Skip to main content

Table 5 Subgroup analysis of prediction performance of various models in training set

From: Development and validation of non-invasive prediction models for migraine in Chinese adults

Subgroup

Model

N

AUC(95%CI)

Sensitivity

Specificity

Youden index

Cut off value

Gender

 Male

Model1

27

0.80(0.63–0.97)

0.647

0.800

0.447

0.744

Model2

27

0.80(0.63–0.97)

0.765

0.800

0.565

0.507

Model3

27

0.85(0.70–0.99)

0.882

0.800

0.682

0.512

Model4

27

0.91(0.78–1.00)

1

0.800

0.800

0.365

Model5

27

0.91(0.78–1.00)

1

0.700

0.700

0.412

Model6

27

0.92(0.80–1.00)a

1

0.800

0.800

0.333

Model7

27

0.94(0.84–1.00)abc

1

0.800

0.800

0.377

Model8

27

0.95(0.87–1.00)abc

0.765

1

0.765

0.793

 Female

Model1

161

0.83(0.77–0.90)

0.838

0.700

0.538

0.621

Model2

161

0.86(0.81–0.92)

0.802

0.840

0.642

0.640

Model3

161

0.88(0.83–0.93)a

0.811

0.820

0.631

0.653

Model4

161

0.88(0.83–0.93)a

0.883

0.760

0.643

0.584

Model5

161

0.89(0.84–0.94)a

0.829

0.840

0.669

0.698

Model6

161

0.90(0.86–0.95)ab

0.721

0.960

0.681

0.830

Model7

161

0.91(0.86–0.95)ab

0.775

0.900

0.675

0.762

Model8

161

0.92(0.88–0.96)abcd

0.928

0.800

0.728

0.494

Age

 > 35 years

Model1

89

0.85(0.76–0.93)

0.652

0.870

0.521

0.837

Model2

89

0.89(0.82–0.96)

0.818

0.870

0.688

0.731

Model3

89

0.90(0.84–0.96)

0.712

1

0.712

0.852

Model4

89

0.92(0.86–0.98)a

0.955

0.739

0.694

0.569

Model5

89

0.91(0.84–0.97)

0.773

0.913

0.686

0.808

Model6

89

0.93(0.87–0.99)a

0.788

0.957

0.744

0.841

Model7

89

0.93(0.87–0.99)a

0.879

0.913

0.792

0.780

Model8

89

0.95(0.91–1.00)a

0.848

0.957

0.805

0.815

 ≤ 35 years

Model1

99

0.81(0.72–0.90)

0.742

0.730

0.472

0.621

Model2

99

0.83(0.75–0.91)

0.694

0.865

0.558

0.640

Model3

99

0.85(0.77–0.93)

0.855

0.730

0.585

0.512

Model4

99

0.84(0.77–0.92)

0.774

0.784

0.558

0.571

Model5

99

0.88(0.82–0.95)ab

0.790

0.865

0.655

0.622

Model6

99

0.88(0.81–0.95)ab

0.839

0.757

0.595

0.541

Model7

99

0.88(0.82–0.95)abd

0.935

0.703

0.638

0.377

Model8

99

0.89(0.83–0.96)abcd

0.887

0.784

0.671

0.494

  1. aindicated that after Delong test, the AUC was significantly better than model 1
  2. bindicated that after Delong test, the AUC was significantly better than model 2
  3. cindicated that after Delong test, the AUC was significantly better than model 3
  4. dindicated that after Delong test, the AUC was significantly better than model 4
  5. AUC Area under the receiver operating characteristic curve