Skip to main content

Table 2 Results of penalized Cox models in determining the factors affecting time to relapse in patients with pulmonary tuberculosis

From: Factors associated with time to relapse in pulmonary tuberculosis patients using penalized Cox models

Variable

LASSO

MCP

SCAD

Variable

LASSO

MCP

SCAD

Beta

Beta

Gender

Injection drug addiction

 Female

Reference

 No

Reference

 Male

0

0

0

 Yes

0

0

0

Age at diagnosis (years)

History of contact with a patient with TB

 0–35

Reference

 No

Reference

 More than 35 years

0.33

1.02

0.99

 Yes

0

0

0

Marital status

HIV infection

 Single

Reference

 No

Reference

 Married

0.24

2.14

2.14

 Yes

0

0

0

Education level

Diabetes

 Illiterate

Reference

 No

Reference

  

 High School

0

0

0

 Yes

0

0

0

 University education

0

0

0

    

Residence

Chronic renal failure

 Urban

Reference

 No

Reference

 Rural

0

0

0

 Yes

0

1.68

0

BMI

Smear test results before treatment

 < 18.5

Reference

 1-9Basil

Reference

 18.5–24.99

0

0

0

 + 1

0

− 0.43

− 0.44

 25–29.99

0

0

0.478

 + 2

0

− 0.04

− 0.074

 > 30

0

0

0

 + 3

0

0

0

Currently incarcerated

Smear test results after 2 months

 No

Reference

 Negative

Reference

 Yes

0

0

0

 1-9Basil

0

0

0

    

 + 1

0

0

0

Prison history

 + 2

0

0

0

 No

Reference

 + 3

0

0

0

 Yes

0

0

0

    

Duration of treatment (days)

0

0

0

Chest X-ray

Duration of delay in diagnosis (days)

0

0

0

 Non-suggestive

Reference

Duration of delay in treatment (days)

0

0

0

 Less suggestive

0

0

0

    

 More suggestive

0

0

0

 

LASSO

MCP

SCAD

    

AIC

578.682

570.416

571.778

 

LASSO vs. MCP

LASSO vs. SCAD

MCP vs. SCAD

BIC

581.530

578.129

579.132

LRT

3.71

2.4

1.31

    

p-value

0.29

0.50

0.99

  1. AIC: Akaike Information Criterion; BIC: Bayesian information criterion; LRT: Likelihood Ratio