If you’re not careful with the practical issues presented by your trial master file, you risk losing yourself in the ontological ones. The former run into the latter at the best of times. What else but a TMF needs to be complete and ever-changing, done before it starts, and unfinished until five years past its end?

According to the EMA’s guidance, the TMF must be set up before a clinical trial commences and continually guarantee its legitimacy and integrity by allowing what it calls “operational staff as well as monitors, auditors and inspectors to evaluate compliance with the protocol, the trial’s safe conduct and the quality of the data obtained”. It should be possible to reconstruct the entire proceedings of a trial simply by consulting the TMF. From a regulatory and good clinical practice perspective, a trial without one may as well not have happened at all.

“In a way, the TMF is the most interesting part of the medical product,” explains Francis Crawley, the executive director of the GCP Alliance, “because it’s about the data. If you talk about the truth and reality of a medicinal product, then you’re talking about a TMF.”

That makes effective sponsor oversight of the TMF as vital as it is difficult. It actually consists of a sponsor and an investigator TMF – though aspects of the former can be delegated to the latter – and is as much a way of conceptualising a filing system as something you might find inside one, particularly since it spans whatever distances exist between a trial’s various sites. ICH GCP E6 (R2) has put more of a focus on oversight than ever before, and it’s the TMF that has to demonstrate it. Overall, it’s the TMF as much as anything else that brings a trial together and gives its results meaning.

Even so, in Crawley’s experience, the full scope and role of the file is something of a mystery to many doctors working at investigator sites, let alone the vast population of humans who take medicine. In the common-sense routines of our daily lives, the act of levering a pill from its blister pack and swallowing it with water seems far more ‘real’ than the doubly shadowy metadata created when it’s prescribed.

However, as Crawley explains, “We are in a world that increasingly organises itself around data, and around how we share and operate data. Our reality is data, to a certain extent.” Take a look at your credit score, for instance; or your search results: data and metadata determine what we can see and do.

“It’s a very strange world we live in, but the reality of a medical product sits inside that TMF,” Crawley continues. “The MHRA, the FDA and the EMA make decisions on the basis of the data it contains. And the question of integrity, which is a very deep, personal, even spiritual, question, is also a tremendously important question with regard to data. Ensuring that integrity is the science of the TMF, and it’s the ethics of it, too. There’s no difference. In my mind, good science and good ethics go hand in hand.”

A matter of oversight

They’re not the only things. Sponsors and investigators are increasingly looking to support their TMF, and, by extension, their data’s integrity, through linking up with specialist partners and consultants. By choosing CROs and TMF quality control with experience managing comparable TMFs, it’s possible to lessen the burdens of inspection readiness and regulatory compliance. However, without proper due diligence – the first and arguably the most important step in effective oversight – outsourcing could have precisely the opposite effect, further complicating the trial and making it more difficult to assess whether it was properly carried out.

Bear that in mind, because, at first blush, the TMF manager’s task seems deceptively simple. The EMA’s Clinical Trial Regulation tries to put the issue plainly, but “the clinical trial master file shall at all times contain the essential documents relating to that clinical trial” doesn’t explain much of anything. The tautological use of ‘essential’ would seem to imply that these documents are predetermined, but the most authoritative list of them, in Section 8 of ICH GCP E6, only specifies around half of the information that needs to be filed to prove a trial’s good conduct.

That helps to explain why the most frequent major issue regulators have with TMFs are in relation to record keeping and the accessibility of essential documents. As MHRA inspector Gail Francis has explained, on more than one occasion the body has had to “add extra inspector days in order to complete the inspection due to problems with access to and provision of the TMF.” In the MHRA’s case, this is onerous even when the relevant sites are within the UK. For those outside the EU, “the ability to add additional days is very limited, and therefore problems with the provision of the TMF could result in the rejection of an application due to the inability to fully evaluate the compliance of the trial with GCP requirements”.

$120 million
The estimated cost for a particular sponsor, detailed in the MHRA ‘Inspection Metrics Report’ for 2016– 17, in approval delays due to TMFrelated issues.

“We are in a world that increasingly organises itself around data, and around how we share and operate data. Our reality is data, to a certain extent. It’s a very strange world we live in, but the reality of a medical product sits inside that TMF.”

The best way to avoid these outcomes is for all parties to agree on their respective oversight and documentation responsibilities prior to the start of the trial. Contracts should specify where compliance with the sponsor’s protocols and regulatory requirements take precedence over the contractor or vendor’s internal procedures, as well as the exact duties that are delegated and how they will be overseen.

