Table 1.
Descriptive statistics
| Variable | Complete data set* n=123 975 |
Validation set n=55 164 | Development set n=65 776 | |||
| N missing (%) | % exposed† | % exposed | % experiencing outcome in exposed | % experiencing outcome in unexposed | P value‡ | |
| Female | 0 (0.0) | 49.5 | 49.0 | 0.19 | 0.22 | 0.48 |
| Arrived by ambulance | 0 (0.0)§ | 13.6 | 11.1 | 0.87 | 0.12 | <0.001 |
| Referred by physician | 0 (0.0) | 14.0 | 10.1 | 0.36 | 0.19 | 0.006 |
| Triage priority 1 | 0 (0.0) | 0.8 | 0.9 | 1.48 | 0.19 | <0.001 |
| Triage priority 2 | 0 (0.0) | 13.1 | 14.8 | 0.41 | 0.17 | <0.001 |
| Radiology order in ED | 0 (0.0)¶ | 18.1 | 12.8 | 0.27 | 0.20 | 0.19 |
| Left against medical advice | 0 (0.0) | 5.0 | 5.1 | 0.09 | 0.21 | 0.18 |
| Discharged night-time | 0 (0.0) | 30.4 | 33.5 | 0.18 | 0.22 | 0.36 |
| Discharged weekend | 0 (0.0) | 31.0 | 33.0 | 0.17 | 0.23 | 0.12 |
| Discharged summer | 0 (0.0) | 15.2 | 14.7 | 0.11 | 0.22 | 0.04 |
| Discharged winter | 0 (0.0) | 23.3 | 23.4 | 0.22 | 0.20 | 0.73 |
| Male provider | 3385 (2.73) | 44.2 | 43.9 | 0.24 | 0.18 | 0.09 |
| Junior physician | 3385 (2.73) | 22.5 | 25.2 | 0.25 | 0.19 | 0.22 |
| Non-physician provider | 3385 (2.73) | 7.1 | 14.3 | 0.11 | 0.22 | 0.03 |
| Mortality | 0 (0.0) | 0.15 | 0.21 | N/A | N/A | N/A |
| Median (IQR) | Median (IQR) | Median (IQR) in subjects experiencing outcome | Median (IQR) in subjects not experiencing outcome | P** | ||
| Age (years) | 0 (0.0) | 42.0 (20.0, 66.0) |
31.0 (12.0, 58.0) |
81.0 (71.8, 89.0) |
31.0 (12.0, 58.0) |
<0.001 |
| Comorbidity score | 3035 (2.45) | 0.0 (0.0, 0.0) |
0.0 (0.0, 0.0) |
2.0 (1.0, 6.0) |
0.0 (0.0, 0.0) |
<0.001 |
| ED census (N) | 0 (0.0) | 29.0 (20.0, 36.0) |
30.0 (22.0, 37.0) |
33.0 (25.0, 39.0) |
30.0 (22.0, 37.0) |
0.02 |
| Hospital bed occupancy (%) | 0 (0.0) | 92.0 (87.8, 96.6) |
89.1 (84.1, 93.5) |
90.1 (83.9, 93.8) |
89.1 (84.1, 93.5) |
0.87 |
*N before excluding missing values.
†Proportion of subjects sharing characteristic indicated in ‘variable’ column.
‡P-value for difference in outcome, exposed vs unexposed, non-adjusted, development set. Arrived by ambulance, referred by physician, triage priority 1 and 2, discharged summer, non-physician provider with p<0.05.
§Database-linkage between source table and ambulance dispatches for 14 918 (12.0%) subjects.
¶Database-linkage between source table and radiology orders for 18 435 (14.9%) subjects.
**P-value for difference in predictor distribution, subjects experiencing outcome vs subjects not experiencing outcome, non-adjusted, development set. Age, comorbidity score and ED census with p<0.05.
ED, emergency department.