Case Study: Oxford PharmaGenesis
“Using Elicit, we recently conducted a ‘rapid literature review’ investigating 40 research questions across almost 500 papers. In the past, we would never have been able to conduct a rapid literature review of that scale. This technology allows us to offer clients deeper insights across larger evidence bases than ever before.” Kim Wager, Scientific Director, AI and Data Science Team, Oxford PharmaGenesis

We're excited to introduce our research collaboration with Oxford PharmaGenesis, an independent, global HealthScience communications consultancy that advises 8 of the top 10 global pharmaceutical companies.
Oxford PharmaGenesis integrates their deep understanding of evidence synthesis and medical communications with Elicit's AI-powered literature review. Now, they can deliver literature reviews at unprecedented scale.
Before Elicit: Limited by capacity
Oxford PharmaGenesis faces heavy demand from pharmaceutical clients for a range of literature reviews to inform clinical development plans, identify promising therapeutic targets, analyze their clients’ competitors, and support access to new drugs.
Kim Wager, Scientific Director, said “Many projects you start in the pharma space begin with a literature review of some sort. These can range from comprehensive analyses of every relevant paper, through to rapid reviews to gain insights into a particular issue. Whatever the need, the scale is limited by the capacity of experts to screen and extract the necessary information”.
Marta Radwan, Senior Consultant, gave an example: “A client recently asked for a targeted literature review of treatments for dyslipidaemia. There’s no one definition of the disease, and there are quite a few questions we are hoping to answer. We found 26,000 papers in our initial search”.

This would have placed an extremely high demand on human screeners. Ciaran Wright, Scientific Director, added “literature review projects like this always take a lot longer than the client expects. The discovery and extraction phase consumes a lot of expert resource that could be better spent on analysis”.
To meet clients’ budgets and expectations, the usual approach is to help the client refine their question and scope, but often this isn’t optimal. Wager added: “Sometimes, it’s difficult to refine the client’s research question without actually doing the research. There ends up being a lot of back and forth to identify promising research directions”.
Choosing Elicit: “Elicit’s tech is just better”
Oxford PharmaGenesis knew that AI could transform literature review and tested several AI tools.
However, the other tools failed to meaningfully increase throughput and improve accuracy on the most labor-intensive step of literature reviews: data extraction. After experimenting with Elicit, Tomas Rees, Director of Innovation, believed that “Elicit’s tech is just better” for data extraction. “The results are reliable, and I can easily trace any data point to a primary source to be able to confirm things if we need to. That’s important when our reputation is on the line”.
Wager added, “Historically, PDF parsing has been challenging, which creates all sorts of problems for data extraction from scientific papers, but Elicit seems to have really nailed it.”
Wager also praised the support provided by Elicit’s research-trained customer success team. “They aren’t afraid to dive into the details and make sure we get the results we need”.
Today: Elicit’s AI enables Oxford PharmaGenesis to deliver enhanced literature analytic services
With the help of Elicit, Oxford PharmaGenesis can now deliver literature reviews at an unprecedented scale.
For example, a leading pharma company with a drug in late-stage trials asked Oxford PharmaGenesis for a data extraction of study methods and intervention details across a large set of papers for a competitor analysis. “We initiated the project on a Monday, and they asked for results by Friday” said Wager. With Elicit and human-in-the-loop validation, Wager was able to deliver a first draft of extractions by Thursday. He added, “the client was thrilled to see this much progress on a huge task. This accelerated extraction process allowed our experts to spend more time on in-depth analysis and deliver greater value”.
More importantly, Oxford PharmaGenesis can now offer entirely new kinds of literature analytic services. “There is a whole class of reports where we can be flexible on methodology, as long as results are reliable and robust” Rees said. “For example, literature monitoring, landscaping reviews, targeted literature reviews, and rapid reviews. These can provide critical insights into strategic planning, but the scope is often limited by capacity constraints. With Elicit, we are able to deliver more broad-ranging evaluations and enable more informed planning”.
He added, “Many clients actually want, and are now expecting, us to use AI”.
Earlier this year, a major pharma company asked for all epidemiological data, biomarkers, and prevention therapies related to transient ischaemic stroke, a common neurological condition. Wager said, “Using Elicit, we recently conducted a ‘rapid literature review’ investigating 40 research questions across almost 500 papers. In the past, we would never have been able to conduct a rapid literature review of that scale. This technology allows us to offer clients deeper insights across larger evidence bases than ever before”.

“With Elicit, we can ask detailed questions about study quality like Cochrane’s Risk of Bias and ROBIS assessments quickly and simply, rather than doing it manually.” He added, “Now I can spend that time on quality control and offer a better review than what was possible before. That quality is why clients come to us”.
He was also pleased that it was easy to register this work with PROSPERO: “Our protocol was one of the first AI-assisted literature reviews to be registered prospectively with PROSPERO. It got through fine”.
Looking forward: Scaling Elicit across the company
Beyond enabling larger scale research and new products, Wager also detailed how Elicit makes it easier to collaborate with clients: “We have a client with a highly specific protein inhibitor that could have a considerably better safety profile than alternatives. They want us to explore the nitty gritty of all the genomic, proteomic and interactomic data. This type of large project is highly iterative and discursive. It’s great that Elicit allows us to rapidly pivot the project as the client wants to explore specific leads. This is the kind of analysis that leads you to finding new indications”.
Wager is excited to explore the opportunity of scaling Elicit across the company. “The next step is to build Elicit into workflows of reviewers in our most important divisions, like the Value Demonstration Practice, and our Evidence Review and Synthesis Centre of Excellence”.
Summing up his experience, he said, “I'm very positive, very upbeat. I think we need to use it whenever we can”.