At Janssen, Data Analytics Is Pivotal To Trial Success

By Deborah Borfitz

April 7, 2021 | Janssen is using two homegrown data analytics solutions at scale to optimize clinical trials, enabling internal teams to solve real use cases embedded in the development process, according to Miruna Sasu, MBA, Ph.D., executive director for global development feasibility and advanced analytics at Janssen R&D, during a session at the recent Summit for Clinical Ops (SCOPE). The company is infusing real-world data (RWD) and data science into long-established processes “to increase the operational success of trials and minimize patient burden,” she says.

Trial planning begins by understanding the patient population and determining how to work with targeted study sites and countries, says Sasu. The perspectives of patients and investigators, as well as the competitive landscape, are key considerations. 

The work done by its “feasibility center of excellence” includes program and protocol operational feasibility, protocol optimization, country and site selection, enrollment prediction, and “next best actions recommendations,” says Sasu. Enhancing processes, gathering and integrating intelligence, and a layered and well-timed analytics approach are among the means. Platforms and technology get built to facilitate data-driven decision-making by study teams.

Two solutions—a customized point-and-click “FIDO” platform for analyzing data for trials and a patient-finding tool build in-house using RWD—have been successfully deployed over the last two years, Sasu shares. 

The FIDO platform involves a partnership with a vendor, data scientists, and the information technology team, and is proprietary to Janssen, she says. It is a “one-stop shop” that surfaces complex algorithms for users without any data science expertise. Data standards, an application programming interface, and natural language processing are at work behind the scenes. 

FIDO “follows the feasibility process for every trial,” she continues. “It provides advanced analytics to non-super-users [and] accelerates time to outputs” with easy visualizations. Results of the complex analytics are shared across the enterprise.

A team was tasked with identifying use cases and finding the right vendor partner for the custom build, which integrated a series of operational apps in one place, says Sasu. These include a dashboard providing metrics across different groups as well as tools for program and protocol feasibility, protocol optimization, site selection, country selection, and enrollment modeling.

The platform also has an analytics workbench and clinical data store, but these features are designed more for super-users, she adds. This provides easy access to internal and external data sources for ad hoc and algorithm testing. 

Choosing Sites 

The patient-finding algorithm is an app for site selection, says Sasu. “It allows the [feasibility] team to analyze RWD for trial optimization at the click of a button.”

Team members can learn more about the patient population prior to study launch, for example, to understand the current standard of care and how to position a trial as a treatment option, she says. The data science team does a lot of the sophisticated analytics on the back end, using a trial’s inclusion and exclusion criteria, to simplify the exercise for users. 

Janssen has also licensed a lot of real-world datasets and looks at all of it in a standardized way “without reprogramming and recoding every time,” says Sasu. 

Critical to the initiative’s success was assigning responsibility to an internal team with a project manager and goals, Sasu says. The app built by the feasibility analytics team is now being used at scale.

One way the app gets used is to look at a population of patients in relation to the geographic location of study sites as well as information on their characteristics and medical treatment. Opportunities for improvement can be identified in a matter of days, she says.

Load more comments