Sponsors have combined data from internal and external sources for years, but the pandemic put their data management strategies on steroids, creating a need to gather, process and interpret data in real time to make informed decisions to keep clinical trials up and running.

The pandemic drove sponsors to develop new tools to track and predict trial disruptions and restarts in real time and respond to their sites’ immediate needs.

“We had to reinvent our mechanisms, our infrastructure and our pipelines,” Chandi Kodthiwada, head of product management for R&D IT at Takeda Pharmaceutical, told attendees of the 2021 SCOPE Summit last week. “Everything about how we collected data, how we processed data and how we used data to inform decisions was on the line — to be redesigned to fit this new approach.”

William Illis, global head for collaboration and technology strategy in clinical development at Novartis, said the dramatic shift to reliance on real-time data led to a cultural change for sponsors. “COVID-19 really made us flip the script on that,” he says, adding that sponsors “shifted in a big way from policies informing practices to practices informing policy. In the past, there would be a lot of discussion about what sponsors can do for SOPs or compliance. It’s not that [data science] was unimportant before the pandemic — it just changed the nature of those conversations.”

Last May, for example, GlaxoSmithKline (GSK) began developing a “clinical control tower,” a dashboard that forecasts potential disruption to its trial operations based on data from recruitment, both planned and actual, as well as details on COVID-19’s prevalence in particular countries and its projected infection rate.

“The real-time dashboard highlighted opportunities for us to amend our approach and keep trials moving to plan,” says Eddie Dukes, tech product owner at GSK. He says that for trials that GSK did not outsource to CROs, GSK was able to refresh enrollment data every 24 hours. For outsourced trials, the delay was closer to one week, he says.

GSK’s creation also relied on country- and trial-level surveys, which included questions on specific IRB guidelines and whether CRAs could make on-site visits. The surveys also included questions on whether home nursing could be implemented without a protocol amendment, whether patients could get access to sites and whether electronic signatures were a valid form of consent. He says Johns Hopkins University was the only external data source plugged directly into the dashboard.

Novartis’ Illis adds that enormous amounts of data collected in real time from innovations such as telehealth and virtual patient visits produced “a lot of operational data, which are a source for data science and metrics in gaining a better understanding of what processes are working or where they might be breaking down.”

At AbbVie, an overlay “command platform” was created in early March 2020 that incorporated active site information with COVID-19 public health data; it also relied on site- and country-level information on restrictions for each site.

AbbVie analyst Joseph Yoder says Johns Hopkins was AbbVie’s source for external information about the pandemic on a daily basis. “As the months went on, it was more about how the shutdowns are impacting the sites, which ones are shut down and which ones aren’t. And from there it evolved to the point where we’re now tracking those sites [in real time] as they’re coming back on board.”

He adds that AbbVie is currently focused on what sites need in order to get back to more typical trial activities — how to get investigational product to sites and how to help sites get their enrollment and screening numbers back up to speed.

Yoder said analysts at AbbVie and elsewhere initially thought the pandemic would only last two or three months, not a year or longer. After researchers said they needed to get a handle on the impact the pandemic posed to their sites, the development team came up with its first version of a tool to check on sites within 24 hours.

AbbVie’s dashboard features map images of the pandemic’s impact on sites by country, filtered by active sites, sites in startup, supply management sites and other breakdowns. The maps are driven by data from the Johns Hopkins daily COVID updates overlayed with data provided by AbbVie sites on therapeutic area, enrollment totals, screening data, principal investigator’s name, study phase and priority of the trial. Users can see at a glance whether the resulting data will have an impact on various trial factors, such as meeting enrollment, participant dropout rates and whether the drug must be administered on-site.

The dashboard also presents data on trends and COVID-19 growth predictions by country. Using this information, AbbVie can evaluate individual trials’ and sites’ potential for restart.

“I imagine that vendors that produce these types of tools will be more agile and present better solutions [to sponsors] to be able to handle this sort of scenario. I’m not saying necessarily that the tools that exist currently broke down or fell short completely, they just didn’t have that agility to incorporate such a shock to the system.”

Takeda’s Kodthiwada says pharma companies’ appetite for more data is leading to a shift in hiring resources and upscaling existing resources for more data engineering, algorithms and visualization. “Before this, data science, the skill of wrangling data or interpreting data used to be done in one part of an organization. We now see this trend where almost every team within an organization is capable of owning and interpreting data and having a certain set of the deep data science skills.”

Illis adds that enormous amounts of data now being collected in real time from innovations such as telehealth and virtual patient visits “are producing a lot of operational data, which are a source for data science and metrics in gaining a better understanding of what processes are working or where they might be breaking down.”