Lessons Learned from COVID-19 Clinical Trials

The challenges of the COVID-19 pandemic did bring about some positive impact, accelerating the application of real-world data (RWD) to support evidence generation and public health programs as researchers tried to gain a better understanding of the rapidly evolving virus. HealthVerity recently hosted a webinar that discussed how the technology that made this possible is also being applied to clinical trials and other research to provide a more comprehensive view of the patient journey and advance the future of healthcare. If you missed the webinar, here’s what it covered:

The Technology Progressing Healthcare 

Incorporating RWD into clinical research requires linking the records for the same individual across disparate data sources and over time, while ensuring both accuracy and privacy. This is done with Privacy-Preserving Record Linkage (PPRL). Using our PPRL technology, HealthVerity Census^®, the records containing personally identifiable information (PII) remain with the data owner and are de-identified and assigned a universal patient identifier, known as an HVID, in lieu of PII.

Accuracy is critical to the integrity of the research; however, the inherent inconsistencies in RWD, with nicknames, misspellings and missing fields, can lead to false positives, where different individuals are inadvertently linked as one person, causing inaccurate correlations, or false negatives, where the same person is incorrectly counted as separate individuals, leading to a fragmented view of the patient journey.

HealthVerity Census uses advanced identity resolution techniques to attain the highest accuracy (0.2% false positives and 3-5% false negatives). By using a patient-centric approach where identities are matched to a central repository, rather than dataset to dataset, once an identity is accurately resolved and assigned an HVID, it can be seamlessly linked with any data source in the HealthVerity Marketplace, the nation’s largest healthcare and consumer data ecosystem, representing over 75 unique sources and 150 billion de-identified transactions.

This technology and data ecosystem were used to make the following revelations during the COVID-19 pandemic and are now being applied by life sciences organizations to advance healthcare:

Lesson 1: Leveraging RWD for Real-Time Evidence

The National Cancer Institute (NCI), a division of the National Institutes of Health (NIH), began a study in early 2020 to better understand the risk of COVID re-infection. By combining a variety of data sources, HealthVerity was able to assemble a pilot study cohort of more than 3.2 million people with SARs-CoV-2 antibody tests that NCI could longitudinally follow to compare the infection rate for antibody positive versus negative tests. The pilot was then extended for long-term follow up, with other inputs added, such as vaccine status, for a final study cohort of approximately 25 million patients.

Several life sciences organizations seeking answers to questions about COVID were also able to leverage similar datasets:

• Janssen sought to understand the durability of their vaccine before and during the Delta variant surge
• Pfizer evaluated COVID vaccine breakthrough infections among immunocompromised individuals
• Gilead compared the mortality rate for patients treated with remdesivir versus matched reference patients found in RWD

Accurately linking patients across RWD by leveraging our identity resolution technology allowed these organizations to build the longitudinal patient journeys they needed to answer their questions.

Lesson 2: Resolving Identity to Link Data Across Place and Time

The Centers for Disease Control and Prevention (CDC) wanted to better understand vaccine administration rates and the degree of immunization among the U.S. population, as well as determine areas that could potentially benefit from targeted communications. To do this, they needed to be able to accurately link records across time, with weeks between doses, and often with different locations for each vaccine administration. For example, an individual received her first vaccination in her hometown, then weeks later, went back to work in a neighboring state and got her second dose near the office. She then got a booster months later at another pharmacy back in her home state. The HealthVerity PPRL technology enables resolving this to a single de-identified fully vaccinated and boosted individual versus three partially vaccinated people.  

Just as the CDC has multiple jurisdictions capturing vaccination data that needed to be linked, clinical trials might have multiple locations and other sources feeding data into the trial. This same approach can be used to accurately link those disparate data sources, which can then be linked to a variety of RWD for further analysis.

Lesson 3: Linking Primary and RWD

NCI wanted to understand the impact of the COVID pandemic on cancer diagnosis and treatment. To do this, NCI worked with HealthVerity to link their robust cancer registry data with closed medical claims. NCI analyzed patients diagnosed and treated in 2019 and compared outcomes to patients diagnosed and treated in 2020 to determine if there were changes in care or behavior.

In another study, the CDC wanted to link SARS-CoV-2 genomic sequencing data they were gathering with lab partners to RWD to better understand demographics, clinical characteristics and outcomes associated with different variants. Similarly to the first lesson, HealthVerity was able to assemble a cohort of COVID patients meeting the criteria for the study from a variety of RWD sources and use HealthVerity Census to concurrently de-identify the CDC’s primary SARS COV-2 genomic sequencing data and enable linkage to RWD, delivering an interoperable dataset to address their questions.

On the commercial side, a life sciences client is studying a disease where mobility and the patient experience have an impact on outcomes. Using HealthVerity Census we are able to de-identify and link a partner’s data from wearables for a view of mobility and the client’s patient reported outcomes (PRO) data about their daily living experience. These sources can then be linked to the vast HealthVerity RWD ecosystem to build a robust, interconnected dataset that gives a comprehensive view of the disease’s impact from the patient perspective, the mobility perspective and the clinical perspective.  

The Final Lesson

What we learned from the pandemic is that interoperable RWD, when accurately linked and de-identified, is a powerful tool that can help answer many questions we never thought we could answer before. Click here to see a demo of how you can access the HealthVerity RWD ecosystem or click here to watch a recording of the webinar, Applying Public Health Lessons to Life Sciences, Clinical Trials and Registries.