Like a tsunami, our capacity for rapid data accumulation and data interpretation is advancing exponentially. Machine learning (ML), Natural Language Processing (NLP), and the evolution of electronic health records are revolutionizing the potential availability and use of Real World Data (RWD) sources to improve health. Such data sources are also helping to broaden and define the patient cohort in clinical trials; potentially illuminating the holy grail of responders and non-responders. The question is: can we navigate the divide in the context of Real World Evidence data and that of real-time clinical trials data? This is not an easy path but one that is worth the effort so we can correlate the two whenever possible and better define patient cohorts who will respond to a therapeutic regimen.

Simultaneously, traditional Randomized Controlled Trial (RCT) evidence may be declining as smaller patient populations, related to more personalized medicine (the right therapeutic at the right dosage at the right time for the right patient), make it harder to design studies, and the per-patient and set up costs of conducting RCTs are rising (related to increases in complexity and external standards). Interest in pragmatic clinical trials (PCTs) that combine randomization with more real world circumstances has grown with the potential use of routine data sources to record patient events and outcomes, transforming the costs, size and feasibility of such trials. This changing environment is creating new opportunities for the use of Real World Evidence (RWE).

COVID  — 19 has also accelerated the use of RWD and RWE as clinical trials have rapidly moved from site-based to home-based. Fundamentally, the use of smart sensors and telemedicine for home-based trials has changed the way clinical trials will be managed going forward. Organizations are leveraging aggregated and anonymized EHR data and medical chart data, patient registries, and mobile device information, including smart wearable sensors, to better understand trends and outcomes, leading to improved care delivery.

Real World Evidence and Real World Data

According to the FDA, Real World Data (RWD) refers to all data relating to patients’ health status. This could include Electronic Health Records from hospital treatment, but it can also include data from a wide variety of other sources, including:

• Aggregated and anonymized EMRs/EHRs
• The UK Biobank*
• Healthcare billing records
• Health insurance payer claims records
• Product registries
• Disease & device registries
• Patient-generated healthcare data
• Prospective Observational Data
• Pharmacy Data
• Data from mobile devices
• Mortality Data
• Lab/Biomarkers Data
• Consumer Data
• Survey Data
• Social Media Data

_{Real World Data Sources & Real World Evidence}

Any and all data relating to the health of real world patients can be considered RWD.

Real World Evidence (RWE) is the clinical evidence that’s generated from the statistical analysis of RWD. The FDA considers RWE acceptable clinical evidence if the relevant data meets FDA standards for data fitness, which means that in some cases pharmaceutical companies can use existing data to get regulatory approval for new drugs and find new use cases without the need for expensive clinical trials.

During the current pandemic, this information has become increasingly valuable. Real world data can help inform leaders about anything from high-risk patient populations to the impact of measures like social distancing. It is an invaluable resource, but like any data-related strategy in healthcare, it comes with several hurdles.

For starters, like a tsunami the breadth of this data can easily overwhelm entities without the tools or capacity to make sense of it. And even if they do have what they need to draw comprehensive conclusions and correlations from this information, quickly executing an efficient response is another task entirely.

Real World Evidence has taken a foothold in the clinical trial environment with current use:

1. Drug development: RWE is used to identify targets for the development of new therapies and design the drug development pathway.
2. Regulatory approval decisions: Use of RWE in initial FDA regulatory decisions has been limited to date to circumstances where an RCT is not practical, but the FDA has recently (August 2017) released guidance on the use of RWE to support regulatory decision-making for medical devices and is required to issue RWE guidance for drugs under the 21st Century Cures Act.
3. FDA safety monitoring and safety signals: FDA use of RWE to monitor post approval drug safety is much more established, most recently via the Sentinel Initiative.
4. Health Technology Assessments (HTA) assessments and payer coverage decisions — initial decisions: Payers use epidemiological data, based in part on claims data, at this stage of decision making to generate estimates of the potential patient population they will cover that may require the treatment, and to estimate potential costs and cost offsets.
5. HTA assessments and payer coverage decisions — reassessments: RWE gives decision makers the opportunity to reconsider coverage, discounts and formulary tiering in light of how the products are performing in their relevant patient population.
6. Outcomes-based contracting: Patient outcomes are tracked in order to support contract agreements tying level of reimbursement to real world clinical performance. To date, outcomes-based agreements have not featured prominently in the US health care system because of the difficulty of collecting data to support such agreements, but interest appears to be growing.

While there is great promise in using RWD and FEW, there are some major challenges that must be overcome: potential bias of data sources, incomplete data sets and lack of harmonization of data between RWE data sources, access to such data, and lack of standards in assessing the value of RWE data. Improving RWE data is key for it to gain international traction.

• Evaluation of drug effectiveness, safety, and adherence in real world patients
• Using RWE to evaluate the durability of benefits and side effects over a longer period than studied in RCTs
• Exploring sub-population groups in which clinical benefit is (likely to be) greatest
• Gaining comparable evidence on the new drug and on the comparator (“usual care”)
• Evaluating benefits when used outside of the initial indication
• Leveraging the advantages of pragmatic clinical trials to inform all aspects of the evaluation of drug effectiveness and safety
• Evaluation of comparative effectiveness through indirect comparisons (network metadata analysis) enriched with outcomes from real world patients
• Evaluation of outcomes that are not measured during the standard development process, for example any “other benefits” or wider elements of value such as the impact on productivity
• Evaluation of budget impact and cost-effectiveness in a real world setting.

We can expect Real World Data and Real World Evidence to continue in support of deriving informed efficacy of a drug in a clinical trial. This is a learning experience as we try to correlate RWE with data from actual clinical trial participants. The reality is that patient recruitment and retention in any clinical trial is a real and on-going challenge that cannot be minimized. As we develop standards of use of Real World Evidence, the objective is to broaden the clinical trial patient population as well as its diversity so we can more accurately predict which patient cohort will be a responder or non-responders. Accuracy depends on having a statistically relevant population. Real World Evidence helps to achieve this statistical relevance. The acceptance of Real World Evidence is a process that continues as we more fully understand its contributions and the challenges of mediating hype versus real application. The bottom line, therefore, is that Real World Data/Real World Evidence is here to stay because it does contribute meaningful value in augmenting patient recruitment challenges. It does not replace recruited patients, but we must remember that its intrinsic value is that is derived from real patients. The trick is understanding and mediating the context variances between RWE and real-time data from recruited patients in a clinical trial.

Some have also gone further in the use of synthetic patient populations, but they represent a challenge even though most are derived and extrapolated from real patient populations.

Quantori

Quantori is an experienced systems integration and data science company that can help fill in each of the gaps and disambiguate the data sources.

For hospitals and healthcare providers, we can offer unrivaled systems integration and data management experience. We know the healthcare industry and its unique challenges (such as strict regulatory requirements and difficult legacy systems). We’ve built a wide variety of custom software solutions for data systems integration at healthcare firms, so we know what it takes to get your data prepared and integrated into the RWD/RWE pipeline, regardless of the systemic and regulatory challenges you’re facing.

For pharmaceutical firms, we offer our deep systems integration experience in tandem with a data science team that has unparalleled experience in the field of biostatistics and bioinformatics. Our team members have built advanced analytics teams for pharmaceutical companies, delivered the data science strategy for major research organizations, and conducted a wide variety of RWD/RWE analyses using a variety of languages and machine learning techniques.