Bota et al developed the Infection Probability Score (IPS) to identify patients in the Intensive Care Unit (ICU) who might be infected. This can help stratify patients into risk groups. The authors are from the University of Brussels in Belgium.
Parameters:
(1) body temperature
(2) heart rate
(3) respiratory rate
(4) white blood cell (WBC) count
(5) C reactive protein (CRP)
(6) Sequential Organ Failure Assessment (SOFA) score
|
Parameter |
Findings |
Points |
|---|---|---|
|
body temperature |
> 37.5°C |
2 |
|
|
<= 37.5°C |
0 |
|
heart rate |
> 140 beats per minute |
12 |
|
|
81 - 140 beats per minute |
8 |
|
|
<= 80 beats per minute |
0 |
|
respiratory rate |
> 25 breaths per minute |
1 |
|
|
<= 25 breaths per minute |
0 |
|
WBC count |
> 12,000 per µL |
1 |
|
|
5,000 - 12,000 per µL |
0 |
|
|
< 5,000 per µL |
3 |
|
C reactive protein |
> 6 mg/dL |
6 |
|
|
<= 6 mg/dL |
0 |
|
SOFA score |
> 5 |
2 |
|
|
<= 5 |
0 |
where:
• I could not find a discussion on how patients on a ventilator were handled.
• I would think the significance of a low WBC count would depend on whether the patient was neutropenic to start with, or developed leukopenia as a sign of infection.
total score =
= SUM(points for all 6 parameters)
Interpretation:
• minimum score: 0
• maximum score: 26
• A patient with a score < 14 is unlikely to be infected.
• It might be interesting to see if the change in the score during the ICU stay correlates with infection.
• On page 2580 the authors mentioning using the score in a logistic regression equation to predict probability of infection, but the beta coefficient and constant were not given.
Performance (Figures 1 and 2, page 2582):
• In the development set a score > 14 is used to identify infection; the sensitivity was 74% and specificity 78%. The positive predictive value for > 14 was 53.6% and negative predictive value with < 14 is 89.5%. (I cannot see what a score of 14 means).
• In the validation set a score > 13 was used to identify infection; the sensitivity was 90% and specificity was 89%. The positive predictive value was 72.2% and negative predictive value was 95.9%.
Limitations:
• Much of the discussion in the paper involved the importance of sepsis in critically ill patients. Distinction between a localized infection and generalized sepsis seems a little fuzzy to me.
• The single most important variable was heart rate. This outweighs most of the other variables.
Purpose: To evaluate a patient in the intensive care unit (ICU) for risk of serious infection using the Infection Probability Score (IPS).
Specialty: Critical Care, Emergency Medicine, Infectious Diseases
ICD-10: A00-B99, A49.9,
