The State of Drug Safety Surveillance in the U.S.: Much Improved, More to Come

When a new drug is approved in the United States, it is virtually impossible to know all of the risks that a population may encounter when using that product. Even though the U.S. Food and Drug Administration (FDA) requires drug manufacturers to meet rigorous standards demonstrating the drug’s safety and effectiveness for its intended use, once approved, drugs can be used by many more patients than were studied in clinical trials. This may include patients with unique clinical conditions, differing health status, ethnicity, age, or other characteristics which were not well-represented before the drug’s approval. Further, the drugs themselves can be used in different ways and in different settings than were studied. Until recently, FDA did not have the necessary tools and data access to rapidly and consistently track the risks of serious side effects of regulated drugs after approval. Recognizing this challenge, FDA has developed a pilot system to make the best use of available electronic health data using a new data and research network capable of evaluating the safety of medical products in the U.S. 

Authorized by the Food and Drug Administration Amendments Act (FDAAA) of 2007, this pilot is known as Mini-Sentinel, and is part of FDA’s larger Sentinel Initiative. Sentinel was envisioned as a national electronic system to track the safety of regulated medical products, through the use of existing health insurance claims and electronic clinical data that are generated as part of routine care. In the four years since its inception, Mini-Sentinel has made tremendous progress toward developing this system. Mini-Sentinel is comprised of insurance claims and clinical data from 18 participating data-partners, including some of the largest private health plans in the United States. In order to best protect patient privacy, the data from each partner is maintained behind each individual health plan fire-wall. This “distributed data” approach allows a single coordinating center to distribute FDA safety questions in the form of “queries,” to each of the participating data partners to be run against their own data. Aggregated summary results are then sent back to the coordinating center for final analysis. This process allows FDA to access data that can help in addressing safety questions in near real-time.  

Through Mini-Sentinel, FDA has the capability to better understand the safety outcomes using electronic health care data of approximately 169 million covered lives. This accumulation of data represents the capture of 382 million person-years of observation time and billions of prescription dispensings.[1] Examples of the types of safety questions that have already been addressed by Mini-Sentinel include the following:

• Safety concerns with drugs used to treat high blood pressure and the incidence of angioedema;
• Safety concerns with a new diabetes treatment and the incidence of heart attacks; and
• Impact of FDA regulatory actions (i.e. drug label changes) intended to mitigate serious risks of drugs.

The Mini-Sentinel pilot has demonstrated substantial progress and has proven to be a very useful tool for FDA, largely due to the strong partnerships developed between FDA, collaborating academic institutions, and private health plans. However, in order to ensure continued progress and long-term sustainability, it will be critical for progress to continue in several key areas. 

First, continued methods development and data understanding will be necessary to ensure FDA has access to the most innovative tools. The field of pharmacoepidemiology and drug safety surveillance is still young and the continued development of better study designs and analytic tools to quantify risks of serious adverse events, while accounting for many confounding factors that are inherent on observational data, will be critical. Further, as health reforms impact that way health care is delivered and financed (e.g., development of accountable care organizations and increased use of bundled payments), the electronic health data will change. It will be important to focus efforts on understanding how these changes will impact data used for safety evaluations. 

Second, it is clear that Sentinel’s contributions may extend well beyond FDA’s medical product assessments. The tools and infrastructure that have been developed by FDA over the last four years could be used as a platform to establish a national resource for a more evidence-based learning health care system. This system will enable a better understanding of not only the risks, but also benefits and best uses, of drugs in the post-market settings. 

FDA has initiated steps to ensure the long-term sustainability and impact of Sentinel infrastructure and tools. Within the next few years, FDA has proposed that Sentinel be transitioned into three main components: the Sentinel Operations Center, the Nation Resource Data Infrastructure, and the Methodological Resource for Medical Product Surveillance using Electronic Healthcare Databases. FDA has indicated that while the Sentinel Operations Center will continue to serve as FDA’s portal to the distributed database, the Nation Resource Data Infrastructure could potentially be used by other groups to support broader evidence generation. Potential groups with interest in improving our understanding of the impact of medical products and who could benefit from this framework include the National Institutes of Health, the Regan-Udall Foundation, the Patient Centered Outcomes Research Institute, and other possible stakeholder groups, such as the private industry. 

Collectively, these components will ensure that FDA continues to have the tools to engage in medical product surveillance, while ensuring the long-term sustainability of the system. In just four years, the Sentinel Initiative has laid the groundwork to transform how FDA, and the nation, benefits from electronic health care data. This network continues to foster a community of stakeholders committed the evidence generation, which will ultimately contribute to a learning health care system.

For more information on these issues, including discussion by leaders from Sentinel stakeholders, please visit the Sentinel Initiative Public Workshop event page. There you will find archived video, presentations, and further reading.

[1] http://mini-sentinel.org/about_us/MSDD_At-a-Glance.aspx