The 90% figure from that list is described by the study which listed it as from an unreliable/untested research methodology with a small sample size (18 cases).

The 90% figure from that list is described by the study which listed it as from an unreliable/untested research methodology with a small sample size (18 cases). It looks more like ~20% is a more common upper bound for the figures listed.