Description

Johnson et al reported the Oxford Acute Severity of Illness Score (OASIS) for describing the illness severity of a patient in the intensive care unit (ICU). The authors are from Oxford University and Cerner Corporation.


Patient selection: ICU patients >= 16 years of age, non-burn, admitted >= 4 hours

 

Parameters:

(1) length of stay (LOS) prior to ICU admission

(2) reason for admission

(3) age in years

(4) Glasgow coma score (GCS)

(5) mean arterial pressure in mm Hg

(6) respiratory rate in breaths per minute

(7) temperature in °C

(8) urine output in mL per day

(9) ventilation

 

Parameters

Findings

Points

length of stay prior to ICU

< 0.17 hours

5

 

0.17 to 4.94 hours

3

 

4.95 to 24 hours

0

 

24.01 to 311.8 hours

2

 

> 311.8 hours

1

reason for admission

elective surgery

0

 

other

6

age in years

16 to 23 years

0

 

24 to 53 years

3

 

54 to 77 years

6

 

78 to 89 years

9

 

>= 90 years

7

Glasgow coma score

15

0

 

14

3

 

8 to 13

4

 

3 to 7

10

mean arterial pressure

> 143.44 mm Hg

3

 

61.33 to 143.44 mm Hg

0

 

51 to 61.32 mm Hg

2

 

20.65 to 50.99 mm Hg

3

 

< 20.65 mm Hg

4

respiratory rate

< 6 breaths per minute

10

 

6 to 12

1

 

13 to 22

0

 

23 to 30

1

 

31to 44

6

 

>= 45 breaths per minute

9

temperature in °C

< 33.22°C

3

 

33.22 to 35.93

4

 

35.94 to 36.39

2

 

36.4 to 36.88

0

 

36.89 to 39.88

2

 

>= 39.89°C

6

urine output in mL/day

< 671 mL

10

 

671 to 1,426.99

5

 

1427 to 2543.99

1

 

2544 to 6896

0

 

>= 6897 mL

8

ventilated

no

0

 

yes

9

 

total score =

= SUM(points for all 9 parameters)

 

Interpretation:

• minimum score: 0

• maximum score: 75

• The higher the score the more severe the patient's illness, the higher the ICU mortality and the hospital mortality.

• A score less than 10 was associated with low mortality while a score above 25 was associated with a high mortality.

 

Performance:

• The areas under the ROC curve was 0.90 in validation data set using the multivariable model.


To read more or access our algorithms and calculators, please log in or register.