Did they make risk adjustments for any confounding factors? It seems like all they found is correlation so it doesn't necessarily follow that eating mushrooms was actually the reason for the reduced risk.