The PROWESS trial of recombinant activated protein C in sepsis developed a multivariate equation to predict probability of patient death. This used several clinical and biochemical measures. The PROWESS investigators are from multiple university hospitals in the North America and Europe.
Parameters:
(1) age in years
(2) Apache II score
(3) chronic health points from the Apache II score
(4) serum IL-6 (interleukin-6) concentration in mg/dL
(5) dependency as measured using Katz's activities of daily living (ADL) score
(6) urosepsis
Parameter |
Finding |
Points |
chronic health points |
0 |
0 |
|
>= 1 |
1 |
ADL |
no dependencies in ADL |
0 |
|
dependent in >= 1 activities |
1 |
urosepsis |
absent |
0 |
|
present |
1 |
where:
• The index of ADL involves bathing, dressing, toilet, transfer, continence, and feeding (see Index of Independence of Katz et al in 37.05.05).
• Chronic health points (see Apache II score, above) would be "positive" with a history of severe organ failure or being immunocompromised.
X =
= (0.054 * (Apache II score)) + (0.030 * (age in years)) + (0.211 * LN(IL-6 concentration)) + (0.494 * (dependency points)) - (0.781 * (urosepsis points)) + (0.639 * (comorbidity points)) - 5.75
probability of mortality =
= EXP(X) / (1 + EXP(X))
Limitations:
• The model has not been validated as a tool for making individual patient risk assessment.
Specialty: Critical Care, Emergency Medicine