If Francis’s phrasing wasn’t stark enough, TMF vendor TransPerfect calculates that one such issue detailed in the most recent MHRA Inspection Metrics report (2016–17) may have cost the sponsor involved more than $120 million in approval delays alone. In that case, the MHRA reported, “It was not clearly defined in agreements with the trial sponsors the scope of the TMF that was required to be held by the CRO.” After four months of delays, during which 8,000 documents were located and uploaded, the British regulator still found the two TMFs being inspected to be “significantly incomplete”, and a “robust” corrective and preventative action (CAPA) plan had to be implemented.

Failure to comply

Elsewhere on the MHRA blog, GCP inspector Mandy Budwal-Jagait details a major finding stemming from a commercial sponsor’s failure to record its oversight procedures in the TMF. Although the sponsor retained some electronic documents that could have been used to demonstrate oversight of its CROs and other vendors, they were neither included in the TMF

nor identified in the quality management system. “It was therefore not possible to reconstruct how the sponsor maintained oversight of the CRO and how the trial was being conducted,” Budwal-Jagait writes. What was worse, during a subsequent investigator site inspection, “a number of major findings were identified, which had not been detected through monitoring and of which the sponsor was not aware”.

Presenting on data-protection regulations at TMF events worldwide, Crawley has heard a number of similar horror stories from TMF managers all across the industry. “Often they’re running around trying to get all these different pieces of paper, all the different data sets and memory sticks together in one place so that it’s all interoperable and everything fits together and makes sense,” he explains. “It’s a historical problem, because that’s the way in which things have always been done. But I think we should be setting up the TMF before everything else, and doing the rest through it. It’s like if you want to launch a business today: the first thing you do is put together your website.”

“The question of integrity, which is a very deep, personal, even spiritual, question, is also a tremendously important question with regard to data. Ensuring that integrity is the science of the TMF, and it’s the ethics of it, too. There’s no difference.”

Just like that website, the TMF needs to be easy to use and navigate if it’s going to create any value. In too many cases, however, the MHRA has reported issues with eTMFs that even the sponsors’ ‘eTMF experts’ can’t use to locate all of the essential documents. Moreover, TMF design choices can have significant impacts on the way that documents are handled and monitored during a trial. The more labour-intensive aspects of operating a paper TMF can actually lessen the likelihood that files are duplicated or mislabelled. By contrast, poorly optimised eTMFs often make it easier to upload documents while restricting the user’s ability to either visually compare them or settle on a naming convention.

The devil’s in the data

As Budwal-Jagait also explains, although TMF management is commonly outsourced to a vendor, “the sponsor should be able to demonstrate their oversight of trial activities that have been delegated”. Unhelpfully, this can further complicate TMF management, as it is best done through a sponsor oversight file – a TMF component that always remains with the sponsor and reconstructs oversight proceedings separately from those of the trial.

Like the wider TMF, the oversight file needs to have the internal consistency to fully explain itself. Inspectors look to both for evidence of continual, contemporaneous monitoring and real-time inspection readiness in line with management and quality-control plans. Perhaps no aspect of TMF management has proved harder to perfect over the years. Advanced but unfriendly eTMF systems and oversight procedures often serve to discourage users from keeping files up to date, as it’s easier to manage them in more recognisable, private formats. Although real-time inspection readiness directly translates to risk reduction, inspection notifications often still prompt frantic campaigns of TMF updates.

Sometimes these bursts of activity prove completely counterproductive. “We have seen on inspection that at the end of the trial, the TMF retrieved from the vendor was amalgamated with the sponsor’s oversight file and sponsor oversight documentation removed,” continues Budwal-Jagait. “It was therefore not possible to reconstruct what oversight the sponsor had of the trial while it was ongoing.” In taking it upon itself to simplify and streamline the TMF, that sponsor actually made it more of a riddle for the regulators. It’s worth remembering that the other truth of a clinical trial is its complexity.

As a regular ethics-committee chair, Crawley, a former academic philosopher, brings his own perspective on oversight. “I don’t care if it’s a phase-I product or a phase-III product,” he says. “I want to see a pathway to patients. It’s a complex, complicated pathway, but it’s really in the TMF. It’s the TMF that’s going to tell you what the real value of your medicine is. When people discuss research integrity, they talk about honesty and conflicts of interest. And those are all legitimate questions, but they’re legitimate because they will affect the data. That’s really what counts.”

Sponsors and investigators should think of their oversight responsibilities in terms of plotting and monitoring progress down this winding pathway. They hold the map, and with it the responsibility for staying on track. “GCP is fundamentally a question of responsibility,” stresses Crawley. “Compliance doesn’t eliminate your responsibility, and nor should it really eliminate your freedom. At TMF meetings, I say your data-management plan should be run by your board – by the board of GSK or the board of Pfizer – because data is the whole value of the company.”

But he’s not necessarily complaining. “What I like about the TMF people is that they’re in the right place in their organisations, because they’re the ones who are responsible for the data.”