The randomised clinical trial (RCT) has been the gold standard in drug development for decades, and there is no doubt that it is a powerful tool to determine whether a therapy is safe, effective and worth prescribing – at least in a carefully selected group of people.
Human beings are however a very diverse bunch and it has become equally clear that while clinical trials are important, they provide a poor picture of how a new therapy will fare when it is being used by doctors to treat patients in the real world.
In the controlled environment of a clinical trial patients are carefully selected, often because they have the best chance of showing that a therapeutic intervention will work, and considerable care is taken to exclude the random variables – a lack of compliance, concomitant use of other medicines or therapies or other chaotic factors – that might affect the result.
So-called real-world evidence (RWE) may not be as effective as RCTs in collecting efficacy data, but the large number of patients involved can provide a better way of monitoring safety across the treated patient population for pharmacovigilance purposes and can also provide additional pharmacoeconomic insights, for example looking at the use of resources or financial considerations.
Moreover, RWE can also provide information on the effectiveness of clinical interventions on a broad cross-section of patients – above and beyond the inclusion or exclusion criteria deployed in the RCT setting – as well as other variables such as how well patients adhere to therapy, how it slots into clinical workflows, whether there are any reimbursement barriers to treatment.
In essence, a clinical trial can tell us what a drug does, while RWE can provide the context that tells us whether what it does actually matters.
So what exactly do we mean by RWE? In essence it means the collection of safety and efficacy data in a non-interventional or observational setting, generally in a post-marketing setting, and the practice discipline lends itself to countries with well-structured public healthcare systems, particularly those that have adopted electronic health records (EHR).
For many proponents of RWE, the most valuable approach to generating data is the patient registry, which provides a database of patients receiving a particular medicine or therapy, including their demographic information, other treatments or management approaches they may receive and outcomes, both those reported by the patient and their physician.
Last year, Merck & Co launched a global registry to evaluate the experiences and outcomes of approximately 20,000 patients with type 2 diabetes in real-world settings over three years, taking into account not only clinical measures such as glucose control but also use of healthcare resources, adherence to medication, quality of life and patient-reported well-being.
Similarly, the industry-sponsored PREFER and GARFIELD registries looking at the use of novel oral anticoagulants (NOACs) in patients with atrial fibrillation – reported last year – revealed that prescribing of NOACs was rising, with a sharp decline in the use of older drugs like warfarin. Overall, anticoagulation was still only being used in around two-thirds of patients – contravening clinical guidelines and raising the risk of strokes.
A clinical trial can tell us what a drug does, while real-world data can provide the context that tells us whether what it does actually matters
Meanwhile, it seems companies are increasingly starting to build RWE into their product life cycle management. Pfizer recently set up a specialised unit – called global health and value – that combines and integrates real-world data, health outcomes research, market access studies and pharmacoeconomics under one roof in order to support reimbursement negotiations, said Geno Germano, president of the global innovative pharma unit group, at the Morgan Stanley Healthcare conference last year.
Furthermore, Marco Taglietti, chief medical officer of Forest Laboratories (now part of Actavis), said last year that the company has started to routinely carry out health economic studies along with the use of real-world data. The aim is “to help the commercial organisation achieve its goals. That means creating data … that helps market access,” he explained.
And contract research organisations are also starting to offer real-world evidence capabilities to pharma clients. ICON has been working on the development of an informatics ‘hub’ that will “harness clinical and real world data from both existing and new sources,” according to chief executive Ciaran Murphy. Other CROs such as UK-based Cisiv have developed tools specifically for design, data capture and management of observational studies.
There is certainly no intention that RWE could replace RCTs
Complementary role
There is certainly no intention that RWE could replace RCTs. “Just like precious metals – gold, silver, and platinum – the different types of evidence all have value,” said Jennifer Graff, who is director of comparative effectiveness research at the National Pharmaceutical Council in the US.
Trials can confirm something that we learned from RWE, she noted. For example, a recent finding that eating peanut butter in early life reduces the risk of allergies later on were hypothesised by RWE and then confirmed by an RCT.
In other cases, both RCTs and RWE may be used side-by-side. For example, we know a lot about how treatments for lung disorders work from the regulatory approval studies based upon RCTs, according to Graff. However, RWE can help to confirm treatment benefits and how they might extend treatment effectiveness to older patients.
“In other situations, RWE may be the only ethical approach to getting information,” she told PME, referring to a recent study in the Journal of the American Medical Association (JAMA) which compared breast cancer survival after bilateral mastectomy with other surgical treatment and covered 99% of all breast cancer cases in the state of California.
“You couldn’t study that via a randomised trial because few patients or providers would sign up for a trial with a 50:50 chance of undergoing a major procedure that takes into account many patient-specific preferences,” said Graff. “It wouldn’t be ethical. Here, RWE is likely the best we will ever get.”
