and “chaining” processes, called Smart Contracts, and can enforce the level transparency, traceability and control over clinical trial sequences.

According to Wikipedia, “Smart Contracts are computer protocols that facilitate, verify, or enforce the negotiation of a contract” [33], and their execution can be implemented by using cryptographic hash chains. Practically, Smart Contracts enable the validation of a step with the only condition that every preceding step has been fully validated. For example, the chain of successive blocks could verify that the designed methodology has been followed, and the material presented to publishers would consist of the publication itself and the set of blocks that constitute the Smart Contract, whose correct execution indicates proof that the study was well conducted.

Figure 1 shows that the Smart Contract represents a piece of code that holds a programmatically written contract between as many parties as needed, without any trusted third party, and that executes algorithmically according to the terms provided by the contracting parties. Examples of Smart Contracts are allowing for patient inclusion with the only condition that they have consented or for enabling data analysis with the only condition that the database is frozen. Each of the clinical trial steps detailed in the figure can be chained together in a preceding order, consolidating a transparent trial and preventing a posteriori reconstruction or beautification of data.

In a proof-of-concept experimental study, we implemented a Blockchain system to collect participant consent for a clinical trial ([34] (under review), [35]). Indeed, the US Food and Drug Administration reports that almost 10% of the trials that they monitor feature issues related to consent collection: failure to obtain written informed consent, unapproved forms, invalid consent document, failure to re-consent to a revised protocol and missing institutional review board approval to protocol changes [36, 37]. Precisely, in a fake experimental study, we timestamped each patient consent on the Blockchain and asked again for consent renewal with each revision of the protocol. We obtained a unique master document that holds, in a single data structure or piece of code called Chainscript [38], all the consent collection data, each bound to a version of revised protocol versions. In fact, these data are “hashed”, that is, formatted into a sort of cryptographic form of the real consent and protocol document data. Of importance, this master document represents a secure, robust proof of existence of the whole consent-collection process because of a strict one-to-one correspondence between hashed data and effective consent data. Also, this proof of existence can be checked on any dedicated public website.

Blockchain technology is a major opportunity for clinical research: it can help in structuring more transparent checkable methodology and, provided a set of core metadata is defined, can help check clinical trial integrity, transparently and partly algorithmically. Ultimately, Blockchain can lead to the structuration of some kind of community-driven Internet of health data, gathering researchers and patient communities, social networks and Internet of Things data flows, at a global dimension, with features of individual granularity, decentralisation and security and with transparent interactions to ensure easier and more transparent analysis.