Transforming Real-World Evidence Generation in Biopharma

As the demand for Real-World Evidence (RWE) continues to swell, many organizations in the biopharmaceutical sector are ramping up their internal capabilities to meet these evolving needs. Companies are pouring significant resources into developing cloud-based advanced analytics platforms, implementing self-service cohort builders, and forming dedicated teams of RWE scientists, statisticians, and programmers. These professionals are tasked with designing and executing analyses and studies that provide critical insights into patient care and treatment outcomes.

However, despite these substantial investments, the growing demand for RWE often outstrips the internal capacity of biopharma companies. To keep pace with the rapidly changing landscape of RWE, there is an urgent need to modernize existing tools and processes.

The Limitations of Existing Technologies

In recent years, a variety of self-service tool suites have emerged within the industry, each promising to deliver two primary benefits. Firstly, these tools aim to simplify cohort definition and feasibility analyses while automating routine tasks, such as descriptive statistics for specific cohorts. Secondly, they claim to empower a wider range of stakeholders to analyze real-world data through user-friendly point-and-click interfaces, eliminating the need for complex coding skills in languages like SQL or R.

While these self-service tools have made strides in expediting cohort definition and generating commonly requested analyses, they fall short of truly democratizing RWE access. The accessible interfaces often require extensive training, leading users down a challenging learning curve. Furthermore, many solutions are offered as hosted services outside the biopharma environment, which brings forth integration challenges, complicates data control, and raises scalability issues—especially given the frequent use of per-user licensing models. As a result, the ideal of genuinely democratized RWE generation remains just out of reach.

Enter Generative AI: Changing the Way We Generate RWE

The arrival of Generative Artificial Intelligence (GenAI) has triggered a significant shift in the landscape for generating RWE, affecting business practices across various sectors, including biopharma. GenAI empowers users to engage with real-world data (RWD) in a more conversational manner, allowing them to “talk to their data.” Yet, harnessing this potential is not as straightforward as simply asking a foundational large language model (LLM) a question and receiving an instant answer.

While foundational LLMs are adept at natural language reasoning, they can occasionally produce misleading outputs, or "hallucinations." They often lack a native understanding of critical components such as data schemas, clinical code systems, and temporal logic. Furthermore, LLMs are generally not built to satisfy the specific audit, traceability, and nuanced demands of RWE. This makes the case for a purpose-built GenAI solution quite compelling.

Introducing RWE Agent

In response to these evolving needs, Deloitte has collaborated with Amazon Web Services (AWS) to develop RWE Agent—a sophisticated conversational assistant specifically designed to empower a wide array of stakeholders. With RWE Agent, users can analyze RWD, derive insights, and ultimately move closer to the vision of democratized RWE.

Recognizing the intricacies of RWD and RWE generation, the architecture of RWE Agent employs a multi-agent system. This design features specialized agents that are responsible for particular tasks, such as rules, reasoning, and analytics. When a user inputs a natural language query, a supervisory agent breaks the question down into smaller tasks and directs each component to the corresponding specialized agent. As these agents work collaboratively, they ensure that the user prompt is comprehensively understood and tackled, allowing for a fluid workflow and a high degree of accuracy in the results.

Through the innovative deployment of RWE Agent, biopharma companies can leverage the power of GenAI to streamline real-world evidence generation, ensuring that they are better equipped to meet the growing demands of today’s healthcare landscape. The era of truly democratized RWE may finally be on the horizon, driven by smart tools that bridge the gap between complex data and actionable insights.