Real-World Data in Healthcare: Hype vs Reality
TLDR
Real-world data and real-world evidence have become the new catchphrases in healthcare. Due to a combination of new evidence requirements and exponential technology innovations, several use-cases have emerged for applying real-world data in healthcare research and commercial settings for real-world. In this deep-dive, I provide a primer for the field followed by focus on two special aspects- RWE from patient generated data and RWE governance mechanisms worldwide.
Primer on Real World Data and Real World Evidence
Before we get into the details, lets first understand what’s all the fuss about.
What is Real World Data?
Any data that gives us more information about the patient’s health is real-world data. The EU GetReal initiative gets more specific by defining it as ‘an overarching term for data on the effects of health interventions (such as benefits, risks or resource use) that are not collected in the context of conventional randomised controlled trials (RCTs).’
What is Real World Evidence?
Put simply, any evidence we collect using real-world data is real-world evidence. For example, in a standard randomized controlled trial, we will go about collected specific data points for specified patient population as defined by the protocol. In a RWE trial, on the other hand, we will depend on collecting information from real-world data sources with lesser restrictions, in a more heterogenous patient population and with wider data inputs.
It is quite peculiar to note that many a time we hear the terms RWD and RWE used as if they are one and the same. As the definitions above show, real-world evidence is using data from real-world settings to generate evidence.
This infographic from JAMA 2014 quite accurately shows the varied sources of real-world health data.
Why is there so much buzz about Real-World Data/Evidence now?
This is a good question and can be answered from two perspectives that concern gathering of evidence in healthcare-shortcomings of existing methods and enablers for newer methods like RWE.
- Over the years, it has become obvious that although randomized controlled trials bring rigor and high degree of scientific validation to the hypothesis being studied, there are several drawbacks to the RCT methodology. Some of these challenges stem from the very strengths of a RCT- these include controlled study environment (not reflecting real life situations), focused patient selection (leading to evidence being applied in a narrow range of clinical practice) and multi-country/multicentric nature of modern-day RCTs (making them super-expensive).
- Evidence requirements from regulatory agencies and insurer/payors have changed over the years, going from a few specific milestones for evidence generation and review to an almost lifelong requirement for most healthcare interventions.
- Several earlier methods have been in practice which have approached shortcomings of RCTs, such as patient registries and observational studies. These methods, however, work in limited settings and come with their own challenges, both operational as well as scientific/methodological and sometimes mirror worst of both the worlds.
- Most of these challenges have been well known and so have the possible solutions. What was missing was a clear path towards the solution-opening up the RCT framework and incorporate as much data gathered from ‘normal/routine’ clinical practice as is possible. This clear path is now provided by technology adoption across various healthcare processes and organizations. One of the major change is availability of additional data sources including electronic health records (making it easy to look at large data sources), transactional/operational health data (going beyond typical health data and helping identify additional influences), and patient generated health data (data collected via sensors, wearables and other sources). The second major change comes from availability of computational speed, tools and newer analytical methods that can look at and analyze these huge, longitudinal data sources with greater accuracy and speed.
With these basics taken care of, let us now look at some of the applications of RWE.
RWE brings several benefits to various healthcare stakeholders
Real-world evidence promises something beneficial for almost every group of healthcare stakeholders-
- For organizations bringing new products and services to market (biopharma, medical devices, digital products), availability of RWD to use as evidence-generating mechanism provides several opportunities; these range from improvements in routine R&D processes such as better patient identification, additional data collection and ongoing pharmacovigilance activities to more complex applications such as identifying novel endpoints and ‘pragmatic RWE’ trials.
- For regulators, RWE provides additional sources to assess the risks and benefits of new therapies, comparing them in real-world situations with existing therapies and continue the lifecycle reviews in real-world settings. Both US FDA and EMA have advanced several RWE initiatives towards this end.
- For government and private payors, RWE brings new elements to support existing frameworks for health technology assessments (HTA), cost effectiveness analysis and pricing decisions. There are also signs that RWE analyses can help newer payment models such as value based payment and outcomes based payments.
- For healthcare providers & physicians, RWE provides a fantastic avenue towards truly personalized medicine, improving decision making processes and methods.
- Finally for patients themselves, all of these innovative methods should culminate into better healthcare outcomes, transparent payment models and hopefully improved affordability.
RWE has already started getting ‘real’
All of these benefits and improvements are not just possibilities anymore. Over the past few years, initiatives such as US FDA’s Sentinel Initiative and EMA’s collaboration with GetReal show that regulators are being more open towards use of this data for various purposes already.
By one estimate, 1 in 2 approvals by US FDA in 2018 were supported by a RWE study for safety and/or effectiveness data.
The good news is that while being competitive, the ecosystem is highly inter-connected and several partnerships have formed, all working towards the common goals of RWE applications.
