ClinicalStudyConnect.com

Plain-Language Clinical Study and Ingredient Research

Absolute Versus Relative Risk in a Trial: Working Through the Same Numbers

posted on

Absolute Versus Relative Risk in a Trial: The Short Answer

Relative risk tells you how many times as likely an event was in one trial group compared with the other. Absolute risk tells you how many people actually had the event. The same result can be a big relative change and a small absolute change, so you need the baseline risk and the time period before judging it.

By ClinicalStudyConnect.com Research Desk. Updated September 26, 2026.

Educational information only. This article explains how to read trial numbers. It is not medical advice and does not say whether any treatment is right for anyone. Do not start, stop, or change any medicine because of a study result. Talk with a licensed clinician or pharmacist about your own situation. In an emergency, call your local emergency number.

Terms to Know Before You Read the Numbers

  • Risk: the chance that an outcome happens, shown as a percentage or a fraction. The Cochrane Handbook describes risk as the probability that a health outcome will occur.
  • Baseline risk: the risk in the comparison group, the people who did not get the treatment being tested. It shows what usually happens without the intervention.
  • Time period: how long people were followed. A risk of 4% over one year is not the same as 4% over ten years.
  • Relative risk (risk ratio): the risk in one group divided by the risk in the other. A value of 0.75 means the risk in the treated group was three-quarters of the comparison group’s risk.
  • Absolute risk difference: the risk in one group minus the risk in the other. It shows how many more or fewer people had the event.
  • Number needed to treat (NNT): how many people would need to receive the tested treatment instead of the comparison for one additional person to avoid (or have) the event in a given time frame. The Cochrane Handbook ties this to a time frame.

Why Baseline Risk Changes What a Relative Risk Means

A relative risk hides how common the event was to begin with. The Cochrane Handbook says the clinical importance of a risk ratio cannot be judged without knowing the typical risk of events without the intervention. Its own example is a risk ratio of 0.75, which could mean a drop from 80% to 60% or a drop from 4% to 3%.

Same relative number, very different real-world size. That is the core reason to look for both.

Working Through the Same Numbers Two Ways

These are practice numbers for arithmetic only, not results from a real trial. Assume both groups were followed for the same 12 months. Both examples use a relative risk of 0.75.

Example A: A Rare Event

  • Comparison group: 40 of 1,000 people had the event (4%).
  • Treated group: 30 of 1,000 people had the event (3%).
  • Relative risk: 30 ÷ 40 = 0.75. This is often described as a 25% relative reduction.
  • Absolute difference: 40 − 30 = 10 fewer events per 1,000 people, or 1 percentage point.
  • NNT: 1 ÷ 0.01 = 100 people treated for 12 months for one fewer event.

Example B: A Common Event

  • Comparison group: 800 of 1,000 people had the event (80%).
  • Treated group: 600 of 1,000 people had the event (60%).
  • Relative risk: 600 ÷ 800 = 0.75, the same 25% relative reduction.
  • Absolute difference: 200 fewer events per 1,000 people, or 20 percentage points.
  • NNT: 1 ÷ 0.20 = 5 people treated for 12 months for one fewer event.

Both headlines could read “25% lower risk.” One is 10 people per 1,000 and the other is 200 per 1,000.

What Trial Reporting Guidelines Ask For

The CONSORT 2025 reporting guideline for randomized trials, item 26, asks authors to report the result for each group and the estimated effect size with its precision, such as a 95% confidence interval. For binary outcomes (the event happened or did not), it asks for both the absolute and the relative effect size. Item 14 asks authors to define the time point for each outcome. The guideline home page is at consort-spirit.org.

The Cochrane Handbook adds two points about presentation. It says natural frequencies, such as “per 1,000 people,” are best understood by consumers. It also says that when people have different levels of risk, absolute benefit should be expressed for each identifiable risk group, with the time period made clear. It notes that relative effects tend to be more stable across risk groups than absolute effects, which is one reason absolute numbers need to be given for each risk group separately.

Risk Translation Worksheet: A Step-by-Step Field Guide

Use this when you read a trial report, abstract, or news story. Fill in each line from the paper, not from memory.

  1. Find the outcome and the time period. Write down exactly what event was counted and how long people were followed.
  2. Find the baseline risk. Look in the results for the comparison group’s count or percentage. Write it as a number out of 1,000.
  3. Find the treated group’s risk. Write it the same way, out of 1,000.
  4. Subtract. Comparison minus treated is the absolute difference, in events per 1,000 people.
  5. Divide. Treated risk divided by comparison risk is the relative risk. Check that it matches the paper.
  6. Check precision. Note the confidence interval. A wide interval means the true size is less certain.
  7. Check who was studied. Ask whether the trial participants’ baseline risk looks like the risk of the person you have in mind.
  8. Check harms. Repeat steps 2 to 6 for any side effects the trial reports.

Uncertainty and What This Method Cannot Tell You

This worksheet translates numbers. It does not tell you whether a result is trustworthy. A trial can be small, short, or poorly designed, and a neat calculation cannot fix that. A worksheet also uses group results, so it cannot say what will happen to any one person. Because absolute effects depend on baseline risk, a result from a high-risk group should not be assumed to apply to someone with a much lower baseline risk.

Conditional Next Step: If the Paper Gives Only a Relative Number

  • If the results table lists events per group, use steps 2 to 5 above to calculate the absolute difference yourself.
  • If only a relative figure appears and no group counts are given, you cannot work out the absolute size. Treat that as a reporting gap, since CONSORT asks for both, and look for the trial’s registry results or full text.
  • If the numbers relate to your own health decision, bring the worksheet to a clinician or pharmacist and ask what your own baseline risk is over the same time period. Many trials post results on ClinicalTrials.gov, which has a guide on how to read study results.

Sources Used

Clinical Study Connect is an independent educational publication. It does not sell or recommend products, and this article is general education, not a substitute for advice from a licensed clinician.