Airaksinen et al reported a model for predicting work disability following injury. The authors are from the Finnish Institute of Occupational Health, University of Helsinki, University of Turku, Harvard University and University College London.
Patient selection: injured worker
Outcome: work disability in 10 years
Parameters:
(1) age of the patient
(2) self-rated health
(3) number of sickness absences in previous year
(4) socioeconomic position
(5) number of chronic diseases
(6) difficulty falling asleep
(7) body mass index (BMI) in kg per square meter
(8) smoking
Parameter
Finding
Points
age of the patient
< 35 years
0
35 to 39
7.5
40 to 44
14
45 to 49
28.6
50 to 54
41.7
>= 55
50
self-rated health
good
0
rather good
7.5
moderate
50
rather poor
102
poor
137.5
number of work absences
0
0
1
14
2
24
3
36.7
socioeconomic position
manager
0
senior specialist
1.4
specialist
10.2
office worker
10.24
service worker
17.6
process worker
17
other
21.4
number of chronic diseases
0
0
1
7.2
2
14.3
3
17.2
BMI
< 18.5
5.5
18.5 to 24.9
0
25 to 29.9
2.1
>= 30
5.6
smoking
no
0
yes
5.3
total score =
= SUM(points for all of the parameters)
Interpretation:
• minimum score: 0
• maximum score: 192.7
value of X =
= (0.04445 * (score)) - 5.037
probability of work disability at 10 years =
= 1 / (1 + EXP((-1) * X))
Performance:
• The area under the ROC curve is 0.84.
To read more or access our algorithms and calculators, please log in or register.
