Clinical Medicine & Physiology
TRISS: Historical Trauma Outcome Models and Logistic Formula Example
Explore one historical MTOS coefficient set and the logistic function using fictional trauma scores, with explicit limits on interpretation.
A historical model for comparing outcomes
The Trauma and Injury Severity Score (TRISS) method combines an anatomical injury score, a physiological trauma score, age and mechanism in a logistic model. Boyd and colleagues described the method for trauma outcome and quality-assurance comparisons in 1987.
The example below uses the MTOS coefficient set listed as 1990 in the American College of Surgeons National Trauma Data Bank report. Other coefficient sets and later revisions exist; they produce different estimates and should not be conflated. A historical group-based estimate is not a dependable individual prognosis or a treatment or triage instruction.
The model does not include every determinant of outcome, and its performance varies with population, mechanism, missing data and changes in trauma care. The arithmetic shown here has not been clinically validated.
| Mechanism | Intercept | RTS coefficient | ISS coefficient | Age ≥55 coefficient |
|---|---|---|---|---|
| Blunt | −1.2470 | 0.9544 | −0.0768 | −1.9052 |
| Penetrating | −0.6029 | 1.1430 | −0.1516 | −2.6676 |
Compute the example linear predictor
Use fictional values: B2 is BLUNT or PENETRATING; B3 is a weighted Revised Trauma Score (RTS); B4 is an integer Injury Severity Score (ISS); and B5 is age in completed years. The age indicator is 1 at age 55 or older and 0 below 55. The linear predictor is intercept + RTS coefficient × RTS + ISS coefficient × ISS + age coefficient × indicator.
The input checks restrict this example to age 15 or older, RTS 0–7.8408, and integer ISS 1–75. The original historical method used a special coefficient convention for younger children, which this demonstration does not implement. These simple range checks cannot establish that an RTS or ISS was derived correctly from valid clinical observations.
=IF(COUNT(B3:B5)<>3,"Enter all numeric inputs",IF(AND(UPPER(TRIM(B2&""))<>"BLUNT",UPPER(TRIM(B2&""))<>"PENETRATING"),"Choose BLUNT or PENETRATING",IF(OR(B3<0,B3>7.8408,B4<1,B4>75,INT(B4)<>B4,B5<15,INT(B5)<>B5),"Check input ranges",ROUND(IF(UPPER(TRIM(B2&""))="BLUNT",-1.247+0.9544*B3-0.0768*B4-1.9052*IF(B5>=55,1,0),-0.6029+1.143*B3-0.1516*B4-2.6676*IF(B5>=55,1,0)),4))))Apply the logistic function
The second formula evaluates 100 / (1 + EXP(−b)) using the unrounded linear predictor b, then rounds for display. A displayed 0.00% or 100.00% can be a rounding result and does not mean an outcome is certain.
This is a population-model arithmetic exercise only. It is not for diagnosing injuries, assigning care, or forecasting what will happen to an individual.
=IF(COUNT(B3:B5)<>3,"Enter all numeric inputs",IF(AND(UPPER(TRIM(B2&""))<>"BLUNT",UPPER(TRIM(B2&""))<>"PENETRATING"),"Choose BLUNT or PENETRATING",IF(OR(B3<0,B3>7.8408,B4<1,B4>75,INT(B4)<>B4,B5<15,INT(B5)<>B5),"Check input ranges",ROUND(100/(1+EXP(-(IF(UPPER(TRIM(B2&""))="BLUNT",-1.247+0.9544*B3-0.0768*B4-1.9052*IF(B5>=55,1,0),-0.6029+1.143*B3-0.1516*B4-2.6676*IF(B5>=55,1,0))))),2))))Try the example
Use the original values or change them to test your understanding.