Decoding Medical Statistics: Why Relative Risk Reduction Overstates Treatment Benefits

The moment a patient and a physician decide whether a new medication is worth its side effects happens when they look at the trial's efficacy number. That single percentage dictates the perceived value of the intervention. But if that number is framed as a relative risk reduction rather than an absolute one, the perceived benefit is mathematically distorted from the start.[9]

Consider a hypothetical clinical trial for a new heart disease medication, a scenario frequently used to illustrate statistical framing. If two out of 100 people in a control group suffer a heart attack over five years, the baseline risk is 2%. If the drug lowers that rate so that only one person out of 100 suffers a heart attack, the new risk is 1%.[7]

How that 1% difference is communicated changes everything about how the drug is received. The absolute risk reduction—the actual difference between the two groups—is exactly 1 percentage point. But the relative risk reduction, which compares the 1% drop to the original 2% baseline, is 50%.[7][8]

Pharmaceutical marketing and medical headlines almost universally favor the 50% figure. A 50% reduction sounds like a massive breakthrough, while a 1% reduction sounds like a rounding error. Both numbers describe the exact same clinical outcome, but they drive entirely different patient decisions.[6][9]

The Cochrane Handbook for Systematic Reviews of Interventions, the gold standard for evaluating medical evidence, explicitly warns against relying solely on relative measures. Chapter 15 of the 2023 handbook states that relative risk "can overestimate the clinical importance of an intervention" if the baseline risk is low.[1]

To translate these percentages into practical clinical decisions, researchers use a metric called the Number Needed to Treat (NNT). The NNT is calculated by dividing 100 by the absolute risk reduction. In the heart attack example with a 1% absolute reduction, the NNT is 100.[7][8]

The NNT tells a physician exactly how many people must take the drug for one person to experience the benefit. If the NNT is 100, it means 99 people will take the medication, absorb its financial costs, and risk its side effects without deriving any cardiovascular benefit from it.[7]

This mathematical reality is why absolute risk is the only metric that allows a patient to weigh a drug's true value against its harms. If a medication causes severe nausea in 5% of patients, a 50% relative risk reduction sounds like a fair trade. But if the absolute risk reduction is only 1%, the drug will harm five people for every one person it helps.[4][9]

The American Academy of Family Physicians (AAFP) has long advocated for transparent risk communication. In a foundational year 2000 publication, the AAFP emphasized that physicians must present absolute numbers to patients, noting that "patients are more likely to choose a treatment when its benefits are described in relative terms rather than absolute terms."[4]

The distortion becomes even more pronounced in rare diseases. If a condition affects two people in 10,000, and a drug reduces that to one in 10,000, the relative risk reduction is still 50%. But the absolute risk reduction is 0.01%, meaning 10,000 people would need to take the drug to prevent a single case.[8][9]

Conversely, relative risk can sometimes obscure the value of an intervention when the baseline risk is extremely high. A 2019 review in the Journal of the American Heart Association points out that for high-risk populations, even a modest 10% relative risk reduction can translate to a massive absolute benefit, saving thousands of lives.[3]

Public health agencies face similar challenges when measuring disparities across populations. A Centers for Disease Control and Prevention (CDC) methodology report highlights that measuring health disparities using only relative differences can lead to contradictory conclusions about whether a gap is widening or narrowing, depending entirely on the mathematical baseline used.[5]

The Oxford Academic journal Nephrology Dialysis Transplantation published a 2017 supplement arguing that "one cannot be interpreted without the other." The authors stress that relative risk provides the strength of the biological effect, while absolute risk provides the actual clinical relevance to the patient sitting in the exam room.[2]

For patients reading medical news, the defense against statistical distortion is a simple question: What was the baseline risk? If an article claims a new diet doubles the risk of a specific cancer, doubling a 0.1% lifetime risk to 0.2% is a vastly different threat than doubling a 10% risk to 20%.[8][9]

The National Center for Biotechnology Information's Smart Health Choices guide advises consumers to actively look for the absolute numbers in any health claim. If the absolute numbers are missing, the claim is incomplete by definition.[8]

Medical communication is slowly shifting toward greater transparency, with major journals increasingly requiring authors to publish absolute risk reductions and NNTs alongside their relative findings. Until that standard is universal, the responsibility for doing the math remains in the exam room.[9]