Falconer et al reported the Adverse Inpatient Medication Event Model (AIME) to identify a hospital inpatient at increased risk for an adverse drug event. The authors are from the University of Queensland and Princess Alexandra Hospital in Brisbane, Australia.
Patient selection: hospital inpatient
Parameters:
(1) length of stay in days
(2) admission in previous 12 months
(3) INR
(4) serum sodium in mmol/L
(5) number of medications
(6) antipsychotic medication
(7) antiarrhythmic medication
(8) immunosuppressive medication
(9) history of medication allergy
(10) venous or arterial thrombosis or embolism
|
Parameter
|
Finding
|
Points
|
|
length of stay
|
|
0.35 LN(days)
|
|
admission in past year
|
no
|
0
|
|
|
yes
|
0.51
|
|
INR
|
<= 3
|
0
|
|
|
> 3
|
1.13
|
|
serum sodum
|
<= 125 mmol/L
|
1.24
|
|
|
> 125 mmol/L
|
0
|
|
number of medications
|
< 8
|
0
|
|
|
>= 8
|
0.5
|
|
antipsychotic drug
|
no
|
0
|
|
|
yes
|
0.4
|
|
antiarrhythmic drug
|
no
|
0
|
|
|
yes
|
0.45
|
|
immunosuppressant
|
no
|
0
|
|
|
yes
|
0.87
|
|
medication allergy
|
no
|
0
|
|
|
yes
|
0.37
|
|
thrombosis or embolism
|
no
|
0
|
|
|
yes
|
0.86
|
value of X =
= SUM(points for all of the parameters) - 4.56
probability of adverse medication event =
= 1 / (1 + EXP((-1) * X))
Performance:
• The area under the ROC curve is 0.70.
