Building Amgen’s RWD Platform

July 1, 2016

By Robert Israel

July 1, 2016 | Necessity spawns innovation, and innovation succeeds by addressing a vexing challenge with insight, practical approaches, and a way forward.

Amgen’s information systems team, based in Thousand Oaks, California, and headed by director Minnie Chou, did just that with the Real World Data (RWD) Platform. The collective efforts of their team—which included cross-functional input from Amgen’s Center for Observational Research with assistance from an outside marketing consultancy—were recognized for their derring-do and breakthroughs when it received a Bio-IT World’s Best Practices Award earlier this year.

WebLogoAmgen’s RWD team hoped to provide an enterprise-wide capability to better access and analyze available claims, electronic health records (EHRs), and cohort data, and address questions across the clinical and health IT development lifecycle. It was a tall order. They needed to deliver insights for cross-functional teams who could use the data in a timely and cost-effective manner.

The driving force, in Amgen’s case, was the demand for this information.

Cathy Critchlow, Amgen’s executive director at the Center for Observational Research, summed up the necessity for a solution this way: “[There is a] growing demand for real-world data from both internal and external stakeholders,” she said, “[so] we can analyze these big data sets more efficiently to better inform decisions made by regulatory agencies and payers.”

Chou and her team initially found fault in Amgen’s IT infrastructure.

“We found that the underlying IT infrastructure was not adequate for processing the increasing volumes of real world data maintained by Amgen,” Minnie Chou told Bio-IT World. “The cache of real world data was not standardized into a common data model. It was not centrally located. It was not being leveraged in a consistent manner across the enterprise.”

The team further discovered that this lack of an underlying IT infrastructure not only prevented them from processing the growing volumes of real world data, but that they were hampered by excessively long cycle times.

The Amgen team set to work concocting a multi-layered recipe, similar to the multiple layers of a cake.

“By integrating a high-performance technology, converting the information into a common data model, laying on top cohort builders and pre-defined patient populations, and [adding on] a suite of advance analytics, we found we could fully leverage the wealth of real world data on a platform that delivers insight in a timely and cost-effective manner,” said Chou.

The “high-performance technology,” in this case, is Hadoop, the open-source software that breaks data into blocks. These blocks are stored on clusters of commodity hardware. Known for its flexibility, efficiency, and scalability, Hadoop quickly processes data. Amgen calls this layer of their cake Enterprise Data Lake (EDL). It reduces average access and analysis times from several days and hours into minutes. “Sometimes it only takes seconds,” Chou declared.

The next layer of Amgen’s RWD Platform cake is the Data Harmonization Layer.

“We converted claims and EMR data on 150 million patient lives across the United States, the European Union, and Canada to the Observational Medical Outcomes Partnership (OMOP) common data model,” Chou said. “This harmonization provides significant programming and analysis efficiency. By establishing cohorts of patients with target diseases using a common design framework, we can now conduct real world patient data analytics more consistently, and can respond more quickly to time-sensitive questions on Amgen’s medicines.”

Next came a layer Chou calls the Patient Population Identification and Analytics Layer. This further breaks the data down to support a wide variety of business needs. Among these are clinical study design, disease burden, forecasting, clinical effectiveness, and post-marketing safety analyses.

Finally, the team developed a web portal housing insights on target diseases and on Amgen’s marketing medicines across 18 countries, providing a single point of access to search and data visualization capabilities.

The Amgen team conducted a series of pilot tests to see if their innovation worked.

“The value became immediately clear across many domains with Amgen,” Chou said. “And these demonstrated clear efficiency gains.”

The team’s RWD platform is now being used throughout Amgen to enhance its ability to leverage the ever-increasing volume that initially seemed destined to be stuck in an ongoing logjam.

“Had we not been able to respond to such regulatory inquiries in a timely manner,” Chou said, “there is a risk that action might be taken – a label change, for example – on the basis of spurious associations that would not have been supported by the evidence if such evidence had been available. In these cases, the implications for patients and for the company are large.”

Since being awarded the Bio-IT Best Practices Award, the team has continued to make advancements to the platform. According to Chou, they’ve integrated additional databases, and this has led to more advanced analytic and visualization solutions, thanks to the RWD Platform.

“The goal of bringing new medicines to patients faster and providing scientifically rigorous information about the effectiveness and safety of Amgen’s medicines in the real-world has now been enhanced,” Chou said.