Understanding Adverse Event Rates: Percentages and Relative Risk in Clinical Trials
Dec, 16 2025
Exposure-Adjusted Incidence Rate Calculator
Calculate Exposure-Adjusted Incidence Rate
Enter the number of adverse events and patient exposure time to calculate the Exposure-Adjusted Incidence Rate (EAIR). This tool helps you understand how safety data is reported in clinical trials.
When a new drug is tested in a clinical trial, one of the most important questions regulators, doctors, and patients ask is: how safe is it? The answer isn’t as simple as saying "15% of people had a headache." That number might sound low, but if those 15% were on the drug for only a week, while others stayed on it for two years, the real risk looks very different. This is why understanding adverse event rates - and how they’re calculated - matters more than ever.
Why Simple Percentages Mislead
For years, the go-to way to report adverse events was to count how many people experienced a side effect and divide it by the total number of people in the trial. That’s called the Incidence Rate (IR). If 15 out of 100 people got nausea, you’d say the rate is 15%. Easy. But here’s the problem: it ignores time. Imagine two groups in a trial. Group A takes the drug for 3 months. Group B takes it for 24 months. If 10 people in Group A get a rash, that’s 10%. In Group B, 20 people get a rash - still 10%. But the person in Group B was exposed to the drug eight times longer. The simple percentage hides that. It treats everyone the same, no matter how long they were on the treatment. That’s like saying two cars have the same risk of breaking down because both had one repair - even if one car drove 5,000 miles and the other drove 40,000. A 2010 analysis cited by PharmaSUG showed this method underestimates true risk by 18% to 37% in trials where treatment durations vary. That’s not a small error. It’s the difference between thinking a drug is safe when it might actually be dangerous over time.The Shift to Exposure-Adjusted Rates
In 2023, the FDA made a clear move: they requested that a biologics company submit their safety data using the Exposure-Adjusted Incidence Rate (EAIR) instead of just IR. This wasn’t a suggestion. It was a signal. The agency is no longer accepting lazy numbers. EAIR doesn’t just count people. It counts time. It asks: how many events happened per 100 patient-years of exposure? A patient-year means one person was on the drug for one full year. If 10 people were on the drug for 6 months each, that’s 5 patient-years (10 × 0.5). If 3 of them had a liver enzyme rise, the EAIR is 3 events per 5 patient-years, or 60 events per 100 patient-years. This method accounts for:- How long each person was actually on the drug
- When they started and stopped (including interruptions)
- Whether the same person had multiple events
What About Recurrent Events?
Another method, called the Event Incidence Rate (EIR), calculates events per 100 patient-years too - but it counts every single event, not just the number of people affected. So if one person has three episodes of diarrhea over a year, that’s three events. Another person has one. Total: four events over one patient-year. That’s 400 events per 100 patient-years. That sounds alarming. But is it meaningful? Not always. If those three episodes were mild and resolved quickly, they don’t carry the same weight as a single case of liver failure. EIR is useful for tracking frequency - like how often someone gets a migraine - but it can exaggerate risk if you’re trying to understand how many people are harmed. That’s why EAIR is becoming the gold standard. It can be designed to count either the number of people with an event (like IR) or the number of events (like EIR), but always adjusted for time. It gives you flexibility without misleading.
Relative Risk: Comparing Treatments Fairly
Once you have accurate rates, you can compare them. That’s where relative risk comes in. You take the EAIR for Drug A and divide it by the EAIR for Drug B. If Drug A has 45 events per 100 patient-years and Drug B has 30, the relative risk is 1.5. That means patients on Drug A are 50% more likely to experience the event per unit of time. But here’s the catch: you need confidence intervals. A relative risk of 1.5 might sound scary - until you see the 95% confidence interval runs from 0.9 to 2.3. That means the true risk could be no different at all. Statisticians use the Wald method to calculate these intervals, and tools like R’sriskratio function make it easier. Without this range, you’re just guessing.
Competing Risks: Death Can Hide Side Effects
There’s another layer most people ignore. In trials for chronic diseases - like cancer or heart failure - patients often die before they have a chance to develop a side effect. If you’re studying kidney damage, but half the patients die from their disease first, you’re not seeing the full picture. Traditional methods like the Kaplan-Meier estimator assume everyone stays at risk until the end. But if death removes someone from the pool, you’re biased. You think the side effect is rare - when really, people just didn’t live long enough to show it. A 2025 study in Frontiers in Applied Mathematics and Statistics showed that using cumulative hazard ratio estimation instead gives a 22% more accurate view when competing events (like death) happen in more than 15% of patients. This method breaks down the risk into separate parts: risk of death, risk of the adverse event, and how they interact. It’s complex, but it’s the only way to get the truth.