The emergence of RWE also ties in with another of the key trends in clinical research, namely the shift towards patient-centricity. With healthcare changing in a direction which is looking more at patient outcomes and less at targets, the pharma industry has also started to include patients alongside prescribers and payers in its thinking.
“Patients are asking about how treatments work for patients like them, and providers are trying to understand how different treatment options may impact outcomes that matter to their patients – such as number of falls versus a walk test – and what treatment options may best match those needs,” said Graff.
Equally, payers want to know how a treatment works compared to alternatives and the impact on total cost of care, and policy makers want to see the impact of care coordination. “RCTs are unlikely to be big enough, robust enough, or efficient enough to answer all of these questions,” she pointed out.
Informing payer decisions
A recent AccessPoint report on RWE by IMS Health suggested that while there are dozens of case studies involving the use of RWE assessments in recent years, “the application of RWE to really improve the efficiency of healthcare delivery remains uneven and siloed.”
Regulators also seem on board with the idea. Sir Alasdair Breckenridge, former chairman of the UK Medicines and Healthcare products Regulatory Agency (MHRA), said recently that moves towards conditional, early licensing of medicines will require very good collection of post-marketing data “in which real-world efficacy and safety can be judged outside of the clinical trial setting”.
Moreover, a side benefit of the recently-implemented Early-Access to Medicines Scheme (EAMS) – which provides a way for patients with serious diseases to get access to a new drug before a European licence is granted – is that life science companies would gain “greater exposure to prescribers in the UK and the opportunity to collect real world data on the profile of the medicine,” according to the MHRA.
And while industry and regulators seem to be getting on board with RWE, it also seems that health technology assessment (HTA) agencies too are starting to take notice, although IMS suggests that “examples of them accepting industry-generated RWE or working collaboratively with pharma to generate RWE are few.” Overall, just 6% of assessment carried out by HTAs in the UK, Canada, Australia, France and Germany have included a review of observational data, according to research by health economics consultancy Context Matters.
There are some encouraging instances however. In the UK, the British Society for Rheumatology (BSR) established a nationwide register for patients with rheumatological disorders treated with biologic agents which was specifically designed to assess long-term toxicity from the use of these agents in routine practice. The data – collected over five years – was used by the National Institute of Health and Care Excellence (NICE) to develop guidance on the use of TNF-alpha antagonists in the management of rheumatoid arthritis.
Meanwhile, in the US there has also been an increase in the use of real-world data in HTA and reimbursement processes over the last five to six years, In 2009, one of the US’ largest health insurance companies, WellPoint (now Anthem), started talking about RWE in public forums, and in 2012 the Academy of Managed Care Pharmacy (AMCP) Format Committee issued an addendum to their guidance for formulary submissions that outlines the importance of RWD being used by payers.
“Research NPC conducted with researchers at Brigham and Women’s Hospital found that US state Medicaid directors consider real-world data in making coverage policy,” said Graff. “However, the consistency and comfort with this type of information is still evolving, in part because the tools to know what ‘good’ RWE looks like have been evolving,” she added.
Pharma companies remain somewhat risk-averse to generating and sharing RWE with payers – at least in some instances – as they are worried about how the information will be interpreted and used, according to IMS, which also notes that the complexity of extracting and handling real-world data and a lack of consistent methodologies for RWE studies are brakes on adoption.
Over time, the availability of tools and training like the GRACE checklist – which guides the assessment of observational studies of comparative effectiveness in terms of their quality and usefulness for decision-making – will help payers understand when the RWE is credible and relevant to their decisions and how to incorporate it in an HTA process.
There is much interest in seeing how big data could be interrogated to provide RWE
Next steps
In the future, there is considerable interest in seeing how big data projects such as the Clinical Practice Research Datalink (CPRD) – an observational and interventional database that operates as part of the UK Department of Health and has more than 5 million active patient records – could be interrogated to provide RWE. There is also a lot of potential in the UK’s care data initiative – which is trying to extract anonymised patient information from primary care – once privacy issues are resolved.
Mobile health and wearable technologies also provide opportunities to enhance data collection in real-world settings, although face the same questions as in clinical trials – for example is the data credible and accurate, and how can gaps in the data set be handled? – but magnified by the much higher number of patients involved.
“Stakeholders recognise the promise and potential of using RWE in health care decision-making, especially to answer certain types of questions,” said Graff. “However, we have a way to go before the promise of real-world evidence is seen and this type of information is used consistently across health decisions.”
The last word to French surgeon and 1912 Nobel Prize winner Alexis Carrel who put it succinctly: “a few observations and much reasoning lead to error; many observations and a little reasoning to truth.”