The BBC did a great job of covering this story; most of the media did their usual miserable health care work. It's no wonder the BBC's news sites are displacing traditional print media.
Researchers from the Dutch national Institute for Public Health and the Environment in Bilthoven, Netherlands, asked over 17,000 how much coffee they drank each day.I bet this 50% reduction effect doesn't hold up. Seven cups is a lot -- that's getting into my range. Maybe they actually showed a continuous decline in incidence with coffee dose, but if it were a discontinuous effect then I'd be extremely skeptical.
Those who drank seven or more cups of coffee a day-were 50% less likely to develop type-2 diabetes compared with those who drank two cups a day or less.
The association was still seen when factors such as smoking, alcohol consumption and body mass were taken into account...
... This study backs up previous research which showed that when people increased their coffee consumption for 14 days, their blood glucose levels were reduced, but substituting regular coffee for decaffeinated coffee for 20 days did not affect glucose levels.
They claim they controlled for body mass, but all the mega-coffee drinkers I can think of are relatively slender types. I suspect what they did was isolate a group of adults with ADD traits who manage that predisposition by using a legal and safe stimulant -- coffee. This group is notoriously fidgety and restless, and hence prone to a lower average lifelong body mass. Even if they controlled for present body mass, I'd wonder if they missed a difference earlier in life.
That doesn't mean there isn't some substance of interest in the coffee bean related to insulin activity, but I bet the effect is much less than a 50% reduction in onset of DM II.