The Newly Available FAERS Public Dashboard: Implications for Health Care Professionals

MedWatch reports, after their receipt, are deposited into various databases for further analysis by the Food and Drug Administration (FDA). The FDA’s Adverse Event Reporting System (FDA AERS or FAERS) is one such database which contains reports of adverse events (AEs) associated with drugs, biologics, and certain other medicinal products.¹ The FDA received more than 1.8 million new AE reports in 2017, and the number of reports in FAERS exceeds 15.9 million.¹ Drawing meaningful and actionable conclusions regarding safety of medicinal products from FAERS data is a challenging, intricate, and ongoing process, and the FDA thus far had refrained from making the FAERS data publicly available (notwithstanding availability of these data via FDA’s quarterly data files or by making freedom of information requests).

In a landmark move, the FDA recently made the FAERS data publicly available through an online dashboard.¹ Anyone with access to Internet via a computer (or smartphone) can browse through the AEs reported for various medicinal products from 1968 till the latest quarter (June 30, 2018). Open availability of this resource has several implications for health care professionals (HCPs), as listed below:

1. Awareness: That such a resource exists and what it does is the most obvious implication; many HCPs are likely to be caught unaware of its availability and utility.

2. Dashboard functionality: Unfamiliarity with the interface functionality and navigational tools may hinder its use by HCPs. Lack of familiarity with Medical Dictionary for Regulatory Activities (MedDRA) terminology is likely to frustrate HCPs in the task of retrieving reports of specific AEs.

3. Interpretation of information: Even after becoming aware and gaining familiarity with the interface, interpretation of available safety data is likely to be daunting for HCPs. As an example, a keyword search for a product Xarelto® (rivaroxaban) via FAERS public dashboard shows that 12414 deaths have been reported after use of this product. Interpretation of such a headline statistic, without an understanding of pharmacovigilance concepts and processes, is likely to be challenging for HCPs.

Public accessibility of the FAERS data also has ramifications for treatment compliance and patient-HCP interaction. The modern Internet-savvy patient has access to several online information resources, including private- and government-funded educational websites, patient support groups, blogging platforms, and social media outlets. The newly available FAERS public dashboard is likely to join this list of online resources available to patients who may decide not to take their medications after seeing reports of serious AEs (and deaths) in the FAERS database. Patients are also likely to address consequent questions to their HCPs. To provide educated responses to patients’ questions, HCPs need to grasp the scope and limitations of such publicly available safety data.

The FAERS public dashboard can be considered as an interface which allows one to view redacted versions of previously submitted MedWatch reports related to prescription drugs, biologics, over-the-counter drugs, and homeopathic drugs (AEs involving dietary supplements and vaccines are not available for viewing via this dashboard). Some of the openly viewable data fields include the following:

• Suspect product(s)

• Concomitant product(s)

• Name(s) of the AE(s)

• Seriousness assessment of the AE (ie, serious or nonserious)

• Patient age and sex

• Country of incidence of AE

The notable unavailable data fields include the following:

• Product dosage

• Patient’s medical history

• Case narratives

On a broader level, this dashboard allows categorization of AE reports for a particular medicinal product by various criteria (eg, report source, age group, gender) and visualization of trends. For example, by performing a keyword search for Januvia® (sitagliptin) and utilizing the options available in right-hand side dropdown menu, one can observe that

• The largest number of AE reports for this product was submitted in 2007.

• A slightly higher proportion of these reports involved female patients (noteworthily, gender was not reported in a large number of these reports).

• Pancreatitis was the AE with the largest numerical count of submitted reports.

Each of the broader data categories can be subjected to further review. For example, by highlighting the year 2007 from the above example and again utilizing the right-hand side dropdown menu, one can observe that

• Increased blood glucose was the most reported AE (n = 344) in association with Januvia® in 2007.

The FAERS dashboard may also have utility for learning about safety issues which are not detailed in professional resources (eg, product package inserts). As a hypothetical example, consider a patient interested in knowing about cases of hematoma after injection of Lantus Solostar® (incidentally, the Lantus Solostar® package insert does not include hematoma as an adverse reaction²). By doing a keyword search for Lantus Solostar® and by choosing the reaction group of “General disorders and administration site conditions” and then further searching for “Injection site hematoma” in the reactions list, the HCP can see details of 9 such reports.

Similarly, consider a poison-center pharmacist interested in reports of accidental ingestion of Zoloft® (sertraline) tablets by toddlers. The FAERS output shows 94 reports of AEs associated with this product in patients aged between 2 months and 2 years. By selecting the reaction group of “Injury, poisoning and procedural complications” and by further choosing the reactions groups of “Accidental overdose” or “Accidental exposure to product” or “Accidental exposure to product by child,” the pharmacist can review details of 14 such cases.

