Skip to main content

Table 4 Proportional Hazard model fit statistics with alternative income terms, healthy restricted population, age 18–64, NHANES mortality follow-up cohort, 1988–2001

From: The non-linear risk of mortality by income level in a healthy population: US National Health and Nutrition Examination Survey mortality follow-up cohort, 1988–2001

 

Women (n = 6307)

Men (n = 5426)

All-cause (women = 169, men = 181)

Likelihood ratio test

df

p-value

Likelihood ratio test

df

p-value

   No income

26

12

0.01

67

12

1 × 10-9

   Linear income

29

13

0.007

69

13

1 × 10-9

   Log income

37

13

0.0005

70

13

9 × 10-10

   Non-linear income

53

15

0.000003

80

15

8 × 10–11

heart disease (women = 30, men = 46)

      

   No income

8.6

12

0.7

27

11

0.004

   Linear income

12

13

0.5

36

12

0.0003

   Log income

10

13

0.7

30

12

0.003

   Non-linear income

14

15

0.5

39

14

0.0003

Cancer (women = 63, men = 50)

      

   No income

20

11

0.05

36

11

0.0002

   Linear income

21

12

0.05

36

12

0.0003

   Log income

20

12

0.07

37

12

0.0003

   Non-linear income

27

14

0.02

38

13

0.0003

Injury (women = 11, men = 32)

      

   No income

11

11

0.5

19

12

0.09

   Linear income

11

12

0.5

19

13

0.1

   Log income

11

12

0.5

20

13

0.09

   Non-linear income

12

13

0.5

23

14

0.06

  1. All models adjust for age (continuous), marital status (2 category), race/ethnicity (4 category), occupation (5 category) and education (as smoothed continuous term).
  2. Note: Degrees of freedom between models for men and women may differ because when no events occurred within a particular strata, the covariate was omitted from all models within that gender and cause of death (this occurred for cancer mortality among women of the "other" race/ethnicity category, injury mortality for women of occupational category 3, heart disease mortality for men of the "other" race/ethnicity category, and cancer mortality for men of occupational category 5).