In 2007, several high-profile drug safety problems triggered Congress to mandate the FDA to develop an active surveillance system, utilizing electronic health data, to improve tracking and analysis of drug safety after drugs reach the market [1]. During the drug approval process, clinical tests are conducted to show drug safety and efficacy for treating a specific indication and patient population. The limited size and scope of clinical trials leads to side effects and safety problems that were not detected in the clinical trials to emerge once the drug is on the market, prescribed to a large number of patients, and prescribed for off-label indications [1]. Prior to the release of the Food and Drug Administration Amendments Act (FDAAA), the FDA relied on passive surveillance activities to monitor adverse events and safety issues. Manufacturers submit mandatory adverse event reports, and voluntary reports are collected from user facilities, physicians, and consumers. The FDA also uses published studies and post-market studies conducted by pharmaceutical manufacturers as additional sources for drug safety information. However, the passive surveillance process of reporting adverse events through the FDA’s Adverse Event Reporting System (FAERS) and MedWatch has several weaknesses. First of all, physicians and consumers only voluntarily report adverse events. It is estimated that only about 10% of adverse events are actually reported to FAERS [1]. Despite the low reporting rate, the FDA receives too many adverse event reports to validate. In 2013, just over 1 million adverse event reports were submitted [1]. Consequently, it is not until the FDA observes a strong indication that a problem has occurred that the related reports are thoroughly evaluated. Furthermore, many reports are incomplete and do not allow the reported event to be fully assessed.
With increased use of Electronic Health Records (EHRs), big-data analytics, and coordination and alignment among stakeholders in the healthcare industry, the FDA was able to respond to the FDAAA mandate by launching the Sentinel Initiative in 2008 [1]. Through the Sentinel Initiative, the FDA began working with the public and private sectors to develop the Sentinel System to allow it to use existing electronic healthcare data in near-real time to more quickly evaluate and understand safety issues. The FDA awarded 10 contracts to address issues of data infrastructure, privacy, methodological issues, and stakeholder engagement [1]. Harvard Pilgrim Health Care, Inc. was awarded a contract to create a miniature, working version of the future full-scale Sentinel System [2]. The pilot project is referred to as the Mini-Sentinel and it concluded in 2014.
Over the course of the Mini-Sentinel, over 20 collaborating institutions acted as data partners and Harvard Pilgrim established a Mini-Sentinel Coordinating Center (MSCC) to test scientific methods and develop new activities to create the Sentinel System [2]. Instead of requiring all participating institutions to submit their data to a central computer repository, Sentinel System functions as a distributed data network. Thus, all data analysis is performed by each data partner behind its own secure firewalls. The data partners are required to transform their data into a standardized format, which allows the MSCC to write a single analytical software program for a given safety question and supply it to each data partner to run on its own standardized data (The Sentinel Initiative). To evaluate post-market safety issues, the process of active surveillance utilizing the Sentinel System is as follows: when the FDA identifies a potential link between a drug and an adverse event or safety risk, it submits a query to the Coordinating Center (CC) [1]. The CC develops the software program the data partners will use to conduct an analysis within their own systems. The CC sends the program to the data partners to evaluate the safety question on the standardized data in their local, secure database. Each data partner then returns HIPPA-compliant summaries of the results to the CC [2]. The CC compiles all of the summary results, evaluates their validity, and ensures the program ran correctly. Finally, the CC gives the FDA the compiled results and the summary results from each data partner.
Since the completion of Mini-Sentinel, the FDA has announced the formal launch of the full-scale Sentinel System and granted $150 million to Harvard Pilgrim to continue developing the system over the next five years [1]. Despite taking this next step in active surveillance, the FDA still faces unresolved issues regarding the Sentinel System. Under Mini-Sentinel, 137 drugs were assessed as a result of queries submitted by the FDA to the MSCC. However, of those 137 assessments, only 18 have undergone further investigation by the FDA. This indicates that the system is not being used to its full potential. In addition to technical and methodological problems that still plague the system, conflicting findings from different databases and studies have led to confusion and decelerated FDA action [1].
The Sentinel System is far from completion and requires further refinement to address the complex technological, methodological, legal, and operational challenges associated with this program [2]. The FDA is continuing its efforts to overcome these challenges. For example, in April 2015, it unveiled a $1 million grant to the Reagan-Udall Foundation to transform large amounts of electronic health record data into usable information for evaluating post-market drug safety [3]. While the Sentinel System does not directly affect pharmaceutical manufacturers, its results will indirectly affect their post-market activities, such as increasing the additional safety studies the FDA asks them to complete on an approved drug. Therefore, pharmaceutical manufacturers should stay up to date on the progress made toward realizing the full-scale Sentinel System, so they can be prepared to respond to the new methods for monitoring drug safety.
References
- Health Policy Brief: The FDA’s Sentinel Initiative. Health Affairs. 2015. Available from: http://www.healthaffairs.org/healthpolicybriefs/brief.php?brief_id=139.
- The Sentinel Initiative. FDA. 2010. Available from: http://www.fda.gov/downloads/Safety/FDAsSentinelInitiative/UCM233360.pdf.
- Freund, L. FDA Unveils $1 Million Postmarket Surveillance Grant. FDA News. 2015. Available from: http://www.fdanews.com/articles/170869-fda-unveils-1-million-postmarket-surveillance-grant.
