Skip to main content

Table 3 ROC analyses for determining AUC (95% CI) and for estimating cut-off values for ‘success

From: Follow-up score, change score or percentage change score for determining clinical important outcome following surgery? An observational study from the Norwegian registry for Spine surgery evaluating patient reported outcome measures in lumbar spinal stenosis and lumbar degenerative spondylolisthesis

  

ODI

   

EQ-5D

   

NRS

Leg pain

  

NRS

Back pain

 
 

AUC

(95% CI)

Max corr. Class

Cut-off

AUC

(95% CI)

Max corr. Class

Cut-off

AUC

 

Max corr. Class.

Cut-off

AUC

 

Max corr.

class.

Cut-off

Spinal stenosis

Follow-up score

(points)

0.90

(0.89–0.91)

82%

≤24

0.87

(0.85–0.88)

78%

0.692

0.87

(0.86–0.89)

81%

≤4

0.89

(0.87–0.90)

82%

≤4

Change score

(points)

0.86

(0.84–0.87)

78%

≥13

0.76

(0.74–0.78)

73%

0.105

0.83

(0.82–0.85)

77%

≥3

0.82

(0.81–0.84)

75%

≥2

Percentage change (%)

0.91

(0.90–0.92)

83%

≥30

    

0.86

(0.85–0.88)

81%

≥40

0.86

(0.84–0.87)

79%

≥33

Degenerative

Spondylolisthesis

Follow-up score

(points)

0.92

(0.89–0.94)

85%

≤24

0.92

(0.89–0.94)

84%

≥0.692

0.86

(0.82–0.89)

79%

≤3

0.90

(0.88–0.93)

83%

≤4

Change score

(points)

0.86

(0.82–0.89)

80%

≥13

0.76

(0.76–0.81)

76%

≥0.105

0.81

(0.77–0.91)

76%

≥3

0.84

(0.81–0.88)

79%

≥3

Percentage

change (%)

0.92

(0.90–0.94)

85%

≥30

    

0.84

(0.80–0.87)

78%

≥40

0.88

(0.85–0.91)

80%

≥33

  1. The area under the curve (AUC) with 95% confidence interval (CI) describes a candidate score’s ability to classify patients who replied ‘completely recovered’ or ‘much improved’ on the GPE scale into ‘success’ and those replied ‘slightly improved’, ‘unchanged’, ‘slightly worse’, ‘much worse’, and ‘worse than ever’ into ‘non-success’ at 12 month follow-up. The larger the AUC, the better the accuracy of the score [range from 0.5 (no ability) to 1.0 (perfect ability)]. A cut-off corresponds to the threshold that gave rise to the maximum percentage of patients correctly classified (max corr. Class.) into ‘success’ and ‘non-success’. Results are given for ODI, EQ-5D, NRS leg pain, and NRS back pain for spinal stenosis and for degenerative spondylolisthesis. Because the EQ-5D questionnaire values ranged from −0.6 to 1.0 on a categorical scale, it was not mathematically possible to evaluate the percent change score