Real world evidence: a reality check for pharma?

Rising demands on the industry has pharma firms calling for a more-effective, fit-for-purpose methodology for generating clinical data. Could the answer be real world evidence? Originally published in the November issue of Pharmafocus, Matt Fellows examines this promising research practice and how it could revamp the pharma R&D landscape.

The pharma industry is increasingly coming up against resistance as it struggles to fulfil the ever-increasing requirements of ‘wellness stakeholders’; a term referring to those individuals and organisations with a vested interest in the success and efficiency of the pharma lifecycle, from governments to payors to regulators to patients. The pressure is on for pharma companies, perhaps more than ever, to justify so many aspects of their products to these stakeholders. With cost-constraints tightening their stranglehold on R&D freedoms and changing demographics in the form of aging populations and rising chronic disease rates shifting the rules of the game, the industry is under pressure to evolve to keep pace with the demands of all of today’s wellness stakeholders.

According to the white paper Real-world evidence: A better life journey for pharmas, payers and patients: “Patients are looking for a better end result with their treatment. Providers are looking for data-oriented proof that the prescribed drug helps to optimise patient treatment, and brings added cost-efficiency and better profit margins. Payers (both government and private) are asking providers and manufacturers to prove and promote the benefits that they will reimburse for in their healthcare systems. Regulators, from the overall public health and well-being perspective, are also looking for evidence in a real-world environment.”

Expectations are high to deliver on these needs, and that calls for a solution that can meet the challenge, as Abhimanyu Verma, head of real world evidence and big data solutions at Novartis, points out:

“It is imperative that outcome-based and collaborative models be applied across healthcare systems to take a data and evidence-driven approach to ensure the optimal utilisation of healthcare resources to tackle the rising challenges and deliver the optimal patient outcomes. An approach is needed that brings together the different stakeholders – pharma companies, payors, regulators and governments working together collaboratively , rather than in siloed approaches , with the right joint incentives to deliver the common goal of positive outcomes for patients.”

All of these instances, defined by the requirements of modern healthcare, call for one thing: more evidence. The traditionally small-sample results presented by the tried-and-trusted clinical trial just won’t cut it. And that’s where real world data comes in.

According to the Association of the British Pharmaceutical Industry (ABPI), real world data is any data outside the controlled constraints of conventional randomised controlled trials to evaluate routine clinical practices – information harvested organically in the ‘real world’. This data is then assessed on its organisation, analysis and interpretation to generate real world evidence.

“It’s really any data that’s relevant which is derived from anywhere outside a controlled environment,” explains Andrew Roberts, director of market access and external affairs at Napp Pharmaceuticals. “It’s any data which could be relevant to the therapy area or the product that you’re looking to examine. I would define it as being that distinction as lacking a controlled environment such that you would find framed by a randomised controlled trial.”

So the nature of real world evidence is quite at odds with the data generated by the traditional clinical trial, as Roberts outlines:

“In a randomised controlled trial you’ll set quite distinct control boundaries that you’ll build into the design methodology of the study, and that’s always been purposeful to narrow the specific area that you’re looking to explore. Real world data is the complete opposite; it allows you to really explore all forms of data that may be relevant to the disease area.   It allows you to assess how effective a medicine has been in a real world setting. So you’re not setting boundaries by definition because you want to see how a product might behave with patients that lead a particular lifestyle with a particular condition.”

Real world data is nothing new. The industry has dabbled in it for a short while, but it is for the most part still finding its feet with the methodology. But only now is the notion really starting to become an area of great promise.

“It’s a huge opportunity,” Roberts tells us. “It’s starting now and it is a relatively new area, but it’s developing very quickly. Real world evidence has been something the pharma industry has been using for some time; I’d say the last five years we’ve started to see real world data used very effectively, certainly in the UK, to discuss disease with healthcare professionals. The opportunity now is how that expands.”

Why is it so promising?

