Table 3

 Work related assault: impact of policy logistic regression models

Policy componentModel 1Model 2Model 3
OR95% CIOR95% CIOR95% CI
Model 1: Univariate.
Model 2: Multivariate: adjusted for: Workplace [type of facility (Q1), ownership of facility (Q2), location of facility (Q3), total beds at facility (Q4)], Administration attitude [administration’s expectation of violence (Q24A), administration’s corrective measures against violence (Q24B)], Department/unit/area (Q8), and Types of patients [age of population (Q5), race of patients (Q12A), patient gender (Q12B), patient mental status (13B), average length of patient stay (Q15)].
Model 3: Adjusted and weighted multivariate: adjusted for all variables in Model 2, and also non-response and non-selection.
Zero tolerance0.370.26 to 0.510.540.35 to 0.840.700.38 to 1.51
Types of prohibited behaviours0.450.31 to 0.650.540.33 to 0.900.560.32 to 1.43
Confidentiality in reporting0.700.49 to 1.010.830.50 to 1.381.050.54 to 2.76
Flagging charts0.740.54 to 1.020.760.50 to 1.140.660.39 to 1.30
Consequences for using violence at work0.810.53 to 1.220.970.56 to 1.711.2010.50 to 2.94
How to report non-physical violence0.940.57 to 1.521.010.53 to 1.921.700.57 to 4.53
How to report physical violence0.950.59 to 1.520.700.36 to 1.340.820.35 to 3.02
Required training policies1.000.74 to 1.340.930.63 to 1.371.110.62 to 1.93