Television Watching, Diet Quality, and Physical Activity and Diabetes among Three Ethnicities in the United States

Table 3

Effect of diet and physical activity on likelihood of viewing more than 4 hours/day of TV.

Independents	B	S.E	df	P	OR	95% C.I. OR
						Lower	Upper
Quartiles HEI	—	—	3	0.014	—	—	—
Q1	0.80	0.29	1	0.006	2.22	1.25	3.92
Q2	0.25	0.29	1	0.337	1.29	0.74	2.25
Q3	0.69	0.28	1	0.015	1.98	1.14	3.45
Quartiles physical activity (MET/wk)	—	—	3	<0.001	—	—	—
Q1	1.18	0.28	1	<0.001	3.25	1.86	5.68
Q2	0.80	0.29	1	0.005	2.23	1.27	3.91
Q3	0.39	0.30	1	0.195	1.47	0.82	2.64
Ethnicity	—	—	2	<0.001	—	—	—
AA	1.66	0.27	1	<0.001	5.27	3.09	8.97
HA	−0.42	0.27	1	0.125	0.66	0.38	1.12
Diabetes (yes)	0.45	0.19	1	0.020	1.56	1.08	2.27

Abbreviations. BMI: body mass index; B: coefficient; SE: standard error; Q: quartile; HEI: Healthy Eating Index; AA: African American; HA: Haitian American.
Notes. Cuban Americans were the reference. The 4th quartile was the reference for HEI, physical activity, and BMI. Physical activity was measured in metabolic equivalents (MET) per week. MET = 1 kcal·kg⁻¹·h⁻¹. One MET is the equivalent of sitting quietly. Television watching was reported to the nearest 0.25 hours. The dependent variable is ≥ 4.25 hours of TV versus ≤ 4.00 hours of TV. Education and currently married were tested. Education remained in the final model.
The model was adjusted for age (𝑃=0.058), gender (𝑃=0.705), BMI quartile (𝑃=0.207), currently smoking (𝑃=0.221), and education (𝑃=0.019). Compared to persons with at least college, those with less than a high school education were 1.54 times more likely to watch more than 4 hours of TV/day.
Model parameters. χ ² (17) =  147 (𝑁=827), 𝑃<0.001. Nagelkerke R-squared = 0.247 with 79.6% of the cases classified correctly.