Real world evidence has the potential to make a big impact on the industry as a whole and reinvigorate all stages of the pharmaceutical lifecycle: “With the increased availability of data and technologies available now and in the future, real world evidence presents an opportunity to start shifting to a future collaborative and integrated model,” Verma notes. “For pharma companies, the value leverage lies across the entire value chain – from research to commercialisation.”

Insight can be gained into new products at the research and development stage by comparing potential indications against established patient groups. Evidera’s Sreeram Ramagopalan and Dimitra Lambrelli told us: “Real world evidence can help inform development plans at early stages, identify unmet need in specific disease areas and specific patient populations, provide insight as to where to dedicate resource and how to best position new compounds on the market. During the discovery phase, it can be used to assess which indications could potentially benefit from the new compound and the health issues that remain to be addressed for these particular patient populations - for example, their unmet need and disease burden. Real world evidence can help understand which patients could have the most clinical benefit from a new compound. Linking real-world data with genomic or biomarker data can lead to the identification of new therapeutic targets.”

Roberts too attests to this: “Understanding how different people respond who have a particular disease can be very useful in targeting medicines to a particular subset of a disease population who may benefit the most from a product. If you’re looking to really define the cohort of patients that would gain the most clinically or economically from a treatment, then you can define those patient groups more easily using real world data analysis.”

Real world evidence also has the potential to affect the regulatory process, particularly in the rare disease sphere. A rare disease such as Duchenne muscular dystrophy faces an incredibly difficult route to market; because of the rare nature of the disease, drugmakers cannot gather the required evidence across an adequately sizeable patient base to satisfy the increasing need to prove treatment efficacy to regulators, meaning the drug never gets to market and cannot generate post-market data. Real world evidence can circumvent this issue by providing an existing pool from which to draw efficacy data.

It also has the potential to shorten approval times, as Ramagopalan and Lambrelli point out: “The adaptive licensing pathway by the EMA has the potential to lead to early conditional approval for a treatment without full efficacy data. Full approval would follow through the collection of real-world data after conditional launch showing the treatment being effective.”

And when a drug does make it to market, real world data post-launch can be used to assess the safety and efficacy of treatments, and can even be used to determine pricing strategies. With the current industry climate being one of caution and restraint, and with patients and regulators rightfully requesting treatments be sold at a fair and demonstrable price tag, real world data can provide drugmakers with the evidence they need to make potentially risk-free financial decisions, as outlined in A better life journey for pharmas, payers and patients:

“Payers in both government and private sectors are challenged by increasing healthcare costs. To counter this challenge, they are overly cautious regarding how they spend constrained budgets and are seeking return on investment justification in advance of making any outlays. Many are establishing their own therapeutic guidelines over and above what is required by regulation. To land in the right plan and price category, pharma companies need to provide payers with supporting evidence based on real-life data to document drug treatment efficiency and cost-effectiveness.”

Roberts highlights how stakeholder pressure continues to be a dominating factor in the industry and has the potential to shape drugmaker strategies, particularly when the ominous issue of cost-effectiveness is concerned:

“If pharma is going to be continuously challenged by payers to demonstrate cost-effectiveness, then understanding which patients the treatment will be most cost-effective in is very helpful when you’re talking particularly to payers. Most clinicians now have payer behaviour in some shape or form, so for them to understand which is the right type of patient to use the product, it’s very helpful information for them. 

“The one opportunity that real world data might do is allow us to advocate for a higher price for a medicine if we know categorically that using that medicine in a particular way with a particular patient reduces cost to society elsewhere,” he continues. “There’s some very good examples of how it took us 15 to 20 years to understand how particular medicines completely remove the need for surgical intervention. Take proton pump inhibitors for the treatment of acid-related disease. The use of proton pump inhibitors has meant some surgical procedures now never have to take place. It took us 15 years to understand that – real world data might have shown that within two. There’s a lot of examples in diabetes and in cardiovascular medicine where knowing that data early would allow healthcare to be transformed because you’d be targeting particular healthcare interventions to only the patients who need it as opposed to more empirical treatment which is largely how we treat now.” 

