These findings suggest practical implications for optimising resource allocation ... yielding a final dataset of 2 512 445 documents. Titles and abstracts were processed to extract time-related data ...