A PI started a study and collected some data. The study was originally meant to evaluate the effectiveness of a treatment. After a year or so of collecting data, the PI decided to do some analyses. When I looked at the design of the study, I informed the PI that this design will not be able to provide any meaningful evaluation of effectiveness.
Some time later, I came across a novel method to analyse this data and tried it out. The results were quite interesting and with the PI's agreement, I decide to try and publish it.
My question is this: Since the PI was not involved at all in the conceptualisation of this novel method and did not contribute in the interpretation of the results, should he be granted authorship? According to ICMJE guidelines, the answer should be no but I am unsure.
Addendum: I would like to add, while thinking about this problem, I thought of big studies like the Framingham Heart Study, the Global Burden of Disease study. The PI of these studies are not automatically authors on every paper that came out of the data they collected. Would this be any different from my situation?