Start with a clearly hypothetical example
Imagine a made-up trial in which 10 of every 100 participants in one group and 8 of every 100 in the other have a defined event over the same follow-up period. The absolute difference is 2 percentage points. Dividing that difference by the starting 10% gives a 20% relative reduction. These are two descriptions of the same hypothetical numbers.
The arithmetic does not tell us whether the event was a headache, hospitalization or death, how long follow-up lasted, or whether the result is reliable. Those are essential parts of the evidence. A percentage stripped of its outcome and setting can sound precise while communicating very little.
Then examine the published event proportions
FDA’s 2024 Wegovy announcement reported major cardiovascular events in 6.5% of the treated group and 8% of the placebo group. Subtracting those rounded proportions gives 1.5 percentage points. That calculation describes the observed difference; it is not a forecast of your personal benefit.
Our SELECT evidence review supplies the population, follow-up and treatment context. Those details matter because an event proportion accumulates over a period of observation. A trial result over several years cannot be relabeled as an annual event rate without an appropriate analysis.
A hazard ratio also uses the timing of events
The SELECT paper reports a hazard ratio of 0.80 for its primary outcome. This is a time-to-event comparison, not the ratio obtained merely by dividing the rounded percentages in a news story. Keep the authors’ measure and its label together.
If the headline says risk reduction, check whether the underlying analysis reports a risk ratio, odds ratio or hazard ratio. You do not need to reproduce the model to notice when a summary silently changes the metric. If the distinction affects a decision, ask a qualified clinician or researcher to explain the interpretation rather than treating different measures as interchangeable.
Look for uncertainty and the exact endpoint
A point estimate is one part of a study report. Look for the uncertainty interval, the planned primary endpoint and the comparator. A result for a combined endpoint does not automatically mean that each component changed by the same amount. A subgroup result also needs its own context rather than inheriting certainty from the overall trial.
Keep benefits and harms in separate, clearly labeled rows. The proportion stopping treatment, the proportion reporting a symptom and the proportion having a cardiovascular event describe different outcomes. Subtracting them from one another can create an impressive-looking number with no established clinical meaning.
Five questions before you repeat the claim
Write down the event, the two groups, the observation period, the reported measure and the population. If one is missing, follow the link to the primary source. Our news-headline checklist adds the publication and approval dates so an old result is not mistaken for a new announcement.
Finally, ask what the study did not test. A result about one medicine, formulation and population does not establish equivalent performance for another preparation or a commercial provider. This page’s paid CoreAge Rx placement is separate from the research explanation and is not supported as a clinical outcome claim by these calculations.
Common questions
Is a percentage point the same as a percent change?
No. A change from 10% to 8% is a fall of 2 percentage points and a 20% relative reduction in that hypothetical example.
Can I calculate my personal benefit from the trial’s average difference?
No. A group-level result does not determine an individual outcome. Discuss the applicable evidence and your circumstances with a clinician.
Sources and fact-checking
Sources checked 2026-09-24. Company pages document their own published terms; government and research sources provide context. This is a review of published information, not a hands-on service test or a clinical evaluation. Prices and availability can change.