A new working paper has found that one-third of all data points collected in 105 phase 2 and 3 trials were not needed for the studies’ key analyses, while also highlighting how an increase in clinical data collection is translating into an increased burden on patients and trial sites.
The study was conducted by the Tufts Center for the Study of Drug Development (Tufts CSDD) in collaboration with 14 biopharmaceutical members of nonprofit TransCelerate BioPharma, including Merck & Co., Novo Nordisk and Gilead Sciences.
As a preprint, the study has not yet been peer reviewed. TransCelerate announced the results in a Sept. 15 release.
The findings give “a lot of ammunition to clinical teams to really discuss and look at how they can optimize the collection of that data so that they can minimize patient burden, site burden and all of those downstream effects,” Kenneth Getz, executive director of Tufts CSDD and lead author of the study, told Fierce Biotech.
Trial protocols have grown increasingly complex over the last 25 years, Getz explained, becoming more customized, adding more eligibility criteria and endpoints and enrolling more patients at more sites. This, in turn, has caused poorer trial performance, making it more time-consuming to recruit, screen and retain patients and to clean and analyze mountains of data.
Previous work from Tufts CSDD found that the number of datapoints collected in phase 3 trials jumped from 929,203 to 3,560,201 between 2012 and 2020.
“We have studies where patients have to go to three or four different locations within an institution just to have different assessments performed the same day,” Getz said. “All of that adds time, adds burden.”
In the new study, Getz wanted to examine two different types of potentially unnecessary data. Data that are not required by regulators or used to support key primary and secondary endpoints are deemed “non-core.” If investigators have already collected enough of an important data type, but keep gathering more of that data anyway, that excess data is called “non-essential.”
Using algorithms to assess patient burden, which includes factors like travel time and lifestyle changes required by the study, the team determined that between 25% and 30% of the burden placed on trial participants was due to non-core and non-essential data collection.
While investigators may feel drawn to gather more and more exploratory data in order to inform future clinical development, Getz said, trial sponsors need to recognize that doing so has tradeoffs.
“That kind of curiosity is at the heart of the way we typically innovate,” he added, “But it's creating a balance between feeding our curiosity and also recognizing that it does increase burden.”
TransCelerate member companies are now working on a framework to optimize their data collection based on the results, Getz said, which will be published and shared with the broader industry. And Tufts’ findings also align with regulatory guidelines, he added, specifically the January 2025 guidelines for good clinical practice (PDF) from the International Council for Harmonization of Technical Requirements for Pharmaceuticals for Human Use (ICH).
“It calls for fit-for-purpose, intentional protocol designs that weigh patient and site burden, that really look to optimize data volume,” Getz said of the ICH guidelines. “We've even got regulatory encouragement for the kind of behaviors that we're advocating for in the paper.”