Niskanen et al developed 2 models based on logistic regression to predict in-hospital mortality for patients admitted to the intensive care unit (ICU) as a result of a gastroenterologic emergency. This can help identify patients who may benefit from more aggressive interventions. The authors are from 6 university hospitals in Finland.
Parameters |
Model A |
Model B |
---|---|---|
APS on day 1 |
0.164 * APS |
0.143 * APS |
TISS on day 1 |
0 |
0.041 * TISS |
age in years |
0.030 * age |
0.030 * age |
postoperative admission |
add 0.076 |
add -0.097 |
liver disease present |
add 0.944 |
add 0.938 |
cardiovascular disease present |
add 0.559 |
add 0.672 |
respiratory disease present |
add 0.416 |
add 0.441 |
renal disease present |
add -0.216 |
add -0.104 |
immunocompromised state present |
add 0.674 |
add 0.560 |
gastrointestinal bleeding present |
add 0.364 |
add 0.423 |
sepsis present |
add 0.690 |
add 0.555 |
neoplasm present |
add 0.659 |
add 0.751 |
respiratory insufficiency after surgery |
add -0.563 |
add -0.492 |
hemorrhagic shock present |
add -0.292 |
add -0.470 |
bowel perforation or obstruction present |
add -0.061 |
add -0.084 |
constant |
-4.217 |
-5.088 |
where:
• The Acute Physiology Score (APS) was calculated from APACHE II.
A =
= SUM(items present in Model A)
B =
= SUM(items present in Model B)
probability for in-hospital death using Model A =
= 1 / (1 + EXP((-1) * A))
probability for in-hospital death using Model B =
= 1 / (1 + EXP((-1) * B))
Performance:
• The ROC plots show both models to be slightly better than APACHE II (Figure 3, page 591).
• The addition of TISS in Model B did not provide much benefit.
Limitations:
• The authors point out that the results of the calculations should be used with caution since the number of patients was too low to guarantee the accuracy for the outcome prediction.
Purpose: To predict the in-hospital mortality for a patient admitted to the intensive care unit (ICU) following a gastroenterologic emergency.
Specialty: Critical Care, Emergency Medicine
Objective: severity, prognosis, stage
ICD-10: R57,