Research reading group · call recap
The cardiovascular trial, and relative against absolute
13 April 202518 attendedHeld by Sunil
SELECT — Lincoff AM, et al. New England Journal of Medicine 2023;389(24):2221 to 2232. Our biggest reading group to date, largely because this is the paper members are quoted at most often. The headline everybody had met was a twenty per cent reduction. Most of the hour went on what that number is a reduction in, and why the absolute figure is the one that tells you something about a person.
What the paper is
A cardiovascular outcomes trial in people with established cardiovascular disease and overweight or obesity, and without diabetes. Primary endpoint a composite of cardiovascular death, non-fatal myocardial infarction and non-fatal stroke. Semaglutide against placebo, on top of usual care, over a period of years.
The room did the population properly, which is where reading pays off. These were people who already had cardiovascular disease. That is not most of the people who take these medications and it is not most of the people quoting the result.
Relative and absolute
This took half an hour and was worth it. A relative reduction of roughly a fifth in event rates is not the same statement as a fifth of the people in the trial avoiding an event. The event rates in both arms were single-figure percentages over the trial, so the absolute difference between them is a few percentage points, and that is the number that answers how much this changes for one person like the trial population.
Nobody was made to feel stupid, which in a room with a statistician and two clinicians in it is the actual achievement. Ruthie said it was the first time in two years that a percentage had stopped feeling like a slogan.
What the room concluded
That the paper is important and the way it gets used is often not. It supports a cardiovascular benefit in a defined high-risk population. It does not support the general claim that these medications protect anybody's heart, and several members had been told exactly that by people who meant well. Two took away a specific question about whether their own risk profile resembles the trial's at all.