The Evolving Role of RWD and AI-Driven Insights In Pharma Strategy

Medical affairs professionals now have access to a vast amount of diverse data, as well as AI tools that can analyze it quickly at scale. As a result, traditional limits on what teams can achieve have shifted.

Beforehand, medical affairs and commercial teams might present a static picture of how a product has been used, based on the analysis of real-world data (RWD). Today, instead of simply delivering information, teams can provide actionable intelligence. This could include details of trends as they occur and more nuanced detail about how usage is fluctuating across demographics, geographies and time periods. With insights such as these at their fingertips, they can identify factors influencing uptake and suggest strategies to address or mitigate them.
This raises a crucial question: as our insights become more immediate, strategic and responsive, how do we ensure that they provide the right information to improve patient outcomes and address unmet needs?

Moving Beyond Traditional Real-World Evidence Approaches

Real-world data can be an extremely valuable resource for organizations that want to understand how their product is performing in the wider marketplace. For example, it can be used to identify regions where uptake differs from the mean, or how treatment persistence has varied amongst patients. It can highlight potential benefits and risks, and identify anomalies that warrant further investigation.
Traditional approaches would harness RWD drawn from various sources to produce real-world evidence (RWE). This would include clinical evidence on patient outcomes across various populations.
RWE is essential in demonstrating product safety and efficacy to regulators, and can be used to inform future clinical trials and product communications. However, it has its limitations when it comes to informing actionable strategy during the product lifecycle. This is due to two factors:

• Speed of evidence collection: Medical affairs teams draw on a range of information, from prescribing data to laboratory results, market research, claims databases and electronic health records (EHRs). This information can take months or even years to analyze, meaning that reports often appear months or years after data collection. As a result, reports tend to take more of a retrospective view, rather than presenting evidence that could inform timely decision-making.
• Lack of insight into rationale: While RWD might highlight a discrepancy in a certain region or demographic, the data alone is not sufficient to explain why this occurred. A variation in prescribing rates might be attributable to a gap in knowledge or any number of other factors such as administrative issues, lack of access or biases amongst clinicians. Without this information, analysts cannot construct effective and targeted strategies to address issues and improve outcomes.

The Role of Artificial Intelligence

Artificial intelligence (AI) tools can now analyze vast quantities of data, helping teams make new connections and discover issues that may have otherwise been overlooked. AI can be used to:

• Rapidly identify emerging trends: Potentially transformative insights may be buried in documents such as unstructured field notes from medical science liaisons (MSLs). These may take some time to analyze. AI can read and categorize notes and other data far faster, giving teams the opportunity to identify shifts and opportunities as they emerge.
• Develop predictive insights: By identifying patterns in various sources, AI can help organizations predict patient behaviors. This could include identifying socio-economic, geographical or medical risk factors that may increase the likelihood of a patient dropping out of therapy.

However, it is important to be aware of its limitations. AI will only draw insights from the information it is given. If it is primarily trained on information from certain demographics, socio-economic groups or geographies, it will demonstrate an algorithmic bias that favors those groups over others.
To develop actionable and valuable insights, it is vital to note that a patient’s experience of modern healthcare can vary significantly due to a variety of factors. As a result, any recommendations must be validated by those who have experience of the circumstances they face.

The Importance of Incorporating Clinician Insights

While RWD can provide ample information about what is taking place, it cannot provide definitive answers as to why this may be happening. Adding AI to the mix can help teams draw out nascent trends and unexpected connections faster, but may not be able to identify precisely why behaviors differ across demographics or localities.
To obtain this information, it is vital to engage with healthcare professionals (HCPs), who can deliver the context that completes the picture. Clinicians, nurses and other HCPs can share valuable insights into the day-to-day realities of care in their area and how localized trends, bottlenecks, patient behaviors and other practical realities are shaping the decisions they make.
Many of these issues may not be incorporated into clinical trials, or anticipated by algorithms. However, amplifying them is essential if organizations wish to truly understand how healthcare is evolving.

Improving Insights In Obesity Care

By combining RWD, AI capabilities, demographic data and HCP insights, teams can develop a clearer view of product performance, emerging trends and strategic interventions that can address bottlenecks and unmet needs.
This is particularly valuable in therapeutic areas such as obesity, where a number of factors can influence treatment persistence and HCP behavior. For example, a patient may respond positively to treatment initially, but may relapse due to a psychological or social trigger that may not be evident from observing raw quantitative data.
In such situations, data alone cannot tell the story. Organizations must learn about the patient experience through conversations with HCPs, and discover why patients and clinicians make the decisions they do. When this qualitative information is introduced, the resulting data produces a much clearer view of why certain situations occur, and what can be done to address them. In this way, an actionable narrative can be created for obesity care by integrating different forms of data and expertise:

Collecting Data

Teams may begin by collecting data about product usage. It should be sourced from diverse locations and demographics to ensure that it captures the broadest experience of the product. This may include MSL field insights, conversations with relevant HCPs, claims data or EHRs.

Analyzing Data

Analysis using AI and machine learning (ML) tools may uncover unexpected connections. For example, the raw data may indicate that treatment persistence for an obesity treatment is shorter in a particular region. This issue may be addressed in MSL field notes, which attribute it to a certain localized bottleneck. AI may also identify shifts in clinical prescribing patterns or patient behaviors that indicate an emerging trend, or a lingering question around the efficacy of the treatment.

Local Validation

At this point, it is important to validate this information with the help of subject matter experts (SMEs). For example, data analysis may indicate that patients in certain locations struggle with a particular obesity treatment at around the six-month mark. By discussing the data with HCPs in the area, teams can determine what is actually happening on the ground and the perceived reasons for the issue. These conversations also introduce qualitative information about why patients may be struggling with their care, or why clinicians are making certain prescribing choices. These may be due to resolvable issues with access, biases in clinical treatment or more complex psychological or social factors.

Expert-Informed Strategy

Organizations can then begin to build strategies with recommended approaches that address bottlenecks and optimize patient care. These strategies may be facilitated by AI-driven insights, but the medical strategies themselves are developed by SMEs, who deliver clinical judgments based on the evidence at hand.
At this point, organizations can confidently deliver a picture of emerging trends and issues, contextualized by HCP professionals and SMEs who can explain behaviors and offer clinical solutions.

Building the Future of Strategic, Decision-Ready Insights

This approach enables Medical Affairs and commercial teams to deliver much more actionable insights into how to maximize the product’s impact, much earlier in its lifecycle. Organizations can:

• Highlight trends and changes in behavior
• Identify issues that are impacting patient uptake and treatment persistence
• Offer tailored and clinically informed insights into how to address particular issues
• Adapt strategies and communications in response to current trends and events

Organizations are using AI and ML to analyze data on an unprecedented scale to discover more nuanced and unexpected insights about how their products are performing. However, this information is only truly valuable if it is contextualized and validated by SMEs and HCPs with local insights. By including the voice of the patient and the clinical expertise of HCPs, we can develop strategies that truly make a difference to patients everywhere.

Alexis Cogswell is vice president of real world insights at Inizio Medical.