Hacker News new | ask | show | jobs
by recurseP 215 days ago
This study is just another case of confusing correlation with causation, wrapped in a scary headline to grab attention.

The problem here is that they compared people who were already sick enough to need long-term melatonin prescriptions with those who weren’t. That’s not testing melatonin’s effects, it’s just showing that people with serious health problems (like chronic insomnia, depression, or anxiety) tend to have worse outcomes. And surprise, those same conditions are already known to increase heart risks.

Here’s the kicker: in the US, melatonin is over-the-counter. So their "non-melatonin" group probably included plenty of people using it anyway (they just didn’t have a prescription on record).

No info on doses, no explanation of how it might actually cause heart issues, and it’s not even peer-reviewed, it's just a conference abstract. Even the AHA expert they quoted sounds pretty skeptical (but of course, the press release still makes it sound like melatonin is the villain).

Honestly, if you wanted to design a study that would produce misleading results, you’d do exactly this: use observational data, ignore selection bias, and skip adjusting for how severe people’s conditions were. The real takeaway is that people with chronic insomnia have worse health. Groundbreaking stuff (not), applause.