This nested case-control study within Korea National Health Insurance Database (NHID) analyzed administrative data from 2004-2018 (N=109,230) and a 1:4 sample of age and sex-matched controls. Urbanicity and age 0-29 years were associated with an increased risk of PTSD diagnosis. Among younger-age individuals, the estimated association was larger at both extremes of SES.

The question of the potential impact of living in urban environments on psychiatric problems is important, and as the authors note this has only been examined in relation to PTSD in two prior studies. Neither of those studies examined potential moderating effects of SES, which may be important. The use of an administrative database that includes 98% of the host country's population over a 15 year period and includes a population-based index of urbanicity of the community in which each individual resides and a proxy for their SES are strengths of the study. Although the report is well written and the findings are potentially of interest, several concerns

My major concern is that reliance on an administrative record of a clinical diagnosis to identify individuals with PTSD may create bias due to inaccurate or undercounted detection of PTSD, as the authors note. The very low incidence (<0.01% in 2004 to 0.02% in 2018) indicates that many cases of PTSD may be mistakenly considered to be no-PTSD controls, because these incidence levels are low compared to epidemiological research-based past-year prevalence estimates that tend to be greater than 1% in industrialized countries. Thus, the findings may reflect the greater degree of detection of PTSD rather than PTSD incidence per se. This may be due to greater access to healthcare or providers who have more training on or familiarity with PTSD in urban than rural settings, rather than differential exposure to stressors or trauma in urban settings. This may also help to explain the curvilinear moderation by SES, with lower income individuals potentially served by organizations or providers that are attuned to the impact of stress and trauma on their patients and higher income individuals more able to access specialized PTSD treatment and healthcare generally.

The evidence of greater psychiatric comorbidity amongst the PTSD cohort suggests that a sub-group of the PTSD cohort has more complicated clinical issues than PTSD alone. Were the findings comparable for those comorbid PTSD individuals to the findings for individuals diagnosed only with PTSD?

A related concern is that many individuals who are classified as no-PTSD controls may have other psychiatric conditions that could be trauma-related (e.g., anxiety or addictive disorders, depression; disruptive behavior disorders in children). Were the findings for the cohort of control individuals who had no psychiatric morbidity comparable to the findings for the cohort of control individuals who had psychiatric diagnoses?

Another concern is that the younger age group combines children, adolescents, and young adults These are quite heterogeneous sub-populations that may have correspondingly different experiences (including types of trauma exposures) of and developmental reactions to urban vs. rural life This makes it very difficult to interpret the finding that younger persons' urbanicity-related risk of PTSD was greater than that of mid-life or older adults -- this may be the case primarily for children, or for adolescents, or for young adults, rather than for all three developmental epochs, and if this is the case it would not be correct to conclude that simply being younger than 30 increases susceptibility to adverse effects of urbanicity.

The urbanicity measure also apparently considers only urban vs. rural as options, but there may be a substantial sub-group of person who live in suburban communities (as briefly alluded to in the Discussion) that do not fit with the definition of either urban or rural. How was this handled? Was the definition of "urban" essentially any community in which agriculture was not a primary use of land or occupation (if so, then urbanicity would be a mix of urban and suburban rather than specifically city-based urban settings)?

It is not clear why only age and SES were examined as moderators. Gender is a particularly relevant. For example, female gender is associated with PTSD and could increase their urbanicity-related risk of developing PTSD; however, on the other hand, males may be more likely than females to be involved in urban community violence, and this could increase their urbanicity-related risk of developing PTSD. A rationale for not considering gender as a moderator should be provided, and this should be noted as important for future study.

Minor edits: Line 155 "SESarate" should be "separate SES"; line 273 - this is a multi-year study but not a longitudinal study of the same persons over time, so I would re-phrase that as "multi-year" rather than "longitudinal."

Very responsive revision, excellent manuscript.