Keeping It Real

Despite being championed by a number of statistical consultants for many years now, real-world data (RWD) tends to be overlooked by global product sponsors. Many in the pharmaceutical industry simply prefer to stick with what they know – the traditional clinical trial.

Clinical trials have always been a vital part of the drug development process, delivering key information about the therapeutic effect and safety of a new drug candidate. Nevertheless, they have their limitations when it comes to gauging the real-life effectiveness of a therapy. For instance, inclusion criteria are typically very narrow, often excluding patients under concomitant treatments, organ dysfunctions, or comorbidities – or those who are outside the age limit to reduce confounding factors and ensure data is applicable to an “average” patient. In reality, however, there is no average – and the patients receiving the newly approved therapy may be receiving treatment for other conditions. Clearly, leaving these people out of trials doesn’t help our understanding of how a therapy will perform in the real-world. Patient adherence is another well-known real-world factor for which clinical trials fail to account; participants tend to be more compliant with instructions during trials than patients in the real-world, who may take their treatment at different times of the day – or even miss doses accidentally. Finally (and crucially), clinical trials can’t factor in differences in healthcare professional (HCP) and patient perception of what constitutes a meaningful impact on symptoms and quality of life.

For treatments that tackle rare diseases or target demographics with ethical concerns – children, older or pregnant people, for example – there are even more challenges; in particular, it can be hard to find enough participants for a clinical trial to ensure adequate representation at all.

In my view, RWD has the potential to fill some of the gaps left by clinical trials – and pharmaceutical companies and regulatory authorities are finally taking note of the possibilities. In 2019, Pfizer’s Ibrance was approved for a supplemental New Drug Application by the FDA using analysis largely based on RWD – an industry-first that demonstrates the value. RWD has also been used throughout the COVID-19 pandemic to provide insight on the success of vaccines.

But where can RWD come from? Well, a variety of sources, including electronic health record data and insurance claims, notes from interactions between doctors and patients, and even fitness data stored on patients’ smart devices. In particular, the de-identified medical records of millions of patients stored in observational longitudinal databases can be harnessed to enhance our understanding of the true effects of treatments and care. The largest of these databases, Truven Health Analytics, features information on 100 million patients dating back 40 years – it’s a treasure trove of data that is far richer than anything a clinical trial could provide on its own.

The long-term information on offer with RWD means we could, for example, track the treatment of rare diseases, changes in pattern of treatment, and concomitant therapies. When used in combination with clinical trial data, RWD can give us a detailed and more complete understanding of how treatments are performing.

But there is still a long way to go before RWD truly comes of age. The data itself – what is collected, where/how it is stored, and how it is used – isn’t standardized globally (or even across different databases in the same region). And different databases have their own merits and drawbacks, making the harmonization process more challenging. For instance, administrative healthcare databases are clean and consistent to ensure straightforward sharing of information with insurance companies for billing; they are easy to study, but restricted in the amount of insight they offer. Electronic medical/health records (EMR/EHR) databases, on the other hand, contain patient medical records from multiple facilities in a single directory. Yes, they can be disorganized and incomplete, but they do offer in-depth insight into how patient attributes – such as lung volume or pain scores – affect treatment response.

Whatever RWD we rely upon (and, at some point in the future, that should be all!), this rich, long-term evidence – when used (analyzed) appropriately – offers unmissable insight into treatment efficacy in the real world.