The information retrieved from the FAERS database in the above-mentioned examples should not be used to draw any conclusions regarding safety of the medicinal product for several reasons. The most important takeaway message for HCPs is that individual reports do not imply causality of the product; these reports could had been made because of temporal association between the product and AE (eg, diagnosis of advanced-stage metastatic breast cancer in a patient 1 week after starting an over-the-counter analgesic). Underreporting of AEs is a well-recognized concern³ and the number of reports submitted to the FDA is not reflective of the true incidence of AEs associated with that product. As an example, a search for product keyword “Domeboro” brings 45 reports for viewing in the FAERS public dashboard. It is not correct to assume that only 45 patients have experienced AEs after using this product. The correct explanation is that while an unknown number of patients may have had experienced AEs after using this product, only 45 of these cases were reported to the FDA.

MedWatch reports submitted to the FDA generally contain a narrative field for describing AE characteristics. Unavailability of these narratives as well as patient medical history, product dosage, and event chronology may severely hinder interpretation of individual reports. The FAERS output provides a binary classification of seriousness of AE reports (ie, as either serious or nonserious) without providing information regarding event severity and laboratory markers which can limit generalizability across clinical scenarios. The FAERS public dashboard simply displays the seriousness of AEs as reported to the FDA. In other words, the seriousness of an AE report is assigned by the person sending the report (and not by the FDA), and two different reporters may have differing opinion of the seriousness of the same AE.

Some of the other notable limitations include the fact that the FAERS database may not contain all reports of AEs from countries other than the United States, the AE reports may be medically unconfirmed, and the FAERS data may not represent all known safety information for a medicinal product. AE reports involving products unapproved in the United States are also not visible in this database. Duplicate reports of the same adverse experience leading to larger than actual report count is another concern.

Open availability of FAERS data has important implications for patient-HCP interaction. The FAERS public dashboard, in its current form, is not conducive to retrieving actionable information and may lead to confusion and apprehension among patients. The phenomenon of probability neglect (a person’s focus on the severity of the outcome instead of the probability of the outcome⁴) can occur and, after seeing reports of serious AEs and deaths, patients may decide not to take their medications. Patients may also try to replace their medications with nonpharmacological treatments or herbal and dietary supplements. HCPs can anticipate questions from patients centered around two major themes.

The FAERS public dashboard prominently places numerical death count on all data output pages, and this headline statistic is most likely to attract patients’ attention. HCPs can therefore anticipate questions around the following theme: “Why so many people have died after taking this product?” or “Can I die after taking this product?”

Patients are also likely to compare products via the FAERS public dashboard. As an example, a patient who is trying to make a choice between the aforementioned Xarelto® and warfarin sodium will find the number of deaths reported after use of the former drug (12414) as more than double of that reported for the later drug (5506). Therefore, HCPs can expect questions such as “Why the number of deaths after taking product A is more than that for product B?” or “Is product A safer/dangerous than product B?”

To effectively respond to above questions and allay patients’ concerns, HCPs may need to provide comprehensive explanations. First and foremost, HCPs need to emphasize that the reports of AEs and death do not imply causality of the product; these reports could had been made because of temporal association between the product and AE. The large number of deaths reported for certain medicinal products may actually be due to progression of underlying disease (eg, cancer). Interpretation of individual reports and generalizability is also difficult without knowing the medical history of patients described in these reports and clinical context.

Second, HCPs need to inform the patients that their “chance” of experiencing an AE cannot be calculated from the FAERS database. The reasons behind this limitation (unavailability of actual number of patients who experienced an AE and the total number of patients exposed to the product) may need explanation. Also, the total number of reports for a product is not an indicator of comparative safety of that product.

Third, HCPs may need to present a synopsis of previously known safety data gleaned from other resources (eg, product package inserts, professional databases [such as Lexicomp®], published results of clinical trials and meta-analyses, Cochrane Reviews). The HCP can inform the patients that while AE rates and causality cannot be established from the FAERS data, these set of AEs are the most frequently reported in other resources.

Fourth, HCPs need to assure patients that various stakeholders (HCPs, pharmaceutical firms, and the FDA) are jointly and constantly involved in monitoring the safety of medicinal products. HCPs can inform patients that reports of AEs are diligently reviewed by the FDA which has robust procedures in place to take action upon recognition of any safety concerns.

Finally, HCPs need to stress to the patients to not stop taking their medication without consulting their doctor. HCPs may need to involve patients’ doctors in the consultation process to allay any lingering concerns. After the consultation, HCPs may need to follow-up with the patients to ensure their continued compliance with prescribed medications.

The newness of the FAERS public dashboard also has the obvious implication that this resource is unlikely to be taught as part of current health professional curricula or discussed in professional textbooks. Professional colleges as well as textbooks may therefore need to incorporate instruction in use of FAERS public dashboard. Such updates to curriculum may be addressed as part of initiatives such as science of safety framework.⁵ Hospitals and health care institutions may also need to conduct training sessions to familiarize their staff with this dashboard.