I'll say the mildly concerning things are the 78% recommending brain fitness programs despite any evidence that they do anything besides cost money, 12% recommending a very expensive amyloid PET scan despite it not being ready for routine clinical care with as of yet poor predictive accuracy.
More concerning are the numbers suggesting that the MMSE and MoCA are used for diagnostic purposes, for which they are very poorly suited for, and that most are starting out MCI type patients with an acetylcholinesterase inhibitor, which the data suggests is useless for anything besides relieving constipation.
But, all in all, not too surprising, seems to take a while for clinical data to actually start impacting clinical care in certain fields, and the management of AD is no exception. Medicine is full of "I do this because this is what I was taught" rather than "I do this, because this is what the best data available suggests I should do."