The RECOVERY Trial’s IT team, based at Oxford Population Health, are responsible for the study’s computer software programs and databases. These perform a wide range of functions, including randomly allocating treatments to participants; recording participant outcomes; keeping records of training; and identifying any missing information about participants. Normally, Andy, Dave and Richard work on clinical trials related to perinatal and maternal health, whilst Bob supports other clinical trials led by Oxford Population Health, including studies on treatments for diabetes, cardiac diseases and kidney failure. In combination, they have more than 50 years’ experience of developing IT systems to support clinical trials.
What are the IT systems in the RECOVERY Trial?
Andy: First we have the randomisation database, which I manage. It is very important that there is no bias in how participants are allocated the different study treatments, as failure to prevent selection bias could invalidate the results from the trial. The clinicians fill out a simple online form with the patient’s basic details and any conditions that would make certain treatments unsuitable. The computer then randomly allocates them to a study treatment, or usual care (the control treatment).
Dave: The OpenClinica database (my role) handles clinical information entered at hospital sites about each participant, including which treatment(s) they received, the treatment doses, whether they required mechanical ventilation, etc. Other clinical information is sourced from the main NHS databases where routine data is recorded. This helps keep RECOVERY-specific data collection to a minimum: doctors and nurses generally only need to fill out the randomisation form and a follow-up form after 28 days to confirm the participant’s vital status (whether they are alive or dead, still in hospital or discharged).
Richard: Finally, we have the Trial Administration Database Application (TADA), which is my responsibility. This supports the administration of each hospital site, including the monitoring of data collection, the clinician training records, and the management of the study treatments. A popular feature is that it sends out an automatic email summary each week to every hospital site that lists any outstanding paperwork that needs to be chased up.
Bob: All of the systems need to be tested to confirm that they are working properly to ensure that the data we collect are complete and accurate. Testing is always really pressured because it is the final stage before the system goes live. Sometimes there’s only a very limited amount of time, but it’s important to get it right.
One concern was that our systems weren’t originally designed to handle the sheer volume of data that RECOVERY was generating: at one point, the RECOVERY Trial recruited a thousand people in a single 30-hour period.
How did you all become involved in the RECOVERY Trial?
Andy: I can trace it to the exact conversation. It was Thursday 5 March 2020, and Dr Alan Young, Director of Information Science at Oxford Population Health and a friend of mine, asked if I could meet with Professor Martin Landray that afternoon. At that meeting, Martin explained his idea to set up a clinical trial across all UK hospitals to test potential COVID-19 treatments for patients severely ill with the disease. It was very clear we needed to make a decision as soon as possible. As Martin put it, we were three weeks behind Italy and if the study wasn’t in place before the first COVID-19 wave hit the UK, it wouldn’t be possible after that. I said that if we dropped everything else there was a chance we could get all the IT systems in place within two weeks. But it wasn’t just my decision of course. Since we all manage separate parts of the system, Richard and Dave had to be on board as well, and we needed Bob available to validate the systems.
Dave: When Andy told us Martin’s idea, we could see it was an exciting opportunity. Crucially, we also had support from our managers to concentrate on it fully. We couldn’t have done it if we had been working on anything else.
What happened next?
Bob: It was an incredibly intense period to set everything up, and in the end we did it just nine days after becoming involved. Normally, we set up the IT systems for a trial over several months, with meetings every few weeks. But now we had daily meetings, and whereas before we might have said ‘We can discuss that at the next meeting’, we were firing off emails to get a decision within the hour. Each one of us was working late, often into the early hours of the morning – but it was definitely worth it.
Did you feel under a lot of pressure for everything to work?
Andy: Absolutely. We were always aware that if any of the systems went down, the whole study would be affected. For instance, if the randomisation process stopped working, no new participants could be entered into the trial. Another concern was that these systems weren’t originally designed to handle the sheer volume of data that RECOVERY was generating. Often, the trials we work on recruit around two or three people a week; at one point, the RECOVERY Trial recruited a thousand people in a single 30-hour period. Thankfully, everything has been remarkably robust so far. Also, we were incredibly fortunate that none of us got COVID-19 during those crucial early phases or we could not have launched so rapidly.
Apart from its scale, how else has RECOVERY been different to the clinical trials you usually work on?
Dave: I’ve never worked on a trial that has undergone so many changes. Usually, when a trial launches, the study protocol remains the same throughout, except possibly for one or two very minor changes. But with RECOVERY, the protocol was constantly being revised in really major ways as treatments entered and left the study, and we went from a one-step to a multiple-stage randomisation. So we’ve had to be more hands-on throughout. There were also a lot of extra tasks needed when we launched the international arm of the study, including having to translate the whole system into Vietnamese!
Richard: I would also say that we are not usually so involved with the wider study team for a clinical trial. For RECOVERY, everyone gets together for the regular team meetings (almost all of which have been online), which really helped since we were working with many people for the first time. The camaraderie and level of communication has been brilliant. I think this is a key reason the trial has been so successful, and it’s meant our contribution feels really valued.
Has RECOVERY changed the way you work on clinical trials?
Dave: I’m particularly proud of the processes we set up for clinicians to complete the essential training for the study (for instance, how to obtain a patient’s consent to take part, how to complete the randomisation form, etc). Normally, this would be delivered through in-person workshops or lengthy online courses, neither of which were possible during the pandemic. A series of simple videos on the RECOVERY Trial website have been developed that clinicians can watch; then they tick a box to self-certify that they have seen it.
Richard: This is captured in the TADA database, so we have a complete record of the training everyone has done. We are now starting to introduce similar self-certified training systems for many of the other clinical trials led by Oxford Population Health.
Using a self-certified training system is much more cost- and time-effective, and means the clinicians can watch the videos as many times as they need to, or have a refresher whenever they want.
Bob: We’ve shown that it’s possible to do things differently – and not just in a pandemic situation. We should certainly take the opportunity to learn from RECOVERY for future trials.
What has been the real highlight for you from working on RECOVERY?
Richard: It’s really great to think that, even in a small way, we’ve been working as part of the global effort against COVID-19. Even my friends outside work recognise how important RECOVERY is, whereas they haven’t heard of any of the other studies I’ve worked on!