When a new drug hits the market, the work isn’t over. In fact, that’s when the real monitoring begins. The U.S. Food and Drug Administration (FDA) doesn’t just approve drugs and walk away. It keeps watching - closely, constantly, and with increasingly powerful tools - to catch problems that clinical trials simply couldn’t see. Why? Because real people use these drugs in real life, with other medications, chronic conditions, and unpredictable genetics. What looks safe in a trial of 3,000 people might cause harm in a population of 3 million. That’s where the FDA’s postmarket safety system steps in.
How the FDA Finds Hidden Risks
The FDA doesn’t rely on luck or guesswork. It uses a mix of passive and active systems to hunt down safety signals. The backbone is the FDA Adverse Event Reporting System (FAERS), a database holding over 30 million reports since 1969. Doctors, pharmacists, patients, and drug companies send in reports when something bad happens after someone takes a medicine. These aren’t proof the drug caused the problem - just clues. A sudden spike in reports of liver damage after taking a new diabetes drug? That’s a signal. The FDA doesn’t just count reports. It uses statistical tools like Empirical Bayes Screening and Proportional Reporting Ratios to separate noise from real danger. But FAERS has a big flaw: underreporting. Studies show only 1% to 10% of actual adverse events make it into the system. Many patients don’t know how to report. Many doctors don’t have time. That’s why the FDA built something better: the Sentinel Initiative. Launched in 2008 with $120 million in funding, Sentinel taps into electronic health records and insurance claims from over 300 million Americans. Instead of waiting for reports, it actively scans for patterns - like a sudden rise in heart attacks among people prescribed a new cholesterol drug. It’s like having a real-time dashboard on how millions of people are responding to medications.What Happens When a Problem Is Found?
Finding a signal is only step one. The FDA’s Office of Surveillance and Epidemiology (OSE) brings together doctors, pharmacologists, statisticians, and data scientists to evaluate it. They look at the strength of the signal, the number of cases, the severity of outcomes, and whether other agencies saw the same thing. If the evidence points to a real risk, the FDA doesn’t hesitate. It can update the drug’s label to warn doctors and patients. It can require a Risk Evaluation and Mitigation Strategy (REMS) - a set of strict rules like mandatory patient education, restricted distribution, or special monitoring. As of early 2024, 78 drugs had active REMS programs, affecting about 20 million patients. In extreme cases, the FDA can pull a drug off the market. That’s rare - less than 0.1% of approved drugs - but it happens. Take fenfluramine, once used for weight loss. It was linked to heart valve damage in the late 1990s. After thousands of reports and analysis, the FDA ordered its withdrawal. That’s the system working as intended: catching harm before it becomes a national crisis.
Who Reports, and Why So Few Patients Do
You might think patients are the main source of safety reports. They’re not. According to a 2023 study, healthcare providers filed 63% of reports to FAERS. Drug manufacturers made up 31%. Patients? Just 6%. Why? The process isn’t easy. On MedWatch, the FDA’s reporting portal, it takes about 17 minutes to file a report - but most patients don’t know it exists. A 2022 survey by the National Organization for Rare Disorders found that 72% of rare disease patients didn’t know how to report an adverse reaction. Even doctors, who see the most cases, often skip reporting because it’s not part of their workflow. One oncologist on Reddit said she’d only filed three reports in five years - despite seeing many serious side effects. The FDA knows this. That’s why it runs the MedSun program, recruiting 350 healthcare providers specifically to report device-related injuries. But awareness is low. A 2022 FDA survey found 41% of providers didn’t even know MedSun existed. Fixing this gap isn’t just about better tools - it’s about culture. Patients need to know their reports matter. Providers need to see reporting as part of care, not extra paperwork.How Technology Is Changing the Game
The FDA isn’t stuck in the past. In 2019, it rolled out InfoViP, a decision-support tool using natural language processing and machine learning to scan FAERS reports faster and smarter. By late 2023, InfoViP 3.0 cut the average time to detect a safety signal from 14 months down to 6.2 months. That’s a 56% improvement. Machine learning also reduced false alarms by 19%, helping teams focus on real threats. The next big leap is Sentinel 2.0, launched in February 2024. It now includes genomic data from 10 million people through partnerships with biobanks. Imagine knowing that a certain genetic variant makes someone 10 times more likely to have a bad reaction to a drug. That’s the future. By 2025, the FDA plans to integrate data from the NIH’s All of Us program, which tracks health information from 1 million diverse Americans. This will help uncover safety issues in underrepresented groups - something clinical trials have long struggled with.
