Flagged up

The meteoric rise of real-world evidence presents new opportunities and new challenges.

The healthcare industry is experiencing a paradigm shift as real-world evidence (RWE) derived from real-world data (RWD) continues to revolutionise the world of drug development and commercialisation worldwide.

Many practitioners point to the opportunities of RWE to rectify certain limitations associated with randomised clinical trials (RCTs). These include representativeness of the underlying patient population, lack of head-to-head comparators, short follow-up and the ethical concerns of randomising patients.

Regulators and payers, including the FDA, EMA and NICE, increasingly seek high-quality, fit-for-purpose evidence from routine clinical care to support decision-making.

As RWD becomes more ubiquitous, RWE practitioners must ask the difficult question – which data? An innocent and deceptively simple question, its answer is complex.  How do we decide whether data is fit for purpose? How do we effectively identify and evaluate potential data sources? What characteristics are most important?

Much of the best-known RWD around the world has a long history of use in drug development and commercialisation.

Whether a dataset is fit-for-purpose is implicitly context-dependent; there are certain elements of RWD that typically characterise high-quality data including accuracy, completeness, consistency, timeliness, validity, relevancy, integrity, accessibility and traceability.

In general, great general-purpose RWD ideally also has long patient follow-up, a holistic view of the patient experience and linkability between disparate data.
To be effective in this brave new world, RWE practitioners should have a general knowledge of high quality and commonly used RWD to plan and interpret research. Many know that Nordic RWD is world-class, but few can articulate the central reasons why and truly understand its possibilities.

Nordic countries include Denmark, Finland, Iceland, Norway and Sweden with a total population of about 27.5 million. The first registries emerged in the 17th century on paper, administered by the church, collecting information about births, deaths and marriages.

Much later, the Nordic social welfare model began to develop in the 20th century, in which each state developed a broad, governmentally managed social welfare system including universal healthcare.

To administer such a system, governments needed to collect extensive amounts of health and other data on all of their citizens. Strong societal trust in institutions has enabled the collection of this data in a way that would have been controversial in other geographies.

At the same time, the personal identification number (social security number) was introduced in each country, which today enables the identification and tracking of each person’s interaction with various social services.

Furthermore, Nordic countries were early digitisers of health data, which contributed to the availability of today’s data. Each of these factors played a key role in the structure and qualities of Nordic data available for health research today.

Nordic data for health research can be broadly differentiated into administrative, claims-type data with complete population coverage and clinical data including disease registries, electronic health records (EHRs) and biobanks with deep clinical content but often partial population coverage.

Across the Nordics, population administrative data contains healthcare visit data including contact dates, ICD-based diagnosis codes, procedures and diagnostic-related group (DRG) costs, type of specialist and more.

Pharmacy-dispensed medications are available through anatomic therapeutic chemical (ATC) codes, brand names, prescriber specialty, date of prescription, and dispensation and costs. The statistics and social insurance agencies maintain information about income, education attainment and work loss.

Additional clinical data can be found in roughly 200 disease or medication registries across the Nordics, which often contain measures such as disease severity and quality of life. These registries can be deterministically linked at the patient level using personal identification numbers to the administrative data sets.

Registries differ from administrative data in their coverage of the underlying patient population, which is variable. Electronic health records and biobanks are also linkable data sources, including information such as patient-reported outcomes, lab tests and genetic information.

Beyond the administrative and clinical registry data, practitioners can make use of population surveys, environmental data and researcher-initiated cohorts.

To answer certain questions relevant for healthcare decision-makers, data on patients is not enough. Luckily, these sources of patient health data can be combined with total population registers containing all residents in each country, to create control groups without the disease of interest.

In addition, linkages can be made to family members in order to identify genetic risks or to understand caregiver burden.

Many of the beneficial characteristics of Nordic data can be found in other datasets around the world but the power of Nordic data lies in their combination, which is globally unique.

Nordic data has many applications throughout product development in pharma and medtech. There are several particularly compelling ways it can be used across regulatory and commercial use cases.