“If you believe the trend in most health systems is towards greater value-based pricing models,” Roberts continues, “then real world data is going to be absolutely crucial for you to demonstrate the true value of your medicine over time if you are to maintain price or even seek price adjustments. You’re going to need very strong data to do that, and that won’t necessarily come from randomised controlled trials, it’s going to have to come from real world data.”

Real world evidence is also of particular utility to the growing field of personalised medicine. As the industry strives towards this new and exciting field, real world data collection can greatly help to target specific patient groups and deliver on truly bespoke treatment to a much greater degree than would otherwise be possible.

“Personalised medicine allows you to explore outcomes that are derived from using that medicine in different patients who are at different stages of the same disease, and then the data you get from that you can then evaluate and turn into evidence to support the case that’s come from the randomised controlled trial, or to expand it,” Roberts tells us. “I think for personalised medicine in particular, understanding outcomes is critical, so if real world evidence greatly helps you understand outcomes then it’s going to be very important to have personalised medicine developed.

“Having the constraints of a scientific randomised study will give you a set of results, but as we’ve shown they’re not always easy to replicate, if you designed the same study but selected a different, random number of patients. So real world data allows you to examine far more data, but still try and draw useful comparisons. You’re free to explore far more data, which can be useful if you’re looking at personalised medicine because you’re looking to examine as much data as possible to understand how a particular product may behave in a much broader number of patients. It’s definitely very useful for personalised medicine because by definition you are wanting to find exactly where your medicine will be most effective.”

The tools at hand

Like in many other areas of pharma, new technologies have the potential to energise real world data collection in a way that would not be possible previously. Data can be pulled from a wide range of sources that offer retrospective observational information including disease registries, national and patient surveys, electronic medical and health records, administrative claims, prospective observational studies, and there is even rising interest in whether social media platforms can be utilised to gather relevant patient specifics. Electronic health records – particularly the UK’s Clinical Practice Research Datalink – present perhaps the biggest opportunity, as Roberts explains:

“Most of us now, whatever health system we sit in, we’ll have some form of electronic patient record. In some health systems it’s very highly developed and it captures just about every aspect of your disease, so every outcome that relates to your disease might be captured and coded in some way, so it makes it much easier to actually follow how particular patients behave through a disease, whereas in a previous era, the only way you could capture that data would be to identify the patient and track them over time using face to face methodologies: bringing the patient in for consultations and asking questions about their healthcare over a period of time. And that is cumbersome, it’s flawed, and it’s also hugely expensive.”

So delving into already established health databases provides a solution to the increasingly outdated data-gathering methods of the past, presenting a method to invaluable amounts of time and money for those in the pharma industry. A wide range of simple software exists to enable users to access and extract this data from existing databases: “Assuming you have access to data systems, then the software that allows you to search using simple keywords, simple diagnostic procedures can be searched for and then that data can be mined and gathered,” Roberts continues.

But it’s not extracting the data that’s the difficult part.

“It’s not a particularly complicated science to actually mine the data; the problem is always getting access to the database because it has to be anonymised and you can’t be extracting data that’s attributable to patients.” Because legislation has not yet evolved to accommodate the advent of real world evidence-gathering technology in the industry, accessing these banks of patient information at all is the main obstacle faced. And until that changes, Roberts believes the potentially huge impact of technology could be essentially neutralised for the foreseeable future, especially in the UK.

“At the moment we have a situation where we don’t have freedom of access to any database,” he explains, “and I think that is raises a very interesting question, because if we did it would transform what we understand about the disease and  about how medicines and other interventions could be used to improve outcomes in those diseases. I think until then you’re not going to see everything leap forward in a particular direction.”

Verma agrees that this is the big task: “The challenge is less of data generation but more of data integration across diverse sources to get a complete picture,” he explains. “It’s not as much a technical challenge, but one of evolving data standards, patient privacy and informed consent; establishing accepted analytical methods; creating collaborative distributed data networks; and taking a scientific and systematic approach to the analysis of the data once integrated. Advances in analytics - machine learning approaches for example - certainly help, but its not a silver bullet.”

