I should probably have used the word efficiency instead of effectiveness because I had in mind the ratio benefits/(costs+risk).

from https://www.sciencedirect.com/science/article/pii/S227458072...

  Our review of the trial publications found high quality evidence of statistically significant group differences, but also recognized that there are mixed views on the clinical relevance of the observed differences and the value of therapy for individual patients. ...
  If patients are treated, then confirmation of AD by positron emission tomography or cerebrospinal fluid analysis and monitoring for risk of amyloid-related imaging abnormalities was recommended, as done in the clinical trials, although it would strain Canadian resource capacity.