Dreyer et al reported a model for predicting 1-year readmission for a younger adult hospitalized with an acute myocardial infarction. The authors are from multiple institutions in the United States, Canada, and Italy.
Patient selection: adult from 18 to 55 years of age
Parameters:
(1) PHQ-9, from 0 to 27
(2) physical health on admission from SF-12, from 6 to 20
(3) percent obstruction in coronary artery disease
(4) diabetes
(5) sex
(6) GRACE score
(7) income (low if <= $30,000 per year)
(8) marital status
(9) hospital length of stay in days
|
Parameter
|
Finding
|
Points
|
|
PHQ-9
|
|
0.0313 * (PHQ9)
|
|
physical health from SF-12
|
|
-0.0210 * (SF12)
|
|
percent obstruction CAD
|
< 50%
|
0
|
|
|
>= 50%
|
0.333
|
|
diabetes
|
no
|
0
|
|
|
yes
|
0.2931
|
|
sex
|
male
|
0
|
|
|
female
|
0.3123
|
|
GRACE score
|
|
-0.00707 * (GRACE)
|
|
income
|
low
|
0.2053
|
|
|
other
|
0
|
|
marital status
|
married or lives with other
|
-0.1208
|
|
|
other
|
0
|
|
hospital length of stay
|
|
0.0348 * (days)
|
where:
• There are 3 possible GRACE scores. The most recent is the updated score for predicting in-hospital mortality ranges from 1 to 375. The first GRACE score ranges from 1 to 263.
• Being employed is list as being a favorable finding, but this was not included in the model.
value of X =
= SUM(points for all of the parameters0 - 0.5853
percent readmission at 1-year =
= 1 / ( 1+ EXP((-1) * X))
