Research Methods
Read the dose before you repeat the hazard ratio
A methods note on two semaglutide trials: 2.4 mg and 1 mg were different assignments, so their hazard ratios are not substitutes.
When you extract a result, write four items before the relative risk: who was enrolled, what they were given, how long they were followed, and which event counted. SELECT fails that test if the note says only “semaglutide reduced cardiovascular events.” The assignment was 2.4 mg once weekly. FLOW fails the same test if the note borrows SELECT’s dose. FLOW assigned 1.0 mg once weekly.
A practical check
Open the methods, not the news headline. Confirm the dose in the randomization sentence. Then confirm that the percentage you want to quote uses that dose and names its estimand or analysis set. If the paper reports more than one analysis, do not pick the larger one because it is larger.
Discontinuation is a method, too
SELECT reports that 16.6 percent of the semaglutide group and 8.2 percent of the placebo group stopped for adverse events. A methods reading asks whether the quoted analysis still counts those people. If you cannot answer from the paper, you do not yet have the result.