It's the best test there is for scientific truth or validity. Anyone deliberately dodging it to the extent Wells does is almost definitely doing a snow job.
There have been studies/articles/theories/books which have been peer reviewed by a community of experts in the relevant field which are/turn out to be invalid.
Of course. The process isn't perfect. If peer review can still admit errors, then any sweeping statement or study that hasn't been peer reviewed is almost sure to contain a huge number of errors, if not outright falsehoods.
There have been studies/articles/theories/books which have not been peer reviewed by a community of experts in the relevant field which are/turn out to be valid
And had such studies been peer reviewed, they would surely have met muster. If not, well, then they're probably lacking in accuracy.
Peer review is the best tool we have for discerning accuracy in the technically complex realm science has become, but obviously, you seem unsatisfied with it. With what mechanism would you suggest it be replaced?