Gajic et al used the Stability and Workload Index for Transfer (SWIFT) score to identify patients discharged from the ICU who be readmitted. The authors are from the Mayo Clinic and University of Amsterdam.
Patient selection: ICU care for > 24 hours
Parameters:
(1) original location prior to ICU admission
(2) length of stay in the ICU in days
(3) last measured PaO2 to FIO2 ratio during the ICU admission
(4) Glasgow Coma Score (GCS) at the time of ICU discharge
(5) last arterial PaCO2 during the ICU admission
Parameter
Finding
Points
source of ICU admission
emergence department
0
from hospital ward or outside hospital
8
length of stay in the ICU
< 2 days
0
2 to 10 days
1
> 10 days
14
last measured PaO2 to FIO2 ratio
> 400
0
150 to 400
5
100 to 149.9
10
< 100
13
Glasgow Coma Score at discharge
15
0
11 to 14
6
8, 9 or 10
14
3 to 7
24
last PaCO2
<= 45 mm Hg
0
> 45 mm Hg
5
total SWIFT score =
= SUM(points for all 5 parameters)
Interpretation:
• minimum score: 0
• maximum score: 64
• The lower the score the healthier the patient at the time of discharge.
• As shown in Figure 2 (page 681) the rate of readmission for given ranges of the SWIFT score varies. In general a score >= 31 is associated with the highest rates of readmission while a score <= 10 has the lowest.
Performance:
• The ROC curves on page 680 indicate less than ideal performance, with AUCs in the 0.70 range.
To read more or access our algorithms and calculators, please log in or register.
