Real-world evidence’s evolution into a true end-to-end capability

Executive summary

Since 2017, Deloitte’s real-world evidence (RWE) benchmarking studies have signaled the biopharmaceutical industry’s efforts to expand the use of real-world data (RWD) and RWE across the product life cycle.¹ Beyond satisfying postmarketing requirements, for instance, biopharma companies have expanded their inclusion of RWE in regulatory submissions. During the COVID-19 pandemic, RWD/E played a key role in enabling biopharma companies to innovate and bring novel vaccines and therapies against this highly contagious disease to market in record time. RWD/E was invaluable to the scientific community to understand the efficacy of COVID-19 diagnostic tests and repurpose on-market drugs to save lives. Major regulatory agencies across the globe also have stepped up their efforts to advance RWE use in regulatory submissions, further expanding its importance.²

RWE has evolved into a true end-to-end capability

Since 2017, our RWE benchmarking studies have signaled biopharma’s growing focus on building capabilities to leverage RWE across the product life cycle.¹¹ This year's survey results suggest RWE has evolved into a true end-to-end capability, with 90% of survey respondents saying their organization is attempting to use RWE for decision-making across the product life cycle.

Historically, biopharma’s health economics and outcomes research (HEOR), epidemiology, and medical affairs experts have used RWE to understand disease progression, monitor patient safety, and assess clinical and cost-effectiveness. Not surprisingly, our survey data shows the bulk of RWE is generated for these functions even today. However, beyond these traditional uses, more than 30% of survey respondents reported that R&D and commercial are the top two functions that their organizations generate RWE for. Top use cases include informing the clinical trial design and site selection, understanding the heterogeneity of treatment effects, and informing pricing and forecasting assumptions.

The growing traction around IEP has also expanded the incorporation of RWE in the evidence planning process for assets. IEP enables cross-functional collaboration to create a common plan to leverage randomized clinical trial data and RWE approaches to meet the evidence needs of multiple stakeholders including HCPs, regulators, and payers, across the product life cycle. This enables companies to look at evidence planning holistically, make more strategic choices, and use resources efficiently to generate the most impactful RWE. Half of our survey respondents said their organization is actively undertaking IEP for one or more assets (figure 1).

Furthermore, future end-to-end evidence generation approaches could be modeled on the industry’s use of RWD/E during the pandemic, which expedited patient access to novel vaccines and therapies (see sidebar, “RWD/E during the pandemic,” for more information).

Technology investments have enabled scaling the use of RWD/E

As part of building an end-to-end RWE capability, biopharma companies have invested in technology capabilities (e.g., RWE platforms and tools) that are beginning to pay off. Our survey data suggests such capabilities are becoming increasingly common, enabling new ways to generate insights and evidence, scaling to more use cases, and reaching a broader user community.

Advanced analytical tools, self-service applications, and knowledge management systems are central components of a mature RWE capability, and most survey participants report that their organizations already have these in place today. More than two-thirds of surveyed respondents stated that their organizations provide a comprehensive suite of analytical tools for statisticians, data scientists, and analysts. The greater availability of self-service analytical applications also has enabled business and citizen data scientists (who are data-savvy but can’t necessarily write code to generate insights) to quickly run RWD analyses (e.g., cohort creation, descriptive statistics, etc.) through user-friendly interfaces. In this year's survey, more than 80% of respondents reported that their organizations have self-service analytical and visualization tools compared to 58% in the 2020 survey.¹⁴

Over the last several years, we have observed a growing use of knowledge management systems across the industry that are democratizing RWD/E use across the product life cycle. These valuable solutions provide transparency into RWD assets and the corresponding RWE that’s generated, enable consistent use of business rules and methods, and streamline reviews and workflows critical to evidence generation. About 65% of this year’s survey respondents reported having a knowledge management system compared with just 35% in our previous study.¹⁵ Furthermore, the remaining respondents indicated that their organizations plan to develop knowledge management systems, which are essential to increasing visibility into available RWD assets and making RWD analysis more transparent across the organization.

Survey findings also show that technology investments have laid the foundation for organizations to realize more significant benefits from using RWD/E. More than three-fourths of respondents reported that their organization's technology investments have created efficiencies in RWE generation (e.g., reducing time to insight) and enabled the use of new analysis methodologies (e.g., the use of AI).

Adoption of RWD/E in R&D has accelerated as the application in commercial expands

In our previous RWE benchmarking study, most respondents expected the importance and application of RWD/E in R&D to increase significantly over the next two years. This year’s survey data shows organizations have already realized benefits from RWD/E use across several R&D use cases (figure 3). More than half of our survey respondents have used RWD/E to enable a data-driven understanding of disease progression in populations of interest, support label revisions related to safety, and make better decisions on development strategy.

Additionally, nearly two-thirds of survey respondents reported that their organization had used RWD/E to assess disease incidence across socioeconomic, racial, and ethnic groups to improve trial diversity. The disproportionate impact of COVID-19 on racial and ethnic minorities fueled an ongoing need for trial populations that better mirror the real-world demographic of the populations affected.¹⁶ Deloitte’s research with Pharmaceutical Research and Manufacturers of America highlighted a sustained commitment among biopharma companies to address clinical trial diversity. Relatedly, the FDA recently released guidance on enhancing clinical trial diversity, making improving research equity a pressing imperative for all sponsors.¹⁷ To ensure that research is diverse and representative, industry leaders should consider taking bold steps to establish data-driven and inclusive trial recruitment processes.

