Skip to main content

Table 2 Indices of model fit from latent class growth analyses of chronic pain in arthritis and cancer

From: General and disease-specific pain trajectories as predictors of social and political outcomes in arthritis and cancer

k

AIC

BIC

Entropy

VLMR-LRT p

Whole sample

 1

177,684.561

177,839.979

–

–

 2

133,343.915

133,476.949

0.871

< 0.001

 3

130,255.155

130,454.705

0.848

< 0.001

 4

126,764.571

127,030.638

0.851

< 0.001

 + Quadratic

124,206.071

124,501.701

0.859

 

 + Cubic

123,577.681

123,902.874

0.869

 

 5

125,988.308

126,320.892

0.901

< 0.001

Arthritis

 1

14,198.301

14,298.893

–

–

 2

11,316.565

11,407.577

0.702

< 0.001

 3

11,200.776

11,339.689

0.702

0.008

 4

11,105.289

11,292.103

0.757

0.001

 + Quadratic

10,942.055

11,148.029

0.841

 

 + Cubic

10,914.137

11,139.272

0.850

 

 5

11,084.221

11,318.936

0.778

0.29

Cancer

 1

6298.329

6384.856

–

–

 2

4662.729

4736.895

0.904

< 0.001

 3

4474.875

4586.123

0.998

0.052

 + Quadratic

4420.884

4544.492

0.998

 

 + Cubic

4390.708

4526.677

0.998

 

 4

4198.800

4347.131

0.998

0.37

 5

4128.545

4313.958

0.947

0.37

  1. In the whole sample and cancer subsample, the five class linear models failed to converge
  2. AIC Akaike information criterion, BIC Bayesian information criterion, VLMR-LRT Vuong-Lo-Mendell-Rubin likelihood ratio test
  3. Statistics highlighted in bold are the best fitting for a particular index
  4. Standard deviations are reported in brackets for continuous variables (BMI, age, GHQ-12 and social engagement measures)