ERT administers a hefty dose of AI to its application development pipeline

Profile picture for user jtwentyman By Jessica Twentyman September 18, 2020
Technology from Dynatrace is enabling ERT’s developers to build software faster, so that its customers in the pharmaceutical industry can streamline the delivery of safe, groundbreaking new therapies to hospitals and pharmacies. 


For years, executives in the pharmaceutical industry have talked excitedly about the prospect of virtual clinical trials. The idea is pretty simple: instead of confining people on whom a new drug is being tested to a hospital ward or requiring them to make regular clinic visits, why not use modern technologies to manage and monitor them in their own homes? Internet-based conferencing and messaging technologies are perfectly capable of supporting physician/patient consultations, after all, while wearable medical devices can be used to collect patient vital signs.

Despite much talk, change has been relatively slow to come to this risk-averse and highly regulated industry. But the COVID-19 pandemic looks set to accelerate uptake of virtual clinical trials at drugmakers keen to get new products out of laboratories and into hospitals and pharmacies. After all, at a time of lockdowns and social distancing, a virtual approach may be their best chance to complete ongoing studies and start new ones, and avoid losing out on highly lucrative market opportunities to more agile competitors. 

That’s good news for Philadelphia-based ERT (eResearchTechnology), a provider of software, hardware and services for running clinical trials - both those run along more traditional lines and virtual ones, too. In a recent survey of around 110 clinical trial professionals conducted by the company, more than four out of five (82%) stated that their organizations are now incorporating elements of virtual clinical trials as a result of the pandemic, up from 33% prior to COVID-19. 

Under pressure

At the same time, this burgeoning opportunity puts ERT under some pressure when it comes to building new software quickly. Each and every clinical trial comes with its own risks and challenges, so the applications that ERT builds for pharmaceutical companies and researchers typically come with a high degree of customization, so they can be tailored for every specific use case. 

That takes a substantial application development effort, which in turn makes the observability of ERT’s application delivery pipeline a top priority, explains Simon Pilar, director of IT service operations at ERT. For this, it has turned to application performance monitoring specialist, Dynatrace. Says Pilar: 

For us, it’s not just about delivering software quickly, in order to accelerate our customers’ clinical trials. It’s also about being able to hit the highest levels of quality, reliability and performance with that software, while continually increasing the speed of innovation. What we wanted was the kind of observability that could give our developers the insights and answers they need to optimize their code and create the best end-user experiences right from the start. Ideally, we can identify and stop bad code from ever reaching production, so it’s never a problem for customers.

Davis and Keptn

The Dynatrace AI engine Davis, he explains, continuously monitors ERT’s DevOps processes and app delivery pipelines, watching for any errors, degradations and regressions. It automatically prioritizes issues to be dealt with based on their impact and ensuring the DevOps teams at ERT can find, understand and resolve issues before they impact clinical trials. 

Meanwhile, ERT is also using Keptn, Dynatrace’s open source control plane for cloud-native continuous delivery and operations. This means that any issues uncovered by Davis can trigger automatic, self-healing actions and its ‘quality gate’ feature prevents bad code from reaching production. 

The end result is a reduction in development times, while ensuring high standards of quality and precision in the software used to manage clinical trials, says Pilar:

So before we had Keptn in place, we had pipelines with a lot of manual steps. So, as an example, we’d have our builds, then somebody would deploy the code and tell the test system, ‘OK, you’re good to test now’. We would execute manual and automated tests, get feedback and so on, and then it would be deployed to the next stage. It was a lot of manual effort. Now we can automate the handover between different environments without all that manual interaction - so that’s faster, and beyond that, developers are releasing code quicker.

A lot quicker, as it turns out: on average, ERT is now able to speed up delivery of new software from an average of six weeks to just four weeks. But more importantly, it can do so with total confidence that this software will perform well for the customer and help them get new, potentially life-saving therapies out the door faster, with no risk to patient safety.