The biggest weakness in the study is that Taxi and ambulance drivers in the dataset died around 64 to 67, versus 74 for other occupations [0]. If Alzheimer's is much more likely to show up later, then lower Alzheimer's related death rates among Taxi and ambulance drivers may reflect earlier mortality rather than any effect from the job.
[0] https://www.sciencemediacentre.org/expert-reaction-to-study-...