Huq et al reported a model for predicting where a hospitalized patient with a brain tumor will be discharged. The authors are from Johns Hopkins University.
Patient selection: hospitalized patient with brain tumor
Outcome: discharge location
Parameters:
(1) age of the patient in years
(2) race
(3) marital status
(4) insurance type
(5) admission source (residence prior to admission)
(6) 5-factor modified frailty index (mFI-5) of Huq et al
(7) Karnofsky Performance Status (KPS)
Parameter |
Finding |
OR |
ln OR |
age |
<=60 |
1 |
0.000 |
> 60 |
2.02 |
0.703 |
|
race |
white |
1 |
0.000 |
black |
1.73 |
0.548 |
|
asian |
2.05 |
0.718 |
|
other |
1.83 |
0.604 |
|
marital |
married |
1 |
0.000 |
not |
1.48 |
0.392 |
|
insurance |
private |
1 |
0.000 |
Medicare |
1.46 |
0.378 |
|
Medicaid |
1.9 |
0.642 |
|
admission source |
home |
1 |
0.000 |
health care |
2.3 |
0.833 |
|
frailty |
0 |
1 |
0.000 |
1 |
1.39 |
0.329 |
|
2+ |
1.61 |
0.476 |
|
KPS |
100 |
1 |
0.000 |
90 |
1.15 |
0.140 |
|
80 |
2.8 |
1.030 |
|
70 |
5.52 |
1.708 |
|
60 |
7.49 |
2.014 |
|
50 |
13.72 |
2.619 |
|
< 50 |
54.47 |
3.998 |
value of X =
= SUM(points for all of the parameters) - 3.75
probability of discharge to a site other than home =
= 1 / (1 + EXP((-1) * X))
Performance:
• The area under the ROC curve is 0.82.
Specialty: Hematology Oncology