Altman used recursive partitioning in the diagnosis of idiopathic osteoarthritis of the knee. The diagnosis can be made based on clinical data alone. The author is from the University of Miami and Miami Veterans Administration Medical Center.
NOTE: The recursive partitioning is similar to CART (classification and regression tree) analysis.
Patient selection:
(1) presence of knee pain
(2) absence of a condition predisposing to degenerative arthritis (see Table I, page 33)
Parameters:
(1) crepitus
(2) bony enlargement
(3) duration of morning stiffness
(4) age
Crepitus |
Bony Enlargement |
Morning Stiffness |
Age in Years |
Group |
N |
N |
|
|
A |
N |
Y |
|
|
B |
Y |
|
<= 30 minutes |
<= 37 years |
C |
Y |
|
<= 30 minutes |
>= 38 years |
D |
Y |
N |
> 30 minutes |
|
E |
Y |
Y |
> 30 minutes |
|
F |
Group |
Diagnosis |
---|---|
A |
no osteoarthritis |
B |
osteoarthritis |
C |
no osteoarthritis |
D |
osteoarthritis |
E |
no osteoarthritis |
F |
osteoarthritis |
Performance:
• The overall sensitivity is 89% and the specificity is 88%.
Age can be used as a surrogate marker for crepitus. In that case the above table can be rewritten as:
Age in Years |
Bony Enlargement |
Morning Stiffness |
Group |
<= 38 years |
N |
|
A |
<= 38 years |
Y |
|
B |
>= 39 years |
|
<= 30 minutes |
D |
>= 39 years |
N |
> 30 minutes |
E |
>= 39 years |
Y |
> 30 minutes |
F |
Purpose: To evaluate a patient for idiopathic osteoarthritis of the knee using the recursive partitioning criteria of Altman with clinical data.
Specialty: Surgery, orthopedic, Immunology/Rheumatology
Objective: clinical diagnosis, including family history for genetics, criteria for diagnosis
ICD-10: M15, M19,