The blockchain technology that underpins the virtual currency Bitcoin has been applied to clinical trial documents, to see if it can stop them being altered to make it look as if medicines are more effective than they are.
According to a paper published on the F1000Research platform, Greg Irving from the University of Cambridge, and John Holden, a GP, converted the documents from a randomised clinical trial and cardiovascular diabetes and ethanol so as to take advantage of the blockchain infrastructure.
Under their system, the original clinical protocol was given a unique digital signature, determined by the documents text using an online tool called the SHA256 Calculator. This was converted into a public Bitcoin key, using Strongcoin, which is another online tool.
The coin was then ‘spent’ to create a transaction that is time-stamped and recorded as a blockchain. Anyone who wants to see the protocol can generate a new Bitcoin key, using the text of the document to which they have access.
The two clinicians behind the paper argue that this addresses a particular problem with clinical trials, which is that researchers can change trial protocols to analyse health outcomes until a positive result is found – and then only report the positive findings.
Although this is not supposed to happen by international treaty, enforcement is difficult. Using the Bitcoin and blockchain approach, however, it is relatively simple to see whether changes have been made to the text of the protocol documents, since if the keys differ this must have happened.
In a statement issued by F1000Research, Dr Irving said: “Trust in scientific research has been diminished by evidence that some data is being manipulated. The declaration of Helsinki states that every clinical trial must be registered in a publicly accessible database before recruitment of the first subject.
“Yet problems, such as differences between pre-specified and reported outcomes, persist.” He argued that blockchain registration could have prevented a particularly “infamous” case, Study 329, which reported that an antidepressant was effective in adolescents.
In fact, analysis of the eight original health measures had shown that the drug was ineffective, but researchers then looked at another 19 outcomes, found four positive effectives, and only reported on these four – giving a false impression of the drug’s value.
There has been growing interest in the distributed ledger technology of the blockchain in recent years – ironically just as its original use, for Bitcoin, has come under pressure because of the computing power required to generate and solve the cryptographic puzzles that form its keys.
Earlier this year, the government’s chief scientific officer, Sir Mark Walport, argued that it could be widely used in government, including in health.
He argued that some of the attractions of distributed ledgers are that they free information from single institutions, make it easier to validate changes, and improve security since “a cyber-attack would have to attack all the copies simultaneously to be successful.”
Despite this, those involved in the F1000Research paper says this is the first time it has been applied to a clinical trial.
Amy Price of Oxford University, who reviewed the paper, said: “The approach could also be used for randomised controlled trials and a whole range of observational and experimental studies where registers are needed but do not currently exist.”