Person-generated health data (PGHD) continues to provide valuable insights into the fight against the most recent coronavirus.

Learning about how people feel, including the extent and severity of their symptoms and whether they have been exposed to known cases or tested positive for COVID-19, can serve as a means to assist with early detection and management of COVID-19. Community-based reports can provide important information not generally available in electronic health records (EHRs) or health insurance claims databases.

To support the fight against COVID-19, IQVIA developed the COVID Active Research Experience (CARE) project, an online registry designed to learn from people about their COVID-19 experience, regardless of whether they sought medical care for COVID-19 symptoms.

Initially, the registry sought information from three groups of people (1) those who had COVID-19 symptoms, (2) those who thought they had been exposed to COVID-19, and (3) those who had tested positive for COVID-19, even if they did not have symptoms.

Once vaccines were available, the registry was then expanded to also include a fourth group -- people who were vaccinated against COVID-19. Since the project’s launch on April 2, 2020, participants have reported through their smart phones, tablets, or computers at regular intervals over a 12-month tracking period.

This scientific study is driven entirely by people reporting about their personal experiences, including any COVID-19 symptoms (presence, severity, and duration), medical history, treatments, use of dietary supplements and non-prescription medications, and other information not generally available in medical records. The CARE project supports insights into COVID-19 symptoms and severity and how they change over time. It also supports investigation into clinical treatment, and other factors related to symptom severity and resolution, as well as the side effects of COVID-19 vaccination.

To date, more than 33,000 people have joined this COVID-19 experience registry. Over the course of the study’s follow-up period, the registry seeks information on

• Demographics and occupation (gender, date of birth, geography, race/ethnicity, healthcare worker) [baseline only]
• Medical history and risk factors (hypertension, diabetes, transplant, autoimmune conditions, cancer)
• COVID-19 symptoms and severity (symptom presence and severity, impact on daily activities, work, and return to normal health)
• COVID-19 vaccine symptoms and health impact (symptom presence, impact on daily activities, impact on ability to work)
• Testing and vaccines (COVID-19 test type, result and date, COVID-19 vaccine date, manufacturer, and whether it was the first shot or a booster)
• Treatments and healthcare (medications, antimalarials, antivirals, antibiotics, biologics, steroids, healthcare encounters including type, date, and relation to COVID-19)
• Over-the-counter medications and food supplements, including vitamins and herbal products
• Vital status (death or hospitalization information through an alternate contact)

In the U.S., a trusted tokenization process links this registry data to real world data (RWD) on prescriptions, and both outpatient and inpatient healthcare utilization via medical insurance claims. This enriched data supports more in-depth examination of potential interactions with medication use, and can also be used to confirm and augment self-reported medical information in a manner that strongly protects individual privacy.

The platform is also adaptable to fit a given study purpose. For example, the platform supports validation of clinical events of special interest by direct participant contact or electronic medical record (EMR) download (U.S. only), as desired. It also supports enhanced follow-up using modest patient incentives and/or call centers.

How is this COVID-19 real word data being used

New insights
Using information directly from people without the intervention of a clinician supports a variety of insights not generally available from traditional resources. First, we were able to confirm early on that people who experienced three symptoms – fever, loss of sense of taste, and smell – were six times as likely to test positive for COVID-19 as those who didn’t have these symptoms. Research is now underway to determine if some patient risk factors that increase the risk of mortality may also be associated with having stronger reactions to COVID-19 vaccines and to COVID-19 itself.

Similarly, the role of flu shots, as well as vitamins and other dietary supplements are being investigated to evaluate whether they lessen the side effects from COVID-19 vaccine or lead to less severe COVID-19 symptoms.

Some critical considerations when using person-generated health data
Person-generated health data requires different management than traditional health research tools. There are both benefits and drawbacks to this voluntary, self-reported information.

For example, the questions must be worded so that a lay audience can understand them, including how medical symptoms are described. Participants can only report like consumers -- they are not clinicians.

Registries such as CARE, as well as any prospective study or trial, requires volunteers and are rarely formal samples drawn from a fully enumerated base population. Like all studies that do not rely on randomized treatment assignment, it is important to understand the characteristics of these volunteers and to recognize the selection factors that influence recruitment and how these might affect the generalizability of results. Also, active and adaptive data curation is necessary to cull out “bad actors” and “bad typists.”

Despite the limitations, person-generated health data offers unique insights into comparisons of various vaccines, including comparing the new class of mRNA vaccines with the traditional virally vectored vaccines, as well as comparisons of COVID-19 symptom severity for vaccinated and unvaccinated people. Subgroups of special interest, such as pregnant women for instance, can also be selectively recruited and studied.

Ultimately, several factors must be considered in determining how and when to strengthen person-generated health data. For example, if a research question is intended to examine transmission rates, it may be important to collect information on the genomics of any infection. For pregnancy studies, it is important to evaluate birth outcomes. For any type of study driven from such data, modest participant incentives will enhance retention.

One of most enduring lessons from this work is the enormous value collectively obtained from listening to people, and how it is possible to couple this kind of data with systematic evidence generation to better understand the benefit and the risks of treatments and vaccines. Researchers need more than one approach to get answers. The idea of waiting until the ideal clinical trial or set of trials have been completed, typically requiring many years before actionable evidence surfaces, is not good enough. COVID-19 has ushered in a new appreciation and use cases for person-generated real world data and its potential will extend far beyond the pandemic.

For more information, the potential of patient-generated was discussed in the recent webinar,
Real World Evidence for COVID-19: Making an impact now and into the future.