Table 2.
Main risk factors of mortality on the day of and day after the Great East Japan Earthquake and Tsunami from logistic regression models (on the day, Mar/11/2011, N = 860) and Cox proportional hazard models (after the disaster, Mar/12/2011 to May/5/2014, N = 827)*.
| On the day of the disaster (95% confidence interval) Adjusted odds ratio** | After the disaster Adjusted hazard ratio (95% confidence interval) | ||
|---|---|---|---|
| (a) Environmental factor | |||
| Distance from the coast | ≥2000 m | 1 | 1 |
| 1000–1999 m | 3.01 (0.56, 16.16) | 0.83 (0.42, 1.65) | |
| 500–999 m | 16.88 (4.33, 65.84) | 0.76 (0.38, 1.51) | |
| 0–499 m | 22.66 (5.78, 88.84) | 0.84 (0.43, 1.68) | |
| (b) Physical factors | |||
| Height | ≥160 cm | 1 | 1 |
| 150–159 cm | 2.20 (0.67, 7.19) | 1.06 (0.52, 2.17) | |
| <150 cm | 2.98 (0.75, 11.84) | 1.31 (0.57, 2.97) | |
| Sex | Women | 1 | 1 |
| Men | 2.74 (0.77, 9.68) | 3.33 (1.44, 7.73) | |
| Age | 65–69 years | 1 | 1 |
| 70–74 years | 2.25 (0.60, 8.44) | 3.53 (1.24, 10.03) | |
| 75–79 years | 5.29 (1.47, 19.04) | 3.41 (1.19, 9.80) | |
| 80–84 years | 2.32 (0.54, 9.98) | 4.65 (1.61, 13.46) | |
| ≥85 years | 5.41 (1.26, 23.16) | 10.90 (3.82, 31.10) | |
| (c) Social connectedness | |||
| Household | Living alone | 1 | 1 |
| Co-habiting with others, but not parent(s) | 3.04 (0.47, 19.74) | 1.05 (0.54, 2.06) | |
| Living with parent(s) | 6.67 (0.83, 53.71) | 0.45 (0.10, 2.12) | |
| Social interactions | Not meeting any friends | 1 | 1 |
| Meeting some friends | 2.06 (0.51, 8.23) | 0.46 (0.26, 0.82) | |
| (d) Health conditions | |||
| Depressive symptoms | Normal | 1 | 1 |
| Mild | 0.79 (0.29, 2.19) | 1.39 (0.81, 2.38) | |
| Moderate | 1.14 (0.29, 4.50) | 1.45 (0.65, 3.26) | |
| Severe | 3.90 (1.13, 13.47) | 1.91 (0.81, 4.50) | |
| Activities of daily living | Independent | 1 | 1 |
| Partially disabled | 0.73 (0.18, 2.89) | 2.44 (1.30, 4.56) | |
| Disabled | 0.32 (0.04, 2.64) | 2.97 (1.43, 6.14) | |
*Models considered pre-disaster baseline characteristics such as sex, age, education, number in household, social interactions, physical height (cm), body mass index, depressive symptoms, activities of daily living, comorbidity (cancer, heart diseases, stroke, respiratory diseases), and health behaviors (smoking, alcohol drinking, and exercise). **On the day of the disaster, to reduce the possibility of biased estimation from maximum likelihood estimation due to relatively smaller numbers of mortality events (N = 33), we applied logistic regression analysis with penalized maximum likelihood estimation.