Skip to main content

Table 3 Model Performance of selected HIT6 and CH-QLQ to EQ-5D and SF-6D models

From: Mapping between headache specific and generic preference-based health-related quality of life measures

   

Predicted EQ-5D Values

  

Abs Diff

Abs Diff

Model

Adj R2

AIC

Mean (SD)

Min

P.25

Median

P.75

Max

MSE

MAE

 < 0.10 (%)

 < 0.25 (%)

EQ-5D Models

 Observed

  

0.6288 (0.2563)

-0.594

0.531

0.7125

0.7680

1

-

-

-

-

 HIT6 to EQ-5D—CLAD (1)

0.0929

 

0.6429 (0.0984)

0.3962

0.5803

0.6540

0.6908

1.0592

0.0550

0.1720

39.51%

78.17%

 CH-QLQ to EQ-5D—CLAD (3)

0.2933

 

0.6229 (0.1938)

-0.5704

0.5467

0.6842

0.7582

0.9274

0.0583

0.1702

45.61%

76.83%

SF-6D Models

 Observed

  

0.6056 (0.1163)

0.3450

0.5350

0.6000

0.6600

0.9220

-

-

-

-

 HIT6 to SF-6D—CLAD (2)

0.2375

 

0.5926 (0.0649)

0.4292

0.5535

0.5939

0.6297

0.8633

0.0102

0.0751

72.32%

97.25%

 CH-QLQ to SF-6D—OLS (2)

0.5267

-691.3988

0.5950 (0.0775)

0.4117

0.5393

0.6009

0.6547

0.7544

0.0086

0.0700

75.22%

98.25%

  1. Dependent variable for OLS and CLAD was EQ-5D utility score
  2. Independent variable(s): Model (1) HIT6/CH-QLQ score, Model (2) HIT6/CH-QLQ score, age and gender, Model (3) HIT6/CH-QLQ score, age gender, squared terms, interactions
Back to article page

BMC Medical Research Methodology

ISSN: 1471-2288

Contact us