Table 3.
Differences between patients with and without diagnostic discrepancy
| Measure | Total n = 755 | Without diagnostic discrepancy n = 662 (87.68%) | With diagnostic discrepancy n = 93 (12.32%) | p $ | Effect Size* | ||
|---|---|---|---|---|---|---|---|
| Type | Estimate | 95% CI§ | |||||
| At presentation to the ER | |||||||
| Age (years; mean [SD]) | 65.14 (18.4) | 64.84 (18.68) | 67.21 (16.17) | 0.199 | Cohen’s d | 0.13 | −0.10 – 0.35 |
| Female (n [%]) | 322 (42.65%) | 278 (41.99%) | 44 (47.31%) | 0.360 | Odds Ratio | 1.23 | 0.79–1.89 |
| Non-specific chief complaint (n [%]) | 165 (21.85%) | 143 (21.6%) | 22 (23.66%) | 0.809 | Odds Ratio | 0.95 | 0.62–1.41 |
| Triage category (n [%]) | |||||||
| See immediately | 43 (5.7%) | 36 (5.44%) | 7 (7.53%) | 0.281 | Kendall’s τ | 0.04 | −0.03 – 0.11 |
| See within 20 min | 316 (41.85%) | 275 (38.82%) | 41 (44.09%) | ||||
| See within 120 min | 367 (48.61%) | 324 (48.94%) | 43 (46.24%) | ||||
| See today | 23 (3.05%) | 21 (3.2%) | 2 (2.15%) | ||||
| Non-urgent | 6 (0.8%) | 6 (0.91%) | 0 | ||||
| Via resuscitation bay (n [%]) | 81 (10.73%) | 69 (10.42%) | 12 (12.9%) | 0.463 | Odds Ratio | 1.28 | 0.66–2.48 |
| Time at ER (hours; mean [SD]) | 6.54 (2.97) | 6.5 (3.02) | 6.87 (2.58) | 0.202 | Cohen’s d | 0.13 | −0.09 – 0.34 |
| Charlson comorbidity index (points; median [IQR]) | 4.28 (2.87) | 4.27 (2.89) | 4.38 (2.76) | 0.711 | Kendall’s τ | 0.01 | −0.04 – 0.07 |
| Number of active medications (median [IQR]) | 11.46 (8.94) | 11.37 (8.86) | 12.04 (9.54) | 0.590 | Kendall’s τ | 0.02 | −0.05 – 0.08 |
| Number of groups of medications (median [IQR]) | 3.03 (2.15) | 3.02 (2.10) | 3.21 (2.40) | 0.547 | Kendall’s τ | 0.02 | −0.05 – 0.09 |
| Outcome | |||||||
| Length of hospital stay (LOS) (days; mean [SD]) | 7.32(7.19) | 6.90 (6.56) | 10.29 (10.14) | 0.002 | Cohen’s d | 0.47 | 0.26–0.70 |
| Mortality (n [%]) | 33 (4.37%) | 25 (3.78%) | 8 (8.60%) | 0.038 | Odds Ratio | 2.40 | 1.05–5.50 |
All percentages refer to number of patients within column
$ For difference between patients with and without diagnostic discrepancy, two-sided p values are reported
* Odds ratio (OR) for dichotomous variables; Kendall’s τ for ordinal variables, effect size d for metric variables;
§ CI is the Confidence Interval, for Kendall’s τ this was determined by bootstrapping with 2000 repetitions