Table 3.
Multivariate associations (ORs) between natural disaster trauma and onset of DSM-IV/CIDI mood and anxiety disordersa
| OR | (95% CI) | |
|---|---|---|
| Model 1b | ||
| Natural disaster | 1.4* | (1.3–1.6) |
| Model 2c | ||
| Natural disaster | 1.0 | (0.9–1.1) |
| Model 3d | ||
| Simple natural disaster | 1.1 | (0.9–1.3) |
| One related traumatic event | 1.3 | (1.0–1.6) |
| Two related traumatic events | 1.6* | (1.2–2.1) |
| Three or more related traumatic events | 1.7* | (1.4–2.2) |
A separate person-year file was created for each of the 9 mood and anxiety disorders, and these 9 files were then stacked. The models were estimated in a discrete-time survival framework with person-year as the unit of analysis using this stacked dataset, thereby forcing the slopes to be the same across the 9 disorders. Each model controlled for person-year, 8 dummy variables for the outcome disorder, country, age, and sex.
The model included a single dummy variable for lifetime exposure to natural disaster.
The model included a dummy variable for lifetime exposure to natural disaster and an additional dummy variable for exposure to each any other traumatic event occurring in the same year as the natural disaster.
The model included one dummy variable for lifetime exposure to simple natural disaster and additional dummy variables for complex natural disasters in which the respondent was exposed to 1, 2, and 3 or more related traumatic events in the same year as the natural disaster.
Significant at the 0.05-level, two-sided test.