The U.S. Food and Drug Administration (FDA) monitors the safety of medical products once they are available to the public. To achieve this, the agency developed the Sentinel System, a national electronic monitoring tool launched in response to the Food and Drug Administration Amendments Act of 2007. The system became fully operational in February 2016. Its purpose is to protect public health by allowing the FDA to analyze real-world data and track the safety of medical products across a broad patient population after approval.
How the Sentinel System Functions
The Sentinel System operates on a distributed data network, meaning the FDA does not hold a central database of patient information. Instead, the system is composed of numerous “Data Partners,” such as health insurance plans and healthcare providers, who maintain control over their own data. These partners keep their information secured behind their own firewalls.
When the FDA needs to investigate a potential safety issue, it sends a standardized query to these partners. The partners run the query on their own data and return only anonymous, summary-level results to a coordinating center. The aggregated, de-identified results are then forwarded to the FDA for review.
This method of proactive data analysis is known as “active surveillance.” Sentinel actively scours existing data to identify potential safety signals that might not otherwise be apparent. This differs from traditional “passive surveillance” systems, like the FDA Adverse Event Reporting System (FAERS), which rely on voluntary reports from patients and providers. Passive systems can be subject to underreporting, whereas Sentinel’s active approach allows for a more systematic and timely investigation of potential risks.
Data Sources and Patient Privacy
The Sentinel System draws its information from sources like administrative and insurance claims data, electronic health records (EHRs), and clinical registries. This information is provided by the network of Data Partners, which includes national health plans and large integrated healthcare systems. The combined data covers a significant portion of the U.S. population, providing a vast and diverse dataset for analysis.
To standardize this varied information, Data Partners convert their data into the Sentinel Common Data Model (SCDM). This uniform structure ensures that queries can be run consistently across the entire network. The SCDM organizes information into standardized tables covering patient enrollment, diagnoses, procedures, and prescriptions.
Protecting patient privacy is a foundational principle of the system’s design. All personally identifiable information is removed or masked before any data is used. The FDA never receives individual patient records and only obtains aggregated, anonymous results that summarize findings from across the network. This distributed data approach minimizes the risk of privacy breaches.
The Role of Sentinel in Public Health
The Sentinel System provides data from a much broader and more diverse patient population than is possible in pre-market clinical trials. Clinical trials are conducted on a few thousand participants under controlled conditions, while Sentinel provides insights from millions of people using medical products in real-world settings. This allows the FDA to identify rare side effects or risks that may only emerge with widespread use.
The system has been applied in numerous public health scenarios. For example, it has been used to evaluate the safety of new vaccines after they are rolled out to the public, allowing for near real-time monitoring of adverse events. It is also used to investigate risks associated with widely used medications, helping the FDA determine if regulatory action, such as a label change, is needed. For instance, a protocol was developed to assess the risk of blood clots in patients with COVID-19 compared to those with influenza.