Part 1 of a 2 part series: Unlock valuable insights throughout the lifecycle
Pharma is undergoing an ‘evidence revolution’ as real-world evidence breaks out from the confines of market access and infiltrates every part of the life of a medicine, from discovery through to late-stage commercial. This revolution has, in part, been driven by technology – by the availability of electronic patient-level data and the ability to analyze it – and, in part, by market trends, as stakeholders clamor for ever more reliable proof of the value of new medicines.
“Few people yet realize the capabilities of the new data-collection technologies and how they can impact not only clinical research but also discovery research. They can transform the entire biomedical research enterprise, making innovation faster, more powerful and radically cheaper,” says Bernard Munos, Senior Fellow at FasterCures, an ‘action tank’ focused on speeding new therapies to market. “The speed and scale at which RWE will catch on, and the resulting cost savings will be unprecedented. Pharma is not prepared for it.”
Nigel Hughes, Scientific Director at Janssen, agrees that the potential of RWE is much larger and more diverse than some people realize, when you’re able to access and manage it. “We need to do more as a society to illustrate that; all too often we forget to start with the question and not the data. What is it we are trying to answer, and what is the best way to address it?”
Questions abound. How are companies currently using real-world evidence? Where do the greatest opportunities for harnessing RWE lie? What barriers to successful implementation are companies facing? What will companies look like once this revolution has hit in full force?
Getting ahead of disease
The problem with our current approach to health is that it starts with disease; we go to see a doctor when we get sick. By the time we receive a diagnosis for a disease that may have been simmering beneath the surface for years, the prognosis may be poor. For most diseases, we have no presymptomatic natural history.
Monitoring patients at home instead of in hospitals could result in cost saving of as much as 80%. The greater convenience and ability to recruit patients further afield from hospitals would speed enrollment, further reducing costs and improving data quality.
“This is a big problem that data from biosensors can help solve,” says Munos. “With Alzheimer’s, for example, many scientists believe that the disease silently undermines the brain for many years, but we have no data. Now, we can start gathering rich, longitudinal data on healthy-looking patients; If some of them later develop the disease, physicians can look at the data in reverse and, perhaps, figure out where things went off-track. We have thrown over 355 drug candidates at Alzheimer’s and still don’t know what causes the disease. If we had been able to collect fine-grained RWE during these trials, we would be much further ahead with our ability to understand the disease process and identify subgroups of responders.”
He adds that analyzing RWE could, in time, be performed by supercomputers like Watson, which could stream patient data in real-time to identify signal patterns that could be matched to diseases. This ability to advance disease discovery also offers the opportunity to enroll pre-symptomatic patients, and refocus data gathering in a way that will help investigators better match and randomize them.
The trend towards patient-friendly endpoints that can be monitored by biosensors (eg, motion, mood, effort) will likely have a huge impact on clinical research, adds Munos, offering opportunities to significantly reduce the $100 billion a year currently spent by pharma. “Monitoring patients at home instead of in hospitals could result in cost saving of as much as 80%. The greater convenience and ability to recruit patients further afield from hospitals would speed enrollment, further reducing costs and improving data quality.” It also offers regulators an opportunity to track safety in real time. “They’ve been very supportive,” he says.
Hub and spoke
Given its transformational potential, it is not surprising that many pharma companies are proactively looking at how they can best harness real-world evidence.
Some companies have created centralized specialist RWE groups, says Thy Do, Head of Data Generation Oncology and IBD, Europe and Canada, at Takeda. “Many functions utilize real-world evidence – medical affairs, health economics, safety, even clinical development. To accommodate this, a hub-and-spoke model is efficient as it offers a standardized approach to the analysis and preparation of evidence. A centralized group can do the data collection and analytics to generate the evidence that is then dispersed within the organization as needed,” he says.
While accessing, exploring and analyzing vast amounts of data sounds complex, the timely generation of real-world evidence is becoming easier thanks to advances in data management and analytics, says Do. Yet, considerable data challenges remain. “Electronic health record and claims data are generally not optimized to support research. Retrospective data pose some pitfalls; to understand the biases – in order to interpret the data – you need to know their provenance and how they were captured. For example, you need to know how reimbursement guidelines might have impacted the data.”
While it can be more expensive to conduct – for a pragmatic trial, you need a larger sample size to find an effect because of the more heterogeneous population – if you can accommodate the larger sample size within a single hospital or center, for example, then it would be less expensive.
There are other issues beyond the validity of the data, he says. “Another challenge is how to holistically interpret the evidence when different methodologies have been used and the data have not been systematically captured. Having direct access to the raw data for re-analyses or additional analyses also remains a challenge. A lot of countries prefer their own data, so a big challenge is anticipating when country discussions with payers or the scientific/medical community, for example, need to be anchored in country-specific data. This can be quite challenging in countries with strong data privacy laws, where access to data is limited.”
Do is a proponent of prospective trials. “A well-designed and well-executed randomized clinical trial provides the gold standard when it comes to evidence, but it can sometimes cause generalizability issues due to strict enrollment criteria and patient types. It is seldom the case that the criteria to enter a clinical trial are exactly the same as those used by healthcare providers to prescribe the therapy once it is marketed. It is this difference that is driving the industry to use RWE to prove the value of their medicines outside the confines of a clinical trial.”
He believes the industry should move towards pragmatic clinical trials. “While it can be more expensive to conduct – for a pragmatic trial, you need a larger sample size to find an effect because of the more heterogeneous population – if you can accommodate the larger sample size within a single hospital or center, for example, then it would be less expensive.”
Results from larger, more heterogeneous patient populations would compare better to routine clinical practice and generalize better to patients in the real world, says Do. “Payers have so much more influence and they are pushing harder for real-world effectiveness, in addition to evidence of treatment efficacy. The impact of this is that the industry must now deliver a broader package of evidence to increase the likelihood of commercial success for a therapy. An efficacious therapy won’t lead to commercial success unless it’s reimbursed, so companies must be forward-thinking when they develop their commercialization programs in order to include relevant RWE in discussions with reimbursement bodies and the scientific/medical community.”
To deliver this, investment must be made in RWE capabilities, both people and infrastructure. “We need to find epidemiologists or health researchers that are used to dealing with real-world data, not just clinical data. This is a growing trend; clinical programs are pulling these people into their core teams because it’s vital to ensure we have a broader perspective at the table. On the analytic infrastructure side, companies are securing their own data servers to facilitate in-house data analytics, such as predictive modelling, which takes a lot of computer resource time. There’s a lot of investment to be made on that side too, with the possibility of partnering externally as well to expand capabilities (eg, with IBM Watson).”
For part 2 of the series, click here.