[flaptrap]

>reduce the risk of developing cardiometabolic multimorbidity in individuals without any cardiometabolic disease,

Get funding for a study, run multiple regression against a large set of interviewees, find an outcome which satisfies the proponent, spin the results: people who drink coffee and do not have heart disease or diabetes get fewer heart disease or diabetes later, ergo, caffeine reduces the risk. People, this is not a double-blind study. It was biased from the start.