Most of the interviewees opined that the pivotal role of RWE during the pandemic, coupled with the increased regulatory openness to its use, is likely to accelerate RWD/E application in R&D. During the next two to three years, more than half of the respondents expect their organizations to leverage RWD/E more frequently for more innovative and impactful R&D use cases. These include building regulatory-grade external control arms (ECAs), supporting label revisions related to product effectiveness, designing adaptive trials, and tokenizing data for long-term patient follow-ups (see sidebar, “Two ways that biopharma is leveraging RWD/E,” for more information). Doing so requires collaboration between industry stakeholders to help streamline RWE use for regulatory submissions (see sidebar, “Developing a collective experience to evolve RWE use,” for more information).

As companies develop more targeted therapies focused on narrower patient populations, enrolling patients in a traditional control arm is becoming increasingly complex from a cost, time, and ethical standpoint. Constructing regulatory-grade ECAs using data collected from a patient’s electronic health record (EHR), administrative claims, and other sources is of particular interest to cut development time, reduce costs, and ensure participant access to lifesaving therapies.²¹ Such external control arms could create a more representative dataset to study drug effectiveness by enabling the inclusion of patient data across diverse populations. However, embedding ECAs more systemically in development processes requires cultivating an appetite for risk-taking among development teams and investing in technologies and processes for ECA design and implementation.²²

Enabling many of the most impactful R&D use cases requires linking disparate data sources (EHR, claims, and clinical trial data) together at the patient level to create a more robust long-term view of the patient. Tokenization technologies enable biopharma companies to anonymously link disparate data sources together at the patient level, creating a rich dataset while maintaining HIPAA compliance. To date, most linkages have been one-off projects within biopharma companies. But some organizations are developing systems and processes to scale this capability across the enterprise. It will be interesting to see the results of these bold moves over the next few years.

Postlaunch, organizations have been expanding the use of RWD/E to better understand how therapies perform in the real world and their impact on patient health outcomes. Most survey respondents reported that their organization had used RWD/E to understand current treatment patterns and patient behavior, such as switching patterns and adherence. For instance, Eli Lilly analyzed registry data to demonstrate higher patient adherence and longer persistence of its diabetes drug Trulicity compared to other type 2 diabetes medications.²³

Half of the survey respondents reported using RWD to understand factors impacting HCP decision-making to better position their products. Some biopharma companies are already analyzing RWD to determine the impact of their marketing outreach to patients and HCPs (see sidebar, “Analyzing RWD to determine brand marketing effectiveness,” for more information).

AI-enabled RWE generation is picking up pace

Given the expanding volume and increasing access to RWD, companies are exploring advanced analytics, including machine learning (ML), deep learning, and natural language processing, to generate RWE. The survey data suggests the application of AI is becoming more common for RWE use cases. Top use cases for AI include understanding patient behaviors, segmenting patients for trial matching, and enabling a data-driven understanding of disease progression. However, the interviewees commented that the use of AI for RWD analysis is still nascent, with biopharma companies having yet to derive true value from its use.

More concerted efforts to apply AI for RWE generation require a strong leadership commitment to AI as a long-term strategic priority. Most survey respondents stated that their organizations either have or plan to build an AI/ML workbench to enable data scientists to build, train, and deploy AI/ML models for RWD analysis. Beyond the availability of such tools, many of the interviewees stressed the need for biopharma companies to form partnerships to access high-quality RWD and build expertise in training AI/ML models (see sidebar, “Expanding access to high-quality RWD,” for more information).

Even more vital to scaling AI/ML use is broadening the internal acceptance of such techniques in the RWE space. RWE studies traditionally have involved biostatisticians, epidemiologists, and other experts using time-tested statistical approaches to validate causal relationships between data points. Opening the proverbial AI black box or at least creating an understanding of how AI arrived at a particular decision can help change perceptions around AI use. 

Call to action: Embedding RWD/E use across the enterprise

This year’s survey results suggest a continued evolution of RWD/E within the biopharma industry. There are several key principles that companies can pursue to further expand their RWE capabilities:

Set an enterprisewide strategy. Developing an enterprisewide strategy could help change how companies utilize RWD/E. For those organizations that already have a strategy in place, it should be revisited frequently, given how quickly this area is evolving.

Continue to invest. In our last survey, we saw that the companies that invested early were starting to see returns on those investments. With the pace at which technology is evolving, new sources of data, demand for talent, and continued investments also will be key to taking an organization’s RWE capability to the next level.

Integrate RWE into processes. While expanding the use of RWE across the organization, it is important to embed it in the relevant processes wherever possible, making it a standard part of the ways of working within the organization. IEP is a great example of embedding RWE into an asset’s evidence planning process.

Govern for the enterprise. This year’s survey results show organizations have defined processes for RWE study design and dissemination of results. However, embedding RWE use across the drug life cycle requires enterprise governance to ensure that stakeholders have frictionless access to data for the right purpose.

As AI-enabled RWE generation grows, it’ll become crucial to define governance to manage the ethical risks associated with AI use. Ensuring AI models are built with a thorough understanding of the disease and patient population being studied can improve the fairness of outputs and avoid biases toward certain subgroups. Algorithm validation should check training data to ensure its representativeness and correct biases that may reflect outputs. Human decision-makers should be given the discretion to override model decisions to promote ethical decision-making.

Be bold and agile. As RWE becomes a critical organizational capability, leaders should encourage teams to experiment with RWD/E to challenge orthodoxies within their companies. For instance, in R&D, leaders could encourage development teams to explore using RWD/E in drug development programs and present ideas at stage-gate meetings. Allowing teams to experiment, gain hands-on experience, learn, and iterate with RWD/E could create an environment that encourages innovation and sparks new ways of working.

As RWE becomes a key future capability, now is the time for organizations to consider steps to embed its use more extensively across the enterprise. Organizations that do so are likely to emerge as leaders and differentiate themselves as they win with RWE as an end-to-end capability.