Learning to handle disparate and complex data sources in clinical trials

In 2019, in an effort to discover more about growing trends in clinical trials, eClinical Solutions partnered with an independent research team at the Tufts University Center for the Study of Drug Development. Their aim was to quantify the increasing use of external data sources and systems among life science companies pursuing clinical trials, and the impact it was having on cycle times, analytics and data sciences capabilities, and AI preparedness.

Both organisations jointly surveyed 149 different sponsor companies across the globe over three months, with the goal of gaining a deeper understanding of the issues currently faced by clinical development teams. The resulting study – the Tufts-eClinical Solutions Data Strategies & Transformation study – revealed much about how life scientists around the world were approaching clinical data strategy and management.

Quantifying trends in clinical trials

“The goal was to quantify some of the industry shifts we’d observed working with our clients over the past few years, many of which work on the cutting-edge of life sciences, developing immuno-oncology treatments and gene therapies for rare disease” says Sheila Rocchio, CMO at eClinical Solutions.

“We’d noticed a big increase in the number and types of data sources these clients were using to run trials – from biomarkers and specialty labs, to genomic data and physiological sensors – and they were all facing significant and similar challenges. All were trying to handle many different data providers and make sense of this massive influx of data. The data in research has changed dramatically, but the tools to handle these data – Excel spreadsheets and SAS programs – have not changed for 20 years.

“Our clients were looking for a technology platform and provider that could integrate their many different data sources and types into one source of clinical data ‘truth’. From this one source, stakeholders could work to review data, share insights and make decisions, while demonstrating comprehensive provider oversight and data quality.”

Cycle times on the up

Several key findings came out of the study that clearly outlined the main pain points for organisations struggling to manage their data sources, while also revealing researchers’ best ways of addressing data chaos. The first was that cycle times for database locks were found to be increasing.

As companies were having to handle more data from more sources, locking databases in clinical trials was found to be taking far longer. “There was a 40% increase from 2017 to 2019 for the ‘Last Patient Last Visit to Database Lock’ cycle time metric for those companies using more than four data sources,” says Rocchio.

“This significant increase in cycle time speaks to the challenges that life science companies face when they add more data sources and providers to their trials. The goal of these data streams is to have a richer set of evidence about patient experience, with a new therapy for analyses required to develop precision medicines. New speciality labs can deliver biomarkers from samples, but they may not have experience providing resulting data transfers in standard formats and in a regulatory compliant manner. The process of assembling clinical data pipelines for review, integration and analysis is still very manual, and was considered ‘difficult’, ‘time consuming’ and ‘labour intensive’ by 93% of study respondents.”

Implementing data strategies

Another important finding was the impact that implementing formal data strategies had on these database cycle lock times. For sponsors using more than four data sources, cycle lock times were shown to fall by ten days on average once a data strategy had been effectively executed. Sponsors were able to overcome some of the obstacles in coping with an influx of new data by implementing a data strategy, and were faring better than their peers without a data strategy across the board. Respondents with a data strategy defining the flow of clinical data throughout their organisation also rated their data operations activities such as ingesting, organising and mapping data as less difficult than those without a defined strategy.

Rocchio adds: “Those companies using newer technologies like a clinical data platform or hub to centralise and integrate disparate clinical trials data earlier on in the process, and build automated data pipelines, experienced fewer delays. Those with data strategies and the latest technology platforms also considered themselves to have more mature analytics, and were ready to take advantage of recent advances in data sciences, machine learning and AI.

Looking ahead

Findings from the Tufts-eClinical Solutions Data Strategies & Transformation study show that smarter data strategies hold the key to more effective clinical trials – an idea brought into sharper focus in 2020 because of Covid-19.

The pandemic has resulted in 61% of clinical trials being put on hold, as life scientists around the world pivot to tackle a more immediate, unknown threat, and bring vaccines to market in record speed without compromising safety.

Covid-19 has also accelerated the need to leverage technology to conduct virtual clinical trials, bringing research to patients to reduce the burden of trials, and to increase participation rates. In this environment, comprehensive digital data strategies that rely on centralised clinical data platforms, like elluminate from eClinical Solutions, become critical for bringing about a new and improved age for clinical research.