Skip to main content

Table 3 Characteristics of absconding incidents (N  = 188)

From: The impact of structured decision making on absconding by forensic psychiatric patients: results from an A-B design study

 

A Sept 1 2010 – Feb 29 2012 ( n  = 89)

Interim Mar 1 2012 – Aug 31 2012 ( n  = 40)

B Sept 1 2012 – Feb 28 2014 ( n  = 59)

 
 

Mdn/M (SD)

Mdn/M (SD)

Mdn/M (SD)

KW H

Duration (hours)

8.5/41.1 (95.5)

6.5/53.7 (174.2)

4.3/51.1 (144.7)

5.33

Method

 

%

 

χ 2

  Off locked unit

12.4

5.0

23.7

7.35*

  From staff accompanied outing

12.4

10.0

28.8

8.57*

  From unaccompanied hospital pass

41.6

65.0

33.9

9.81**

  From unaccompanied community pass

30.3

15.0

13.6

7.24*

Location during leave

 

%

 

χ 2

  Within city limits, outdoors or public place

57.3

45.0

69.5

5.99*

  Own home

11.2

0.0

3.4

--

  Friends/family home

21.4

27.5

22.0

0.68

  Shelter

13.5

12.5

5.1

2.82

  Hospital grounds

9.0

27.5

20.3

7.79*

Substance use (yes)

31.5

35.0

27.1

0.72

Reoffense (yes)

2.3

0.0

0.0

--

Violence – perpetrator

2.3

2.5

6.8

--

Violence – victim

1.1

0.0

0.0

--

Form of return

 

%

 

χ 2

  Self

55.1

65.0

47.5

2.97

  Police

33.7

25.0

28.8

1.08

  Hospital staff

4.5

7.5

17.0

6.84*

  Family member

3.4

2.5

3.4

--

  1. Note. Outliers were removed in calculating the Duration variable: three incidents that lasted 117, 125, and 130 days, respectively. KW H  = Kruskal-Wallis H test.
  2. *p < .05.**p ≤ .01. The Freeman-Halton extension of the Fisher exact probability test (two-tailed) was performed alongside the χ2 computation.
  3. --χ2 not calculated due to > 20% of cells having sample sizes < 5.