Description

Tseng et al reported models for predicting 30-day outcomes following surgery for a patient with disseminated cancer. One model predicts morbidity and the second mortality. The authors are from the University of California at Davis.


Patient selection: surgical intervention on a patient with disseminated cancer

 

Parameters:

(1) DNR status

(2) weight loss

(3) dyspnea

(4) functional dependence

(5) ascites

(6) chronic steroid use

(7) active sepsis

(8) serum creatinine in mg/dL

(9) serum albumin in g/dL

(10) blood WBC count in 10^9/L

(11) hematocrit in percent

(12) surgical acuity

(13) procedure type

(14) age

 

points for serum creatinine =

= 5.7625 * (creatinine)

 

points for serum albumin =

= 122.22 - (22.22 * (albumin))

 

points for WBC count =

= (1.705 * (WBCC))

 

points for hematocrit =

= 39.3625 - (0.78725 * (hematocrit))

 

points for age =

= (1.1075 * (age)) - 11.075

 

Parameter

Finding

Points

DNR status

no

0

 

yes

33.8

weight loss

<= 10%

0

 

> 10%

18.8

dyspnea

no

0

 

moderate

10.3

 

at rest

29.2

functional dependence

none

0

 

partial

41.8

 

total

56.9

ascites

no

0

 

yes

38.9

chronic steroid use

no

0

 

yes

23.6

active sepsis

no

0

 

yes

25.2

surgical acuity

elective

0

 

urgent

16.1

 

emergency

32.4

procedure type

skin or soft tissue

0

 

vascular

29.5

 

biopsy

21.5

 

gallbladder, appendix, lysis of adhesions

33.2

 

gastrointestinal resection

25.8

 

multivisceral resection

19.1

 

other

36.6

 

total score =

= SUM(points for all of the parameters)

 

Interpretation:

• minimum score: 0

• maximum sore: 636.8

 

value of X =

= (0.02065 * (score)) - 7.037

 

probability of 30-day mortality =

= 1 / (1 + EXP((-1) * X))

 

Performance:

• The concordance index is 0.86.


To read more or access our algorithms and calculators, please log in or register.