Page 113 - Reduction of coercive measures

Multidisciplinary reduction of coercive measures
Results
Descriptive analyses
At the time of the start of the study, 202 residents participated in the study, spread over 40 units. Coercive measures were applied to 169 residents, who were spread over 39 units. Analyses were carried out on coercive measures applied to 107 residents, spread over 30 units, as these residents were part of the sample for the entire duration of the study, i.e. these residents were resident in the unit in question until the end of the study.
Before the start and during the intervention period, N = 428 coercive measures were recorded distributed across 41 types. The most commonly recorded types were audio surveillance (on 40.2% of residents), a form of surveillance technology which detects a door being opened (on 25.2% of residents) and locking the outer doors (on 24.3% of residents). Sixteen types accounted for 80% of all coercive measures. Applied coercive measures were distributed over four subsets as follows: seven measures applied to prevent from direct danger, and 22 measures to prevent from indirect danger or disadvantage, 5 coercive measures arising from the use of surveillance technology, and 7 from the use support material. Application of coercive measures concerned 107 residents residing in 30 units.
Program effects
The proportion of coercive measures that were ended during and after the intervention period was 40.4% in the intervention group and 20.3% in the control group. The variance across units was significant (z = 2.05, p = .04); the variance across residents was not significant (z = 1.11, p = .27). The ICC correlation at the level of units was .25 and .06 at the level of residents. An ICC larger than .05 suggests that the dependencies due to the clustering cannot be ignored, and multilevel analysis is indicated (Hox et al., 2017). Analyses focusing on the degree of reduction of coercive measures before and during the intervention period, i.e. all registered cm, show a significant positive effect of the program on reduction of coercive measures (b = 1.42, t = 2.874, p = .009) (table 1), that is, in the intervention group the proportion of coercive
111
5