Understanding Adverse Event Rates: Percentages and Relative Risk in Clinical Trials

When a new drug is tested in a clinical trial, the goal isn’t just to see if it works - it’s to understand how safe it really is. One of the most important ways we measure safety is by tracking adverse events: unwanted side effects, reactions, or complications that happen during treatment. But how do you decide if one drug is safer than another? A simple percentage - like "15% of patients had nausea" - might sound clear. But it can be wildly misleading. That’s where understanding percentages and relative risk becomes critical - especially when the FDA is asking for more accurate methods.

Why Simple Percentages Fail

The most common way to report adverse events is the Incidence Rate (IR). It’s simple: divide the number of people who had an event by the total number of people treated. If 30 out of 200 patients got a rash, that’s a 15% rate. Easy. But here’s the problem: what if one group took the drug for 3 months and another took it for 2 years? The IR doesn’t care. It treats both groups the same, even though the longer-exposed group had far more opportunity for side effects to appear.

This isn’t just a technical glitch - it’s a safety blind spot. A 2010 analysis found that using IR alone could underestimate true event rates by 18% to 37% in trials where treatment durations varied. Imagine a cancer drug that causes fatigue. If patients on Drug A stay on treatment for 18 months and those on Drug B stop after 3 months, the IR might show Drug B as safer. But that’s not because Drug B is safer - it’s because patients just didn’t have enough time to develop the side effect. That’s not safety. That’s timing.

Enter Patient-Years: The EIR Method

To fix this, researchers started using Event Incidence Rate adjusted by Patient-Years (EIR). Instead of counting people, you count time. One patient taking a drug for 1 year = 1 patient-year. Ten patients taking it for 6 months each = 5 patient-years. You then calculate how many events happened per 100 patient-years.

For example: if 25 patients had headaches over 500 total patient-years of exposure, the EIR is 5 events per 100 patient-years. This gives you a rate that reflects how often side effects occur over time, not just how many people experienced them. It’s especially useful for chronic conditions where patients stay on treatment for years - like rheumatoid arthritis or hypertension.

But EIR has its own flaw. If one patient has five separate headaches during treatment, EIR counts each one. That inflates the risk. You’re not measuring how many people are affected - you’re measuring how many times something happened. That’s fine for rare, serious events like liver failure. But for common, mild events like dizziness, it can make a drug look riskier than it really is.

The FDA’s Push for EAIR

In 2023, the FDA made a clear move: they requested Exposure-Adjusted Incidence Rate (EAIR) in a supplemental biologics license application. This wasn’t a suggestion. It was a signal.

EAIR is more complex. It doesn’t just account for time - it also accounts for recurrence and treatment interruptions. If a patient stops the drug for two weeks due to a side effect, then restarts, EAIR adjusts for that gap. If a patient has three episodes of vomiting over 12 months, EAIR captures that pattern, not just the count.

The result? A more accurate picture of real-world safety. MSD’s safety team found that switching to EAIR revealed previously hidden safety signals in 12% of their drug programs - especially for long-term therapies. Roche, on the other hand, found that 35% of medical reviewers initially misread EAIR results because they weren’t trained on it. That’s the catch: better methods need better training.

Relative Risk and the Real Safety Story

Now, let’s say Drug A has an EAIR of 8 events per 100 patient-years, and Drug B has 4. The relative risk - or Incidence Rate Ratio (IRR) - is 2.0. That means patients on Drug A are twice as likely to experience the event per unit of time. But here’s what most people miss: relative risk doesn’t tell you the absolute risk.

If Drug A causes 8 events per 100 patient-years, that’s still only 0.08 events per person per year. Even with a relative risk of 2, the actual increase is small. That’s why you need both numbers: the EAIR to understand the baseline, and the IRR to compare treatments.

The FDA and EMA now expect both. The ICH E9(R1) guidelines, updated in 2020, require safety analyses to consider exposure time and treatment discontinuation. That means if you’re submitting data to regulators, you can’t just give percentages anymore. You need to show how risk changes over time - and how it compares between arms.

Implementation Challenges

Switching from IR to EAIR isn’t just a math problem - it’s a systems problem.

Pharmaceutical statisticians report that EAIR takes 3.2 times longer to program than IR. A typical EAIR analysis might take 14.7 hours versus 4.5 hours for IR. Common mistakes? Incorrect date handling (28% of cases), ignoring treatment breaks (19%), and inconsistent patient-year calculations (23%).

CDISC, the global standard for clinical data, now requires EAIR reporting for serious adverse events in oncology trials (v3.0, 2023). SAS and R have standardized macros for EAIR - the PhUSE GitHub repository for these tools has been downloaded over 1,800 times since 2023. Still, 42% of companies report formatting issues when submitting EAIR to regulators who aren’t familiar with the metric.

What’s Next?

The industry is moving fast. In 2020, only 12% of FDA submissions included exposure-adjusted metrics. By 2023, that jumped to 47%. The global clinical trial safety software market hit $1.84 billion in 2023, growing at 22.7% yearly - all because regulators are demanding better data.

The FDA’s 2024 draft guidance on exposure-adjusted analysis proposes standardized EAIR formulas. The European Network of Pharmaceutical Medicine is testing EAIR across 15 disease areas - early results show it outperforms IR in 11 of them. And the FDA’s Sentinel Initiative is now using machine learning to detect safety signals from EAIR data - with 38% better accuracy than old methods.

By 2027, experts predict 92% of Phase 3 drug submissions will include EAIR alongside traditional IR. The days of reporting simple percentages as safety metrics are ending. The new standard is time, recurrence, and context.

What You Need to Know

If you’re reading clinical trial data - whether you’re a doctor, a patient, or a researcher - here’s what matters:

• Never trust a percentage without knowing the exposure time.
• Look for EIR or EAIR - not just IR - in safety tables.
• Check if the study reports relative risk (IRR) alongside absolute rates.
• If treatment durations vary between groups, ask: "Was exposure time accounted for?"
• Be wary of claims that one drug is "much safer" - small absolute differences can look big in relative terms.

The goal isn’t to scare people away from new treatments. It’s to make sure we understand the real trade-offs. A drug might have a higher relative risk - but if the absolute risk is tiny, it might still be worth it. That’s the nuance. That’s the science. And that’s what the FDA is now demanding.