Assessing patient safety and effectiveness is a key need for regulators and therefore for pharmaceutical companies.

An effective safety study includes RWD from routine clinical care to capture real patient populations, long follow-up to find late-onset safety events, large populations for sufficient power especially for rare events, complete populations for unselected results and robust capture of safety events.

Nordic data provides excellent conditions for these types of studies. Although the Nordics are smaller in population than many other regions, entire patient populations can be extracted from all 27.5 million residents rather than subsamples, many of which are not fully representative.

In parallel, to assess comparative effectiveness, a great deal of progress has been made in causal inference methodology. However, in many cases, the method is only half of the story – the right data is needed to enable these methods, notably to control for confounding.

The Nordic data environment is also unique in its ability to drive forward the promises of personalised medicine. In best case scenarios, the data used to enable personalised medicine research contains many different dimensions including medical, environmental and genetic information to tailor treatment and predict patient risk, amongst other applications.

This is a challenging set of conditions that are simply unrealistic in many regions due to current infrastructure, linkage and access requirements.

In the Nordics, the future is already here. For an example, in Finland, patients’ entire electronic medical record and socio-economic information is available for approximately 15 years in combination with complete genome variant data sequenced using Genome-Wide Association Studies (GWAS) for more than 500,000 individuals in FinnGen.

This data and sets like it, exist within a robust data management and extraction infrastructure to enable research leading to truly unique insights that are simply not otherwise possible on such a large and impactful scale.

This expansive data enables a wide range of applications in personalised medicine from both a subject matter perspective but also methodologically – powerful AI models need diverse, high quality, high-volume data – all of which are available together in the Nordics.

Despite the polarising debate of ‘RWE vs RCT’, there is a particularly important area of overlap – the registry-based RCT (rRCT). The world’s most well-known, large scale rRCT is the TASTE trial conducted in Sweden in the early 2010s and published in the NEJM, bridging the Swedish clinical registry infrastructure and follow-up with the RCT methodology.

This methodology addresses a recurrent challenge of RCTs – the non-representative patient population – by integrating recruitment, treatment and follow-up within routine clinical care, as well as the challenge of establishing causality in RWE. Although not without their own challenges, rRCTs are a powerful approach that can be implemented using the Nordic registry environment.

Due partly to the infrastructure already in place, this pragmatic trial enrolled 7,244 patients at an incremental cost of $50 per patient.

Nordic RWD also excels in rare disease, due to its ability to identify and richly characterise patients using the region’s diverse and linkable data using the entire population as a basis.

Similarly, in chronic disease, long follow-up and holistic data are paramount in understanding patient trajectories over time, medical and economic outcomes and develop robust knowledge about therapeutics over time.

Despite the multitude of powerful use cases for Nordic RWD, there are scenarios when practitioners should look to local data. In particular, research questions that are dependent on the structure of healthcare systems or formularies are unlikely to generalise well from the Nordics to other regions.

Today, Nordic data is often used by pharmaceutical companies in commercial applications where robust evidence is needed to establish unmet need, understand holistic disease burden, analyse treatment patterns and mapping patient trajectories.

Although Nordic data is vastly under-utilised given its potential, its acceptance has already been demonstrated in international regulatory and payer decision-making including by the FDA, EMA and NICE. As the need for all types of medical evidence grows, expect to see Nordic RWD used more often and in more innovative ways.

The nature of Nordic RWD, shaped in part by each country’s historical societal evolution, sets a benchmark in the global RWE landscape.

The deep data quality, long patient follow-up and linkable data infrastructure has positioned Nordic RWD as a globally unique region for generating powerful RWE.
To enable robust evidence generation strategies, RWE stakeholders in the pharmaceutical industry must have a firm grasp on the Nordic data landscape and its opportunities as it provides high-quality solutions to many RWE needs.

As requirements on RWD grow, so will interest in Nordic data, which should be included in all modern global RWE strategies.

Kirk Geale is CEO at Quantify Research. Go to quantifyresearch.com