Industry Adoption: Slow, But Real
Not every company has caught up. A 2024 survey found that 68% of pharmaceutical firms now report EAIR alongside IR, but 42% ran into problems when submitting to regulators who didn’t know how to read it. Some reviewers thought high EAIR numbers meant the drug was unsafe - not realizing it was just accounting for longer exposure. SAS programmers say EAIR takes 3.2 times longer to code than IR. Common mistakes? Incorrect date handling (28% of cases), ignoring treatment breaks (19%), and miscalculating patient-years (23%). MSD found that switching to EAIR uncovered safety signals in 12% of their programs that IR had missed. Roche had to train their medical team because 35% of reviewers misread the results. Thankfully, resources are improving. The PhUSE group released standardized SAS macros for EAIR in 2023. They’ve been downloaded over 1,800 times. Their GitHub repo now includes 37 validation checks - like making sure no one’s exposure time exceeds the study duration, or that event counts match the data.What This Means for You
If you’re a patient reading a clinical trial summary: don’t trust percentages alone. Look for whether the data accounts for how long people were on the drug. If it doesn’t, the safety numbers are incomplete. If you’re a researcher or clinician: insist on EAIR for any trial longer than six months. Ask for the exposure time calculation method. Check if competing risks were handled. Don’t accept a simple percentage as the final word. If you’re in pharma or regulatory affairs: the FDA is watching. ICH E9(R1) already requires exposure time to be considered. CDISC now mandates EAIR reporting for serious adverse events in oncology trials. Your systems need to track start and end dates for every participant - down to the hour. And you need to train your reviewers. Because if you don’t, you’re not just submitting data. You’re submitting misinformation.The Future Is Adjusted
By 2027, experts predict 92% of Phase 3 drug submissions will include EAIR. The FDA’s 2024 draft guidance is pushing for standardization. The Sentinel Initiative is using machine learning to detect safety signals faster using exposure-adjusted data - and it’s 38% more accurate than old methods. This isn’t just a statistical upgrade. It’s a cultural shift. We’re moving from counting bodies to measuring time. From snapshots to videos. From "how many got sick?" to "how long did it take, and how often?" The truth about drug safety isn’t in a percentage. It’s in the details. And those details matter.What’s the difference between incidence rate (IR) and exposure-adjusted incidence rate (EAIR)?
Incidence Rate (IR) is the percentage of people who experienced an adverse event, regardless of how long they were on the drug. Exposure-Adjusted Incidence Rate (EAIR) calculates the number of events per 100 patient-years - meaning it accounts for how long each person was actually exposed. EAIR gives a more accurate picture of risk over time, especially when treatment durations vary between groups.
Why does the FDA prefer EAIR now?
The FDA prefers EAIR because traditional IR underestimates risk when patients are on treatment for different lengths of time. In long-term studies, IR can make a drug look safer than it is. EAIR corrects this by factoring in exposure time, giving regulators a clearer, more honest view of safety. The agency formally requested EAIR in a 2023 submission, signaling a regulatory shift.
Can EAIR miss events that happen early?
No - EAIR doesn’t ignore early events. It counts every event that occurs during the exposure period, whether it happens on day 1 or day 700. The adjustment is only to the denominator (time), not the numerator (events). So early events still contribute fully to the rate. What EAIR changes is how we interpret the rate in context - not what gets counted.
What’s a patient-year, and how is it calculated?
A patient-year is one person being exposed to a drug for one full year. To calculate it, you take the time between a patient’s first and last dose (including any interruptions), add one day, then divide by 365.25. For example, if someone was on the drug for 180 days, that’s 181/365.25 = 0.495 patient-years. This is the standard method used in FDA-compliant analyses.
Is EAIR used outside the U.S.?
Yes. While the FDA has been the most vocal in pushing EAIR adoption, the European Medicines Agency (EMA) accepts both IR and EAIR but requires justification for the method used. The International Council for Harmonisation (ICH) E9(R1) guideline, adopted globally, requires exposure time to be considered in safety analyses - even if it doesn’t mandate EAIR specifically. Most major pharma companies now use EAIR worldwide to meet U.S. standards.
How do you handle patients who die during a trial?
Death is a competing risk - it stops you from observing other events. Traditional methods like Kaplan-Meier assume everyone remains at risk, which distorts results. The preferred approach is cumulative hazard ratio estimation, which separates the risk of death from the risk of the adverse event. This gives a more accurate picture of how likely the side effect is to occur before death. This method is now recommended in recent statistical literature for trials in oncology, heart disease, and other high-mortality conditions.
Next steps for anyone working with clinical trial data: audit your exposure time calculations. Make sure your datasets include precise start and end dates for every participant. Implement validation checks for outliers. Train your team on EAIR interpretation. And if you’re reading a safety report - ask: "Was time accounted for?" If the answer isn’t clear, the data isn’t reliable.
Pawan Chaudhary
December 17, 2025 AT 15:52Linda Caldwell
December 18, 2025 AT 15:02Philippa Skiadopoulou
December 20, 2025 AT 11:39Sam Clark
December 21, 2025 AT 21:25Michael Whitaker
December 22, 2025 AT 21:43Kent Peterson
December 24, 2025 AT 00:52Josh Potter
December 24, 2025 AT 19:43amanda s
December 26, 2025 AT 04:56Peter Ronai
December 27, 2025 AT 12:18Brooks Beveridge
December 28, 2025 AT 12:50Anu radha
December 30, 2025 AT 00:44Jane Wei
December 30, 2025 AT 22:54Jigar shah
January 1, 2026 AT 21:21