Table 3. The Relationship between “End-of-Life” Treatment Intensity and Mortality: Adjusted Odd Ratios and Absolute Risk Differences.
| Mortality Measure and Population | Hospital EOL Treatment Intensity | ||
|---|---|---|---|
| 1SD below average | average | 1SD above average | |
| Adjusted* Odds Ratio (OR); (95% CI) | |||
| 30-day post-admission mortality | |||
| Patients at low† probability of dying | 1.06 (1.04 – 1.08) | 1.0 (reference) | 0.97 (0.96 – 0.99) |
| Patients at average‡ probability of dying | 1.06 (1.04 – 1.09) | 1.0 (reference) | 0.97 (0.96 – 0.99) |
| Patients at high¶ probability of dying | 1.09 (1.07 – 1.11) | 1.0 (reference) | 0.97 (0.95 – 0.99) |
| 180-day post-admission mortality | |||
| Patients at low† probability of dying | 1.03 (1.01 – 1.04) | 1.0 (reference) | 1.00 (0.98 – 1.01) |
| Patients at average‡ probability of dying | 1.03 (1.02 – 1.05) | 1.0 (reference) | 1.00 (0.98 – 1.01) |
| Patients at high¶ probability of dying | 1.06 (1.04 – 1.09) | 1.0 (reference) | 1.00 (0.98 – 1.02) |
| Adjusted* absolute risk difference, %; (95% CI) | |||
| 30-day post-admission mortality | |||
| Patients at low† probability of dying | 1.4 (0.9 – 1.9) | 0 (reference) | -0.7 (-1.1 to -0.3) |
| Patients at average‡ probability of dying | 1.6 (1.1 – 2.1) | 0 (reference) | -0.7 (-1.1 to -0.3) |
| Patients at high¶ probability of dying | 2.1 (1.6 – 2.7) | 0 (reference) | -0.8 (-1.2 to -0.4) |
| 180-day post-admission mortality | |||
| Patients at low† probability of dying | 0.6 (0.2 – 1.0) | 0 (reference) | -0.1 (-0.4 to 0.2) |
| Patients at average‡ probability of dying | 0.8 (0.4 – 1.2) | 0 (reference) | -0.1 (-0.004 to 0.2) |
| Patients at high¶ probability of dying | 1.5 (0.9 – 2.0) | 0 (reference) | -0.1 (-0.5 to 0.4) |
*
Models directly adjusted for patient demographic and clinical characteristics and treating hospital characteristics; Regressions weighted by inverse-probability-of-treatment weights derived from models using patient demographic and clinical characteristics from the current and immediate past prior hospitalization plus the EOL treatment intensity index of the prior hospitalization and specific intensive treatments received during the prior hospitalization.
†
0%,
‡
4.6%, and
¶
41% predicted probability of inpatient death