Tang et al reported a model for predicting polypharmacy in an elderly adult living in the community in China. The authors are from Fudan University in Shanghai.
Pateint selection: community-dwelling adult >= 65 years of age
Parameters:
(1) age in years, from 65 to 105
(2) living district
(3) preference for medical institution
(4) number of visits to tertiary hospital
(5) number of visits to secondary hospital
(6) number of visits to community healthcare center
(7) number of diagnoses
(8) main types of disease
points for age =
= (1.359 * (age)) - 88.368
points for number of diagnoses =
= (4.125 * (number)) - 0.225
|
Parameter
|
Finding
|
Points
|
|
living district
|
countryside
|
14
|
|
|
suburban
|
21.6
|
|
|
central city
|
25.4
|
|
preference
|
community
|
14
|
|
|
secondary
|
33.8
|
|
|
tertiary
|
39
|
|
number tertiary
|
<= 5
|
14
|
|
|
6 to 10
|
63.6
|
|
|
>= 11
|
84.3
|
|
number secondary
|
<= 5
|
14
|
|
|
6 to 10
|
63.3
|
|
|
>= 11
|
100
|
|
number community
|
<= 5
|
14
|
|
|
6 to 10
|
43.4
|
|
|
>= 11
|
96.3
|
|
disease type
|
other
|
11.8
|
|
|
musculoskeletal, connective tissue
|
14
|
|
|
digestive
|
24.4
|
|
|
endocrine, metabolic, nutrition
|
25.3
|
|
|
circulatory
|
40.4
|
|
|
respiratory
|
50.8
|
total score =
= SUM(points for all of the parameters)
Interpretation:
• minimum score: 85.7
• maximum score: 503.6
value of X =
= (0.01881 * (score)) - 3.736
probability of polypharmacy =
= 1 / (1 + EXP((-1) * X))
Performance:
• The area under the ROC curve is 0.78.
