Table 3

Respondent sociodemographic variables and associations with CAM use, 2002 NHIS (weighted)

Characteristic

% of Sample

% of group using CAM in last 12 months

UORa

99% CIb

AORc

99% CIb


Educational attainment

Less than high school

16.6

20.7

1.00

.

1.00

.

High school graduate/GED

29.9

30.2

1.65

1.45–1.88

1.37

1.19–1.57

Some college/Associate's degree

29.1

40.4

2.59

2.29–2.93

1.85

1.62–2.12

Bachelor's degree

16.3

47.9

3.51

3.06–4.03

2.40

2.05–2.81

Master's, Doctorate, or Professional degree

8.2

52.0

4.14

3.49–4.91

2.79

2.27–3.42

Poverty statusd

Below poverty level

11.2

27.0

1.00

.

1.00

.

100%<=ratio<200%

17.0

28.9

1.10

0.94–1.28

1.06

0.90–1.25

200% <=ratio<300%

17.9

33.1

1.33

1.15–1.55

1.14

0.96–1.36

300%<=ratio<400%

15.4

36.8

1.57

1.35–1.84

1.18

0.98–1.43

400%<=ratio<500%

12.8

39.3

1.74

1.49–2.05

1.20

0.98–1.47

500%+

25.7

44.8

2.19

1.89–2.54

1.30

1.07–1.56

Sex

Male

48.0

31.5

1.00

.

1.00

.

Female

52.0

40.4

1.48

1.37–1.59

1.55

1.42–1.70

Age

18–44

52.5

36.6

1.62

1.46–1.80

1.73

1.47–2.04

45–64

31.4

40.4

1.90

1.71–2.12

1.64

1.41–1.90

65+

16.1

26.3

1.00

.

1.00

.

Employment status

Not employed

35.9

32.1

1.00

.

1.00

.

Private sector

47.4

36.4

1.21

1.12–1.32

1.07

0.97–1.19

Government

10.2

44.5

1.70

1.49–1.93

1.12

0.95–1.32

Self-employed/family business

6.5

43.5

1.62

1.40–1.89

1.41

1.19–1.66

Race and ethnicitye

Hispanic

11.0

27.9

0.63

0.56–0.70

1.09

0.95–1.26

Non-Hispanic white

73.3

38.2

1.00

.

1.00

.

Non-Hispanic black

11.4

28.0

0.63

0.56–0.71

0.90

0.79–1.03

Non-Hispanic American Indian or Alaska Native

0.6

42.4

1.19

0.70–2.03

1.37

0.80–2.37

Asian or Other Pacific Islander

3.7

43.4

1.24

1.01–1.53

1.56

1.19–2.03

Region of residence

Northeast

19.3

37.0

1.31

1.17–1.48

1.12

1.00–1.24

Midwest

24.4

37.9

1.37

1.22–1.53

1.20

1.07–1.35

South

37.0

30.9

1.00

.

1.00

.

West

19.3

43.1

1.70

1.52–1.90

1.58

1.39–1.79


a Unadjusted odds ratios. Sets of bivariate logistic regressions were performed to determine associations between each independent and control variable and CAM use.

b More conservative 99% confidence intervals were used because of the enhanced statistical power generated by the large sample size.

c Adjusted odds ratios. Each variable is adjusted for all other variables in Tables 1, 2, and 3.

d "Poverty status" is based on a multiply-imputed total family income variable.

e Non-Hispanic, multiple race sample adults were dropped from the analysis due to a small sample size (n = 45).

Nahin et al. BMC Public Health 2007 7:217   doi:10.1186/1471-2458-7-217

Open Data