Effect of ED-based specialised models of care on practice delivery and costs
Study | Intervention | Comparison | Sample size | Age (years) (mean (SD)) | Outcome | Measure of effect | Effect (intervention vs comparison) |
Grover and Lee29 | Behavioural health unit | Standard ED room | I: 2475 C: 2168 | I: 13.2 (3.28) C: 13.6 (3.14) | Length of stay | Absolute difference | −45 min |
Security man-hours per 100 patients | Mean difference | −36.8 hours (95% CI −48.3 to −25.3) | |||||
Restraint orders per 100 patients | Mean difference | −0.9 orders (95% CI −1.64 to −0.16) | |||||
Mahajan et al 30 | Child guidance model | Standard ED care | I: 561 C: 470 | 12.5 (3.4) | Length of stay | Mean difference | −43.1 min (95% CI −63.1 to −23.1) |
Cost of ED care | Cost per patient per hour | US$18.98 | |||||
Parker et al 26 | Rapid response model | Standard ED care | NTD | NTD | Monthly admissions to inpatient psychiatry resultant from RRM consultation | Mean differences* | First RRM vs pre-RRM: −0.3 RRM termination vs first RRM: 1.2 Second RRM vs RRM termination: −4.9 |
Greenfield et al 31 | Emergency room follow-up team | Standard ED care | I: 568 C: 412 | 14.6 (1.5) | Number of hospitalisations† over a 1-year period | Relative risk | 0.56 (95% CI 0.46 to 0.69) |
Risk difference | − 16.1% (95% CI − 21.8% to − 10.4%) | ||||||
Number needed to treat | 7 | ||||||
Number of ED return visits over a 1- year period | Relative risk | 0.61 (95% CI 0.32 to 1.17) |
*Unable to calculate 95% CIs due to unknown sample size; using study-reported data to clarify the findings, we are confident of a statistically significant difference for the last comparison (second RRM vs RRM termination), but not clear about whether there are significant differences for the first two comparisons.
†Medical or psychiatric hospitalisation.
CI, Confidence interval; NTD, not able to determine; RRM, rapid response model.