Your assertion, such as it is, is poorly formed. This research did not, as you imply, test 200 potential associations and report the one that was the “winner”. A single /a priori/ hypothesis was formulated and then tested using the data, in accordance with standard statistical methodology. The logical extension of your statement is that no association could ever be found since anything found is just a green-jelly bean association. When multiple tests of hypotheses are performed, accepted statistical practice is to the make the criterion for significance more stringent precisely to avoid the green-jellybean effect you are implying has occurred here.

Similarly illogical is pointing out the size of the questionnaire - as though the number of questions a person is asked has any impact whether eggs have an effect on Alzheimer’s disease incidence.