Home » Regions » Global News » Three Questions: Rob DiCicco, GlaxoSmithKline

Three Questions: Rob DiCicco, GlaxoSmithKline

Monday, August 21, 2017

CWWeekly presents this biweekly feature as a spotlight on issues that executives in clinical research face. This week, writer Suz Redfearn spoke with Rob DiCicco, vice president of Clinical Innovation and Digital Platforms at GlaxoSmithKline.

Q: Research indicates that 66% of protocols are amended, and one out of every 10 amendments have issues related to human error. From your perspective, why is this happening?

A: I think it’s part and parcel of working in a document-based environment. The medium lends itself to that. Additionally there is a huge focus across the industry on study startup. What winds up happening is that folks are prone to copy and paste, and errors are likely to be carried over. A number of activities are dependent on the protocol beyond the study design itself—electronic data forms need to be created from it and websites need to be updated with descriptions of the trial from the protocol. Because of the amount of manual effort that gets put into not only creating the protocol but engineering all of the downstream connected processes, there is a lot of opportunity for human error.

The crux of the problem is the document-based nature of how we handle protocols. This, coupled with this cultural impetus to start a study quickly, are probably your leading-edge factors that are driving human error.

Q: What has been done to try to remedy this and why did it fail?    

A: There have been two approaches. Some folks have tried to adapt the application of certain kinds of software to drive protocol authoring and take some of that manual effort out of it. These efforts probably have failed because the level of disruption has not been commensurate with the level of gain. Those software applications don’t really consider the workflow of how a protocol is created or how sites need to utilize a protocol to determine workflow at their institution.

Sponsors also try to review quality backward into their protocol by setting up peer review communities, leveraging third parties to do quality control, etc. But I think there’s a limit to that versus taking a quality-by-design approach. These have been the primary efforts.

Q: What, in your estimation, can fix this?

A: I think the first step has been taken with the evolution of both the TransCelerate Common Template Protocol project, which has representation from 15 of the member companies. We have also been collaborating with the NIH and FDA to harmonize the templates people use to create their protocols. That lays the foundation for a couple of things. It allows a certain consistency, along with suggested language that people can adapt right out of the box. It also lays the foundation for a good amount of automation.

One of the obstacles that has been in the way of those authoring the protocol, as well those re-using information in the protocol, has been too much customization. Once certain key elements of the content are harmonized, now you lay the foundation for leveraging automation to building the document. TransCelerate released its first version in December of 2015 and updated it twice in 2016, and in May of 2017. We jointly harmonized our template with those of the NIH and FDA.

So everyone agreed on what the container looks like and what key elements of content look like. Next, how can we leverage existing technology to not just standardize authoring of study design, but other downstream processes like data collection, study procedures, workflows, budgeting and contracting? All of these things had been manual entry and document-based.
Typical phase III studies may run for several years, which means that when the next protocol needs to be written in that disease area, there’s a good chance the author and team didn’t work on the first study. They will start a whole new protocol, likely copying-and-pasting from that first study, introducing human error versus using an automated approach where content can be added in a controlled way, layered on top of what’s already there, and the info will be consistent from protocol to protocol.

We’ve hit that tipping point now where we can start to export protocols into applications where you take the manual effort out of it, thus taking the risk out of the process.

 

This article was reprinted from Volume 21, Issue 33, of CWWeekly, a leading clinical research industry newsletter providing expanded analysis on breaking news, study leads, trial results and more. Subscribe »

Related Posts