- Emergence of ‘RWE firms’: These firms come in various hues and colors-platforms such as TriNetX provide access to real-world data sources; others such as Aetion and ConcertAI provide analytic platforms and methodology guidance. Traditional contract research organizations such as IQVIA and Parexel have also expanded their offerings with RWE related services to provide access to datasets and providing operational support. There are also interesting specialty focused firms such as Allstripes (formerly RDMD, focusing on rare diseases) which work with specific patient populations, to bring out customised solutions for both patient groups and industry groups.
- Hospital networks and insurer/payors: Several insurer/payors systems such as Optum and Humana have started working with their own data sources and partnering with other players in the ecosystem to analyse data and work that into their business models.
- Biopharma/Medical Device collaborations: All of the major biopharmaceutical and medical device organizations have started collaborating with these firms and starting to create RWE ‘centres of excellence’ within their own organizations to update their drug development practices. These collaborations have also produced some interesting proof-of-concepts ranging from enhanced cycle time through eligibility checks on real-world data to conducting ‘virtual’ clinical trials.
- Regulators and HTAs: I have quoted above about US FDA and EMA starting off their journeys in promoting RWE usage. Several other agencies around the world, with Japan PMDA standing out, have also started working within the ecosystem thereby creating guidelines and early collaborations for incorporating RWE into their decision making.
- Independent research and scientific collaborations: Organizations such as PCORI and Duke Margolis have taken great efforts into shaping up some of the early thoughts in this area.
Special focus I: RWE and patient generated data
One of my own focus areas within RWE landscape is the potential for patient generated data (PGD) for use as real-world evidence. As mentioned earlier, availability of mobile health technology in large sections of society and innovative sensor and wearable products have made it easy for users and/or patients to capture various direct and indirect data elements that can contribute towards real-world evidence.
There are plenty of examples of PGD from every day technology usage: pulse and heart rate captured by mobile phone watches, sleep time and diet data from patient self-reporting in a wellness app, and exercise data such as step counts from fitness wearables all count towards PGD.
These rich data sources have three main applications in RWE setting across the research-development-commercialization spectrum of pharmaceutical and medical device product lifecycles-
- Starting with research, PGD can provide avenues to develop novel endpoints that are difficult/impossible to capture through conventional methods and/or are more accurate and reflective of unmet needs. For example, gait measurement through a wearable device in Parkinson’s disease
- In the development phase, PGD can help in gathering additional data points that can help existing evidence generation mechanism. For example, capturing patient reported symptoms via mobile phone.
- In the post-approval lifecycle management, PGD can provide valuable insights into the usage of the drug/medical device in the real-world setting enabling additional safety, effectiveness and outcomes insights. For example, capturing asthma exacerbations for anti-asthmatic medication usage. They can also provide important insights from commercial perspective about drug/device adherence patterns, analysing subgroups and visualizing real-time impact.
The field of such data usage is still evolving and at each step of identification, validation, capture and review, PGD advantages need to be measured against hand-offs. Some particular methodological challenges with such data include unmeasured confounding, measurement error, missing data, model misspecification, selection bias, and fraud.
Special focus II: RWE Readiness Globally and in India
With the pressures on drug development output, increasing healthcare spending and emphasis on outcomes through product lifecycle, RWE has gathered lot of steam across the spectrum of regulators and HTAs. Several government agencies have started putting together governance mechanisms that guide real-world ‘data supply chain’ including data generation/identification, cleaning & management, aggregation and access/usage.
Major differences exist across these frameworks, chiefly concerning transparency, patient confidentiality, ethics and credibility of data usage. Various countries of course also have operational and technical difficulties such as EHR penetration and usage, interoperability standards, data quality and security considerations to name a few. The following chart, from an OECD study, showcases differences in governance and operational/technical readiness across select countries.
A recent overview of governance mechanisms for RWE of 8 countries by Office of Health Economics provides valuable inputs for an ideal national RWD infrastructure. It calls for a balance between private vs public interests as follows-
When we look at India, the picture is quite different. Due to the nature of healthcare in India with regards to drug/device regulations, pricing mechanisms and intellectual property rights, much needs to be achieved when it comes value-based/outcome-based mechanisms. Although India has a national EHR policy, EHR penetration has been patchy and still at quite an immature stage-the annotation below shows my empirical assessment of incomplete (highlighted in yellow) and missing (highlighted in red) data elements for a typical Indian patient medical record.
A direct consequence of the immature EHR landscape is low-to-nil uptake of interoperability standards. At the same time, overall data protection governance is still evolving with various rules and regulations under draft stage, thereby hampering clarity on healthcare data availability and usage.
Due to all these reasons, India would most probably be positioned in the lower left quadrant of the above graph, with both low levels of governance and tech/operational readiness.
There is some early interest from some biopharma organizations and corporate hospital systems in India towards gathering RWE data but the motivations behind these are short-term at the best. Indian RWD scene thus has a long way to maturity.
As mainstreaming of healthcare innovations accelerates, RWE will continue to become an integral part of evidence generation and review mechanisms, worldwide as well as in India. It will bring its own new challenges to the field, but it has surely started changing the status quo already!