“Personally, I think one of the things we need to do is have a more informed national debate about how we can use the data that we have,” Roberts adds. “The NHS incidentally is often cited as one of the best healthcare systems to undertake RED because we have this huge patient record; the electronic patient record database for patients in the UK with certain chronic diseases is one of the biggest in the world. Understandably people fear their own patient data being used for commercial purposes but we need to have the debate because it’s not as crude as that. There’s a way of regulating how the data is used but at the moment we’re underusing the data that’s sat there, and from a scientific point of view that’s got to move forward. The way the NHS purchases medicines and manages disease could be greatly enhanced if it used the data at its disposal more effectively, and it currently doesn’t do that.”

Similarly, wearable technologies are an up-and-coming technology development which, as in other areas of pharma, represents a huge potential to change the face of real world evidence, allowing data to be captured passively and remotely in real-time on a huge scale. But it may be a while before it lives up to its full potential in this particular application, according to Verma.

“Advances in wearable technologies, passive monitoring will allow for generation and capture of real world data in a more systematic way directly from patients. Having said that, it’s early days yet and these technologies are yet to mature for use in clinical settings; establishing clinical and accepted endpoints from such data is yet not proven.”

Ups and downs

Real world evidence clearly presents many invaluable advantages over the classic randomised controlled trial, tackling many of the shortcomings that the latter has increasingly shown in recent years while presenting a method by which to meet progressively more demanding requirements on all sides.

“The biggest opportunity you have with real world data is you can explore far more information far more cost-effectively, because nearly all of it is by definition derived remotely, so you’ve got potentially small teams of people mining lots of data from across quite a wide footprint, assuming you can gain access,” Roberts tells us. “So you have the opportunity to look at far more data than you could in a randomised controlled trial, because RCTs still rely on face-to-face patient interviews which are expensive and time-consuming, and whilst they are critical to demonstrate safety and efficacy, RWE can complement that by going much more widely and by looking at more patients with a particular disease to understand how their disease behaves and how a particular intervention may improve outcomes in patients with those diseases.”

But, because of the radically different approach employed by real world data-gathering, there are many ways bases on which, by its very nature, it cannot deliver:

“What you lose is the rigour,” Roberts continues. “In a controlled setting, it’s highly observational, so you’re dealing with the opportunity to explore in very fine detail what is happening to a patient. You don’t necessarily have that with real world data because you’re relying exactly on what’s being captured and it’s usually a past event, so you can’t shape the event and then come back and look again, because its real world, it’s already happened.”

Another issue is one that is less overt, and one which ultimately may require much more caution to overcome, as Verma points out:

“A fundamental challenge is that for most of the sources, the data is collected for a different primary purpose and is then used for secondary purposes. This introduces challenges with regard to the evidence that must be assessed carefully within the right context to draw the right conclusions keeping in mind the inherent data quality issues, biases present and the limitations of the methods and analytical approaches taken. A systematic and scientific rigour is essential lest wrong conclusions are drawn. It’s a classical case of being careful to distinguish between causality and correlation.”

So, taking into account the somewhat double-edged nature of real world evidence and the challenges it presents, could it ever replace randomised controlled trials? Roberts believes a more complementary approach is the way forward:

“I’m not suggesting for one minute that real world evidence would replace a randomised controlled trial in personalised medicine – far from it. But what it can do is help healthcare professionals understand more about where the medicine could be used in a targeted setting. So in most forms of medicine, real world evidence is complementary to randomised controlled trials because you still need the trial to prove safety and efficacy, as a principle.”

And this essentiality of the randomised controlled trial to the heart of the pharma lifecycle is echoed by Verma:

“Randomised controlled trials are and will remain the gold standard of establishing efficacy and safety of a therapy.  Real world evidence complements evidence established via randomised controlled trials to gain insights and evidence on how the therapy performs - effectiveness, efficacy and safety - in real world settings,” he remarks. “Real world evidence can also be used to enhance supplementary endpoints and, in some very rare disease conditions, possibly also be used to support registration claims. An effective way forward is to view the body of evidence in a patient population holistically across randomised controlled trials and real world evidence.”