The Big Gaps and the Road Ahead
Despite all this, the system isn’t perfect. A 2021 GAO report found the FDA failed to require postmarketing studies for 37% of high-risk drugs approved between 2013 and 2017. Even when studies are required, they’re often late. Dr. Joseph Ross at Yale found that only 58% of mandated studies were completed on time, with delays averaging over three years. That means patients could be taking drugs with unknown long-term risks for years. Small biotech companies are especially vulnerable. While 92% of large drugmakers use AI tools to detect safety signals, only 37% of small firms do. The cost and complexity are high. But the stakes are even higher. Cancer drugs, which make up just 14% of approved medications, account for 22% of safety studies - because they’re risky. Gene therapies and complex biologics are the next frontier. Their side effects may take years to show up. The FDA’s system is built for pills, not gene edits. Funding is another concern. The Office of Surveillance and Epidemiology is operating at 82% staffing capacity. Without more resources, the system risks being overwhelmed. The global pharmacovigilance market is projected to hit $12.3 billion by 2028, but most of that money goes to drug companies, not regulators. The FDA’s budget hasn’t kept pace.What This Means for You
If you take prescription drugs, this system is your invisible safety net. It’s why your doctor checks your liver enzymes before prescribing a new medication. It’s why some drugs come with black box warnings. It’s why you’re asked to report side effects - even if they seem minor. Your report could be the first clue that leads to a life-saving change. If you’re a caregiver, patient advocate, or healthcare worker, know that your voice matters. Don’t assume someone else will report. Use MedWatch. Tell your doctor. Push for better systems. The FDA doesn’t have eyes everywhere - but millions of patients do. The future of drug safety isn’t just about better tech. It’s about better participation. The goal isn’t perfection. It’s progress. To catch harm faster. To protect more people. To make sure that every new medicine - no matter how promising - is truly safe for everyone who takes it.How does the FDA know if a drug is unsafe after it’s approved?
The FDA uses multiple systems to track drug safety after approval. The main tools are the FDA Adverse Event Reporting System (FAERS), which collects over 30 million reports from doctors, patients, and drug companies, and the Sentinel Initiative, which actively scans electronic health records and insurance claims from over 300 million Americans. These systems look for unusual patterns - like a sudden increase in liver damage or heart problems - that weren’t seen during clinical trials. Statistical tools and machine learning help separate real risks from random events.
Why don’t more patients report side effects?
Many patients don’t know how or where to report. The process can feel disconnected from daily life. A 2022 survey found 72% of rare disease patients didn’t know how to report adverse events. Even doctors, who see side effects regularly, often skip reporting because it’s time-consuming and not part of their usual workflow. Only 6% of reports to FAERS come from patients, while 63% come from healthcare providers.
What is a REMS program?
A Risk Evaluation and Mitigation Strategy (REMS) is a set of safety measures the FDA requires for certain high-risk drugs. These can include mandatory patient education, special training for prescribers, restricted distribution, or regular blood tests. As of January 2024, 78 drugs had active REMS programs, affecting about 20 million patients. REMS are used when the known risks are serious enough to warrant extra controls, even if the drug is otherwise effective.
Can the FDA remove a drug from the market?
Yes, but it’s rare. The FDA can withdraw a drug if the risks clearly outweigh the benefits and no other safety measures can reduce the harm. This usually happens after multiple reports, confirmed by active surveillance, and after the manufacturer has had a chance to respond. Examples include fenfluramine (linked to heart valve damage) and rofecoxib (Vioxx, linked to heart attacks). Most drugs are kept on the market with updated warnings or restrictions instead.
How is the FDA improving its drug safety system?
The FDA is investing in new technology. Sentinel 2.0, launched in 2024, now includes genomic data from 10 million people. InfoViP 3.0, updated in late 2023, uses advanced machine learning to cut signal detection time from 14 months to under 6.5 months. By 2025, the FDA plans to integrate data from the NIH’s All of Us program, which tracks diverse populations. The goal is to move from passive reporting to real-time, predictive safety monitoring.