Table 4.
Clinical prescribing error rates by type and changes over study periods.
| Error type | Control—Paper (n = 9635 orders) | Intervention—First 70 days of eMM (n = 16,734 orders) | Intervention—One-year post-eMM (n = 8891 orders) | |||||
|---|---|---|---|---|---|---|---|---|
| No. Errors | Errors/100 orders (95% CI)a | No. Errors | Errors/100 orders (95% CI)a | IRRb eMM vs Paper | No. Errors | Errors/100 orders (95% CI)a | IRRb One-year post-eMM vs Paper | |
| Wrong dose | 644 | 6.68 (6.19–7.22) | 1288 | 7.70 (7.29–8.13) | 1.22 (1.04–1.43) | 534 | 6.01 (5.52–6.54) | 0.79 (0.68–0.93) |
| Overdose | 308 | 3.20 (2.86–3.57) | 674 | 4.03 (3.74–4.34) | 1.22 (0.96–1.55) | 247 | 2.78 (2.45–3.15) | 0.79 (0.63–1.01) |
| Underdose | 223 | 2.31 (2.03–2.64) | 408 | 2.44 (2.21–2.69) | 1.25 (0.95–1.66) | 183 | 2.06 (1.78–2.38) | 0.83 (0.64–1.07) |
| Wrong frequency | 315 | 3.27 (2.93–3.65) | 522 | 3.12 (2.86–3.40) | 0.92 (0.73–1.16) | 157 | 1.77 (1.51–2.06) | 0.47 (0.36–0.61) |
| Wrong route | 199 | 2.07 (1.80–2.37) | 396 | 2.37 (2.14–2.61) | 1.65 (1.20–2.27) | 77 | 0.87 (0.69–1.08) | 0.44 (0.27–0.71) |
| Duplicated drug therapy | 140 | 1.45 (1.23–1.71) | 457 | 2.73 (2.49–2.99) | 1.82 (1.26–2.63) | 205 | 2.31 (2.01–2.64) | 1.48 (1.05–2.09) |
| Wrong drug | 24 | 0.25 (0.17–0.37) | 62 | 0.37 (0.29–0.48) | 1.82 (0.66–5.00) | 15 | 0.17 (0.10–0.28) | 0.49 (0.11–2.23) |
| Drug–drug interaction | 20 | 0.21 (0.13–0.32) | 71 | 0.42 (0.34–0.54) | 0.69 (0.23–2.09) | 17 | 0.19 (0.12–0.31) | 0.39 (0.09–1.73) |
| Allergy | 20 | 0.21 (0.13–0.32) | 19 | 0.11 (0.07–0.18) | 0.45 (0.12–1.75) | 13 | 0.15 (0.08–0.25) | 3.17 (0.48–20.9) |
| Inadequate monitoring | 19 | 0.20 (0.13–0.31) | 40 | 0.24 (0.18–0.33) | 1.19 (0.34–4.19) | 5 | 0.06 (0.02–0.14) | – |
– indicates too few errors for modelling (Supplementary Table 3 reports remaining error categories).
aCrude error rates and 95% CIs are from simple Poisson models.
bAdjusted incidence rate ratio (IRR) and 95% CIs are from generalised linear mixed models adjusting for study time (immediate post-eMM model only), patient age, and clustering by ward and patient.