My closer look at…. “What the Rise of Real-World Evidence Means for the Pharmaceutical Industry: A Closer Look”

Written by: | Posted on: | Category:

ISPOR’s Value and Outcomes Spotlight recent Sep/Oct 2020 issue contained an interesting Q&A article with Jennifer Graff (VP, Comparative-Effectiveness Research, National Pharmaceutical Council). Overall, it was interesting, high-level stuff - which raised a few Qs in my head. Some extracts from the article:

"Researchers have estimated that the use of real-world evidence could reduce trial costs between 5% to 50% to expedite safety monitoring and simplify trial and data collection."

Does anyone know where this ‘result’ originated from? Like any good statistic it sounds reasonable enough (‘0%’ and ‘100%’ aren’t really options), has an element of alliteration, and is vague enough to avoid serious challenge. Following a quick internet search, the earliest document I could find making this claim was a QuintilesIMS white paper from 2016. Regardless of the source, I wonder if the ‘could reduce costs’ has resulted in ‘has reduced costs’…?

"There are multiple technical challenges with the collection, transformation, and evaluation of real-world evidence. However, we are learning that good data with thoughtful design and analysis yield similar results regardless of the sophisticated statistical manipulations"

The sophisticated statistical manipulation IS the thoughtful analysis!

Data from RCTs is comparatively structured, clean, and robust in relation to RWD. A well designed and conducted RCT would (in most cases) be happily analysed using simple statistical approaches (e.g. t-tests, chi-square). Indeed, this was the general approach until the wide-spread use of computers. In such studies, you tend to get similar results regardless of whether the analysis is simple or complex. I would argue that the same is not true for analysis of RWD. Given the multitude of potential biases this data has, you invariably need to set aside the analytical pop-guns and bring out the big statistical cannons. Good quality RWD does reduce the need for complex statistical approaches, but the lack of randomisation and blinding invariably leads to analytical complexities.

"….the demand for highly trained individuals to design and analyze high-quality real-world evidence studies exceeds the supply. These challenges can be overcome with education, tools, and training."

....or by getting Numerus’s expert personnel to take care of it for you! 😊

By clicking "Accept," you agree to the use of Numerus and third party cookies and other similar technologies to enhance your browsing experience, analyse, and measure your engagement with our content.