Well it suggests? That study had minimal controls for correlating factors, only alcohol and red meat, nothing about careers and lifestyles that correlate to coffee consumption. Less likely to die from "accidents" is an especially suspicious measurement.