Table 4.

Multivariate analysis* of the comprehensive prevention programme on violence rates by admissions characteristics

Characteristics	Pre-intervention trend (per month)†		Change in level‡		Change in trend (per month)§
	IRR (95% CI)	P value	IRR (95% CI)	P value	IRR (95% CI)	P value
Sex
Male	1.05 (0.76 to 1.46)	0.7500	0.59 (0.28 to 1.20)	0.1308	0.95 (0.89 to 1.01)	0.0810
Female	1.27 (0.84 to 1.93)	0.2343	0.35 (0.15 to 0.83)	0.0212	1.00 (0.93 to 1.07)	0.9548
Age (years)
<40	1.11 (0.78 to 1.58)	0.5292	0.43 (0.19 to 0.99)	0.0471	0.96 (0.90 to 1.04)	0.2771
≥40	1.16 (0.79 to 1.69)	0.4107	0.51 (0.24 to 1.08)	0.0730	0.97 (0.92 to 1.04)	0.3601
Waiting time (hours)
≤2	1.11 (0.67 to 1.85)	0.6468	0.39 (0.13 to 1.18)	0.0892	0.96 (0.88 to 1.05)	0.3427
>2	1.12 (0.83 to 1.51)	0.4233	0.49 (0.26 to 0.92)	0.0306	0.99 (0.94 to 1.04)	0.6704
Length of stay (hours)
≤3	1.03 (0.66 to 1.62)	0.8738	0.57 (0.22 to 1.51)	0.2329	0.96 (0.89 to 1.04)	0.2823
>3	1.13 (0.82 to 1.55)	0.4231	0.38 (0.20 to 0.74)	0.0089	1.00 (0.94 to 1.06)	0.9764

*Segmented Poisson regression offset by the total number of admissions at OED per month. RR<1 represents a decline and conversely, RR>1 represents an increase in monthly violence rate.

†Rate of change in monthly violence rate prior to the intervention (ie, time effect).

‡Immediate change in the mean monthly violence rate from pre-intervention to intervention period.

§Change in slope per month following the intervention period.

IRR, incidence rate ratio; OED, ophthalmology emergency department.