Multilevel survival models of overdose outcomes
| Conditional effects | Repeat non-fatal overdose n=16 417 | Drug-related death n=15 998 | Any death n=16 417 | ||||||
| HR (SE) | CI | P value | HR (SE) | CI | P value | HR (SE) | CI | P value | |
| Between-patient (L1) | |||||||||
| Race (White) | 0.76 (0.09) | (0.51 to 1.14) | 0.093 | 0.60 (0.18) | (0.22 to 1.60) | 0.085 | 0.88 (0.11) | (0.58 to 1.34) | 0.317 |
| Sex (male) | 0.88 (0.05) | (0.73 to 1.06) | 0.021 | 0.84 (0.10) | (0.56 to 1.25) | 0.151 | 0.82 (0.05) | (0.68 to 0.99) | <0.001 |
| 25–34 (15–24) | 1.02 (0.07) | (0.80 to 1.29) | 0.801 | 1.75 (0.32) | (0.96 to 3.17) | 0.002 | 1.60 (0.25) | (0.96 to 2.65) | 0.002 |
| 35–44 (15–24) | 0.75 (0.06) | (0.56 to 0.99) | 0.001 | 1.54 (0.31) | (0.79 to 3.01) | 0.035 | 2.20 (0.35) | (1.30 to 3.72) | <0.001 |
| 45–54 (15–24) | 0.70 (0.07) | (0.51 to 0.96) | <0.001 | 1.73 (0.37) | (0.86 to 3.47) | 0.010 | 3.92 (0.60) | (2.38 to 6.48) | <0.001 |
| 55–64 (15–24) | 0.73 (0.08) | (0.52 to 1.04) | 0.003 | 1.24 (0.41) | (0.54 to 2.82) | 0.389 | 6.57 (0.97) | (4.03 to 10.71) | <0.001 |
| 65–74 (15–24) | 0.26 (0.05) | (0.13 to 0.52) | <0.001 | 0.10 (0.10) | (<0.01 to 2.76) | 0.022 | 11.00 (1.67) | (6.67 to 18.14) | <0.001 |
| 75–84 (15–24) | 0.24 (0.07) | (0.09 to 0.60) | <0.001 | 0.33 (0.24) | (0.03 to 3.65) | 0.132 | 16.30 (2.52) | (9.79 to 27.13) | <0.001 |
| 85 plus (15–24) | 0.09 (0.05) | (0.01 to 0.58) | <0.001 | – | – | – | 27.39 (4.28) | (16.38 to 45.81) | <0.001 |
| Repeat non-fatal overdose (no) | – | – | – | 3.02 (0.40) | (1.94 to 4.69) | <0.001 | 1.69 (0.16) | (1.24 to 2.30) | <0.001 |
| Random effects | |||||||||
| Between-county variability | 0.23 (0.05) | (0.10 to 0.51) | |||||||
HR=hazard ratio produced by multilevel survival models. CI refers to 99.9% confidence interval for L1 effects. For categorical variables, reference categories are listed in parentheses. For fatal overdose, adults aged 85 plus were removed from the final model due to no instances of fatal overdose in that age group.
L1, level 1.