Arte SheetsLearn. Try. Understand.

Healthcare & Clinical Formulas

SAPS II Research and a Fictional Logistic Formula

Read about the original SAPS II hospital-outcome model and practice a separate fictional logistic calculation.

The original research and its limits

Le Gall, Lemeshow and Saulnier developed SAPS II using data from medical and surgical intensive care units. The original model combined 17 variables and estimated hospital mortality using a logistic relationship. It was designed for research and comparison in defined cohorts; performance and calibration can change across hospitals and populations.

The published relation uses z = −7.7631 + 0.0737 × S + 0.9971 × LN(S + 1), followed by 1 / (1 + EXP(−z)), where S is the correctly constructed SAPS II score. The full score requires the specified individual variables and component rules; arbitrary grouped inputs are not enough to verify it or establish an individual prognosis.

A separate fictional logistic exercise

The formula below uses invented coefficients and an arbitrary fictional input X. It is not SAPS II and does not estimate mortality or any other clinical outcome. Its only purpose is to show how LN and EXP can transform a number into a value between 0 and 1.

Enter a numeric X from 0 to 100 in B2. With X = 20, the fictional expression rounds to 0.3138. Changing X to 50 produces 0.7451. Missing, text and out-of-range values produce a message.

Fictional input XFormula outputInterpretation
00.0474Invented model value
200.3138Invented model value
500.7451Invented model value
=IF(COUNT(B2)<>1,"Enter a numeric value",IF(OR(B2<0,B2>100),"Enter a value from 0 to 100",ROUND(1/(1+EXP(-(-3+0.05*B2+0.4*LN(B2+1)))),4)))

Read the research in context

The original publication and later calibration study explain how model estimates were assessed in different cohorts. This example cannot diagnose disease, guide ICU care, or determine a person’s chance of survival.

Try the example

Use the original values or change them to test your understanding.

Further reading