If they did the result would be a journal rendered almost unreadable by the large percentage of studies it describes with no significant outcome, and would remain unread.

This depends on how it was organized. Data sets could be maintained, and only checked when papers show interesting results in nearby areas.