Providers, payers, regulators and patients increasingly want to know more about drugs than whether they are safe and effective in randomised controlled trials (RCTs). Having been stung by drugs that shine in RCTs but under deliver in day-to-day practice, these stakeholders want to know what effect therapies will have on people in the real world, where comorbidities, non-adherence and other issues controlled for in trials affect outcomes.
These demands create some difficulties for drug developers. Companies must run highly controlled trials that clearly detect safety and efficacy signals to get drugs approved. Yet these studies are often too short and too restrictive in their inclusion/exclusion criteria to answer all the questions asked by other stakeholders.
“Efficacy studies show us in very basic terms how effective the treatment is from a biomedical perspective. But then when the treatment is out there in the real world, we know it’s affected by many things,” John Weinman, professor of psychology as applied to medicines at King’s College London, said.
The result is a disconnect between the data companies’ need to generate to win approval, and the data they need to persuade people to pay for and use their drugs. Companies have responded by gathering, generating and analysing data outside RCTs to produce evidence of efficacy and value in the real world. More than half the leading life science companies polled by Deloitte in 2017 were developing or significantly improving their real-world data (RWD) capabilities.
These investments are expected to yield data that, when analysed, creates evidence of how a typical patient responds to a drug and the value it provides to healthcare systems.
Generating real-world evidence
Companies are stepping up their interest in real-world evidence (RWE) at a time when access to data from outside RCTs is expanding quickly. Traditionally, drugmakers have run observational studies such as registries to comply with postmarketing requirements and track the long-term progress of patients. Other companies have mined existing data sources, notably administrative claims data from insurers and governments, to assess how their products are used in the real world.
These resources remain important parts of the RWE toolkit but they have limitations. Claims data lacks clinical outcomes. Some patient registries are unreliable and have limited follow-up. Increasingly, researchers are mitigating these weaknesses by combining the resources with data from emerging sources, such as electronic medical records (EMRs).
EMRs and more comprehensive electronic health records (EHRs) have been around for decades but have really taken off in recent years due to government actions and technological progress.
The rise of EMRs raised hopes for a golden era of real-world research but the initial reality was more prosaic. While in theory EMRs make it possible to share, aggregate and analyse contemporaneous, comprehensive medical data, in practice widespread use of free-text fields, a lack of standardisation and barriers to interoperability blunted their impact. However, there are signs this is changing.
“EMRs are becoming increasingly attractive options,” Jason Shafrin, senior research economist at Precision Health Economics, told PME. “Increased use of standardisation and the use of machine learning for the open text fields has made them much more useful.”
Researchers can further enhance the utility of EMRs by combining them with other data sources. The practice of using claims data to fill in gaps in EMRs is well documented. Today, researchers can also pull in laboratory results and data from apps, wearables and other technologies to generate a more detailed picture of how patients are using and responding to therapies.
Using real-world evidence
RWE has distinct applications pre- and post-approval. Before a drug comes to market, sponsors use the resource to evaluate unmet needs, identify subpopulations of patients and perform other tasks that inform their R&D strategies and the design of RCTs. In the future, RWE also looks set to play a bigger role in applications for approval or label extensions. Once a therapy wins approval, companies turn to RWE to understand how the drug is used and show its effect on health and economic outcomes.
The evidence has applications in multiple therapeutic areas. Roche showed the value it places on real-world cancer evidence when it struck a $1.9bn (£1.3bn) deal to acquire oncology EHR provider Flatiron Health. The takeover was driven by a belief that “regulatory-grade real-world evidence is a key ingredient to accelerate the development of, and access to, new cancer treatments,” Daniel O’Day, CEO Roche Pharmaceuticals, said. Yet, RWE originally established itself in chronic diseases.
“It can help differentiate you in a crowded space or, for more complex, chronic diseases, it can show that the benefits seen in a short-term trial with a narrow patient population really do apply to a broader population of interest to the payer,” Shafrin said.
Companies that fail to make the real-world case for their drugs risk having someone else shape the narrative. In the past, providers and payers had a restricted view of how drugs worked in the real world. Each doctor had insights into how their patients responded but the sample size was small and the breadth of data limited. That era is ending. Now, technologies capable of continually capturing data are expanding and democratising access to insights about real-world safety and efficacy.
Diabeter, a Dutch chain of type 1 diabetes clinics bought by Medtronic in 2015, is an example of this trend. The clinics increased the proportion of patients who achieve key outcome targets through the use of an integrated, value-based and technology-enabled care model. One facet of the model is the collection and analysis of RWD from devices such as glucose meters.
“Whatever technology patients are using, we have a way to pull in that data … and then have it come through a pipe so it populates our practice management system,” Sue Willman, vice president of global diabetes integrated care solutions at Diabeter’s parent company Medtronic, told PME.
The first use of the data is to inform decisions about the care of individual patients. Diabeter then goes a step further by building population dashboards to facilitate the data-driven evolution of care pathways. This entails looking at the outcomes that therapies, coaching and other interventions achieve in the real world to guide how patients are treated in the future.
The future of real-world evidence
The challenges of translating the Diabeter model to new geographies and therapeutic areas mean it is likely to remain an outlier in the near term but echoes of its value-based approach are entering the mainstream. Notably, manufacturers and government and private payers are responding to concerns about drug prices by tying payments to outcomes. Novartis secured such an outcome-based model for its CAR-T therapy Kymriah. Spark Therapeutics has a similar arrangement for its gene therapy.
These agreements make real-world efficacy a requirement for full payment for a drug. If the product fails to work as expected, the manufacturer will not recoup the sticker price. The nascent model is likely to evolve over time but the rise of gene therapies designed to provide a one-dose cure and other products with dramatic efficacy and similarly dramatic prices suggests it will stick around in one form or other.
Some people envisage an even bigger role for RWE. One idea floated in academia is to replace phase III trials with real-world studies in which typical patients are randomised to receive the experimental therapy or a control. While that seismic shift is unlikely to happen soon, emerging regulatory acceptance of RWE means the scene is set for a gradual change in how drugs are assessed pre- and post-approval.
When combined with the rising demand for non-RCT data from payers and providers, the evolution of regulatory attitudes will establish RWE as a vital piece of R&D and commercial strategies for years to come.