Facebook versus The BMJ: when fact checking goes wrong