I remember seeing a study on how men admitted to hospitals with a cardiac event actually had surprisingly low LDLs on average. The conclusion? Was it that maybe their correlation assumption was off?

no, of course not! It was "We need to lower the threshold to start administration of statins."