Neuner et al studied the risk of suicidal attempts for psychiatric inpatients. They developed a CART model for predicting the risk of a suicide attempt during a hospitalization. The authors are from the University of Regensburg in Germany.
The basic documentation system for German psychiatric hospitals is referred to as BADO.
Parameters:
(1) history of previous suicide attempt
(2) length of stay in days
(3) suicidal thoughts on admission
(4) compliance with psychotropic medications
(5) age at admission in years
(6) clinical global impression (CGI) on admission
(7) global assessment of function (GAF) on admission
If there is no history of a previous suicide attempt
|
Length of Stay (days)
|
Suicidal Thoughs on Admission
|
Percent with Suicide Attempt
|
|
<= 105.5
|
NA
|
0%
|
|
105.6 to 293
|
no
|
2%
|
|
105.6 to 293
|
yes
|
8%
|
|
> 293
|
NA
|
14%
|
If there is a history of a previous suicide attempt
|
Length of Stay
|
Compliance
|
Age
|
CGI
|
GAF
|
Percent Suicide Attempt
|
|
<= 96.5
|
no
|
NA
|
NA
|
NA
|
14%
|
|
<= 96.5
|
yes
|
NA
|
<= 6.5
|
NA
|
2%
|
|
<= 96.5
|
yes
|
NA
|
> 6.5
|
NA
|
6%
|
|
> 96.5
|
NA
|
<= 41.5
|
NA
|
<= 20.5
|
41%
|
|
> 96.5
|
NA
|
<= 41.5
|
NA
|
> 20.5
|
18%
|
|
> 96.5
|
NA
|
> 41.5
|
NA
|
NA
|
5%
|
where:
• A GAF <= 20.5 indicates a low global assessment of function.
• A CGI > 6.5 is considered high.
• A patient with a low GAF or high CGI on admission were at increased risk for rehospitalization and tended to have a chronic or recurring illness.
• A stay > 96.5 days was considered a long-stay.
Limitations:
• The length of hospital stay may be a retrospective finding.
