Method: Multivariate logistic regression on data from 5,820 patients randomly selected from the 11,630 patients in the Cystic Fibrosis Foundation Patient Registry (CFFPR).
Parameters:
(1) age in years
(2) gender
(3) FEV1 as percent of predicted
(4) Z score for the weight-for-age, based on patients enrolled in the Cystic Fibrosis Foundation Patient Registry (not normal children)
(5) pancreatic sufficiency (not requiring pancreatic enzyme supplementation)
(6) diabetes mellitus
(7) Staphylococcus aureus infection
(8) Burkholderia cepacia infection
(9) number of acute pulmonary exacerbations requiring treatment during year
(10) (number of acute exacerbations) * (Burkholderia cepacia)
|
Parameter
|
Finding
|
Points
|
|
age in years
|
|
(age in years)
|
|
gender
|
male
|
0
|
|
|
female
|
1
|
|
FEV1 as percent of predicted
|
|
(percent as whole number)
|
|
weight-for-age Z score
|
|
(Z score)
|
|
pancreatic sufficiency
|
no
|
0
|
|
|
yes
|
1
|
|
insulin-requiring diabetes mellitus
|
no
|
0
|
|
|
yes
|
1
|
|
infection with S. aureus
|
no
|
0
|
|
|
yes
|
1
|
|
infection with B. cepacia
|
no
|
0
|
|
|
yes
|
1
|
|
number of acute exacerbations during year
|
|
MIN(5, number)
|
|
(acute exacerbations) * (infection with B. cepacia)
|
|
0 to 5
|
where:
• The assignments of 0 or 1 to "no" or "yes" responses is assumed, as I could not find it explicitly stated in the text.
• Z score = ((patient weight) – (mean weight for age)) / (SD of weight for age)
• I interpret interaction term "(number of acute exacerbations) * (Burkholderia cepacia)" to have maximum value of 5 (5 as maximum value for acute exacerbations, B. cepacia as 0 or 1).
X =
= (-0.028 * (age in years)) - (0.23 * (points for gender)) + (0.038 * (percent FEV1)) + (0.40 * (Z score)) + (0.45 * (pancreatic sufficiency)) - (0.49 * (diabetes mellitus)) + (0.21 * (infection with S. aureus)) - (1.82 * (infection with B. cepacia)) - (0.46 * (acute exacerbations)) + (0.40 * (acute exacerbations) * (infection with B. cepacia)) + 1.93
probability of 5 year survival =
= EXP(X) / (1+ EXP(X))