A collaborative solution

This clear consensus is echoed by forward-thinking real world evidence experts throughout the industry; the assessment of both research methodologies presenting distinct strengths and weaknesses has led to the natural conclusion that the most effective system combines the strengths of both disciplines. Ramagopalan and Lambrelli outline succinctly how this can be beneficial:

“Real world evidence can be used to optimise clinical trials. Data can be sought on patient numbers to identify the sites/countries with the highest frequencies of patients to recruit from. Further, real world evidence can provide an overview of the patient population of interest to make informed decisions about inclusion and exclusion criteria for trials. These steps can help to reduce costs of running trials by speeding up patient recruitment.”

This school of thought has given rise to a number of prospective models which merge the two existing practices to form a working model which is fit-for-purpose. One of these is the pragmatic clinical trial:

“Pragmatic clinical trials are somewhat intermediate between real world evidence and randomised controlled trials and could become more and more common in the future to address the bias in real world evidence and the very controlled environments of randomised controlled trials,” Ramagopalan and Lambrelli explain. “Generally, pragmatic clinical trials involve a randomisation of treatment like randomised controlled trials but otherwise all other elements are kept in such a way so that patient care does not change as compared to how they would normally receive it in the ‘real world’. Pragmatic clinical trials incorporate outcomes that are relevant to patients and other relevant stakeholders, have no or minimal exclusion criteria for patients and the trial is monitored as part of standard care with little of the intensive control seen in randomised controlled trials.”

Another is the integrated RCT/RWE model, which shifts the traditional focus from post-launch clinical trials to clinical effectiveness data collection as a means of post-market research, outlined in Revitalizing pharmaceutical R&D: The value of real world evidence:

“Though the traditional randomised controlled trial model has proven successful for many years, it still relies on thinking and technical capabilities from the 1960s. The fully integrated RCT/RWE model emphasises a different approach: using new technological capabilities to generate clinical-level evaluation in a real world patient treatment environment — without predefined patient inclusion/exclusion criteria or artificial patient adherence measures.

“The new R&D model would emphasise achieving a reliable proof of concept with a good safety profile and an efficacy profile that would encourage the pharmaceutical company to continue the remaining evidence generation in the market. In other words, the next step would be collecting clinical effectiveness data rather than generating additional clinical efficacy information in randomised controlled trials. For example, the analysis of recent failures during phase III (late-stage attrition) is supportive of the new R&D model.”

A long but exciting journey

It’s clear that the methodology has the potential to optimise clinical research in a manner that modernises it to meet the demands of today’s industry, but it is equally clear that successful application in the future calls for greater adoption, particularly through standardisation, as noted by Ramagopalan and Lambrelli:

“Ultimately more data, more complete data and better quality data is required from all countries. For ease of analysis, features of individual databases should be standardised using a common data model.”

The road ahead is undoubtedly promising, and real world evidence looks to provide the industry with a silver thread to pull it away from the growing burdens of shrinking budgets, rising costs and impotent research methods. It won’t perhaps be the smoothest journey, and there is a battle to be fought on widespread adoption of the method, but its benefits cannot be understated.

“The industry is already starting to move in this direction and we are in the initial stages of a long but exciting journey,” Verma ponders. “The best way is to work and integrate collaboratively across the different siloes that typically exist in pharma organisations and externally in broader healthcare systems context. Nothing succeeds like demonstrating proof points to build confidence in approaches, methods and overcome what are common challenges to the industry – where consortia and collaborative approaches can be very effective.

“There is currently a tremendous amount of hype and excitement on one end and scepticism on the other on the value of real world evidence. The reality is somewhere in the middle and a balanced, collaborative approach both within companies, within industry and across healthcare stakeholders is a must – especially by engaging patients directly in support of research and outcome goals.  It is a long journey but definitely achievable with a collaborative approach. Balancing the longer term approach with more immediate shorter term drivers and needs is going to be critical for a sustained path to achieve outcomes based models.”  

Source