Yeah, that's one argument - and is certainly true for later (medieval and early-modern) periods where urban disease rates were _much_ higher than rural.

The corollary hypothetical is that Roman cities had well-developed water and sewer infrastructure, whereas isolated forts were drinking out of wells and using latrines, with a higher probability of cross-contamination. They also might have had higher incidental population density (low-ranking troops sleeping in barracks; common meals) than at least some urban districts.

But... I don't know which case is actually true, and am curious what the evidence might show!