January 31, 2019 | Andy Lawton has a career’s worth of experience doing statistics in healthcare. First he worked the UK’s National Health Service as a statistician, helping labs with quality control, consulting with physicians, and assessing clinical trials from six patients up to 20,000. When he joined Boehringer Ingelheim (BI), he began in statistics, but during his time at BI was involved in developing clinical data management systems, safety systems, and administrative systems, as well as special projects such as Data Transparency, BI representative to TransCelerate, and—externally—the launch of CDISC.

In 2016, Lawton left Boehringer to form his own company, Risk Based Approach Ltd. “I could see the risk-based approach was coming in strongly,” Lawton told Clinical Research News. “I'd been working in the risk area since 2002, taking some of our first studies on metrics and risk-based approach.” Today, he’s worked with 26 companies on five continents, and he is applying his expertise to risk-based approaches in trials.

Lawton believes that companies need to shift their perspectives so they are measuring real quality parameters, data quality issues, and safety issues. “That's what we should be measuring, and not just emphasizing critical success factors,” he says. On behalf of Clinical Research News, Lee Yuan spoke with Lawton about what these real quality parameters are, and how companies can address them.

Editor’s note: Lee Yuan, Conference Director at Cambridge Healthtech Institute, is planning a track dedicated to Risk-Based Monitoring at the upcoming Summit for Clinical Ops Executives, SCOPE, in Orlando, February 18-21. Lawton is chairing and will be speaking on the program. Their conversation has been edited for length and clarity.

Clinical Research News: What do you feel are the biggest challenges in implementing risk-based monitoring and risk-based approaches today? Why?

Andy Lawton: Oh that's a long response, I'm afraid. Risk. It’s both the exact term that companies should be looking at, but it's one they shy away from. It’s sometimes difficult to understand because companies take huge risks in developing their drugs. And they seek to minimize in every other area the risk they’re taking. This is probably creating an environment where we don't manage risks as well as we should because we think we're covering them up by doing 100% monitoring. Which is where this all started off.

The risk-based approach was first put out by the regulators in ICH Q9. Back in 2006 and was moved over to the clinical trial area in 2011 with the risk-based monitoring publications from the FDA and the EMA, which were finalized in 2013. These were followed with the ICH E6 revision in 2016.

The aim was originally to take more risk-based approaches to clinical trials, but now we need to take a risk-based approach to everything we do. This calls for pharmaceutical management really to assess risk in all areas—not just clinical trials—but system risks within their companies. And this, again, is a challenge for companies. Many feel that they’ve started to implement it or they are implementing it, but they’re still not covering the full scope of what’s envisaged by the regulators.

I think the challenges for medicine and clinical trials is really number one, to understand what risk is. In doing those risk assessments and following through from that. And then how to best manage it and to get the whole organization to buy in. Part of the key to this is probably quality tolerance limits, which I've spoken on many times, but it is still so misunderstood. As W. Edwards Deming said, "Don't waste efforts on zero tolerance goals”. That is do not seek perfection, identify an acceptable error rate, and try and improve on that. And it's really because companies are dealing with humans that they say, ‘No, we've got to be perfect.’ And we never are perfect. There are always errors; there are always issues. If we're seeking perfection, we never achieve it, so we always fail.

I think the challenge is still getting companies to understand risk, how to measure it, what to do about it, and to move away from maybe some of the parameters they have been measuring, such as recruitment rate, onto really critical areas such as number of adverse events (Are they what you expect?) and number of protocol violations (Are they what you expect?), real quality parameters and data quality issues as well as safety issues. That's what we should be measuring, and not critical success factors.

How can industry be encouraged to adopt defined quality?

Well, in part, I think that's going to have to be down to the inspectors and regulators really to encourage companies. I know it's there in ICH, and that should be encouragement enough. But perhaps inspectors are going to have to say if they're not implementing quality tolerance limits, give them findings on it. Make it quite clear that defined quality is expected.

What inspectors will probably do is that if they're not implementing quality tolerance limits, and actually pre-defining quality, they'll probably find them on minor issues. And just keep giving findings on that because the companies haven't pre-defined quality. But it might be difficult for some of those companies to make that link between, “Oh, well, they've given a minor data finding and we're doing risk-based monitoring, why is that?” Rather than saying, “Well, you haven't implemented quality tolerance limits and pre-defined your quality, which could have helped prevent this finding.”

What innovations in the trial space have you most excited, especially around risk-based monitoring?

Again, an enormously big question and we'll be talking on that in the coming presentations. Quality tolerance limits, although it's not a great technology innovation, is a big process change. That really excites me: the whole area of the change in how we manage the risk-based approach. Because we're going see a big change over the coming five, 10 years. As people understand risk better, they'll move on and change their processes. And we're already seeing some companies, some of the early adopters with risk-based monitoring, actually change the way they do things: making their algorithms more sophisticated, building more intelligence into their algorithms.

Now instead of algorithms which were only 30% successful at detecting issues, we've got ones that are 80% predictive. And once you're doing that, you're achieving the ICH goal of being more effective and efficient so we can get more drugs to patients, develop more, and do it cheaper.

The integration of all of the different data systems that we have is exciting; having those all together is going to be really critical. On top of that I would say anything, any processes where we can make it electronic: informed consent, patient diaries, any of these where we can make the data more immediate and accessible to all are going to be critical areas that we really need to focus on.