Poor Nutritional Habits: A Modifiable Predecessor of Chronic Illness? A North Carolina Family Medicine Research Network (NC-FM-RN) Study

Article Figures & Data

Download figure
Open in new tab
Figure 1.
Racial and ethnic differences in the percentage of respondents reporting selected nutritional habits. P values are based on logistic regression models with given food habit as the dependent variable and race/ethnicity as the independent variable, adjusted for clinic site, age, sex, smoking status, alcohol consumption, physical activity, self-rated health, and BMI. N ranges from 1323 to 1357, depending on missing data for the food habit. Percentages are not adjusted. ∗, P < .05; ∗∗, P < .01; and ∗∗∗, P < .001 for comparison with white, non-Hispanic respondents.
Download figure
Open in new tab
Figure 2.
Comparison of selected nutritional habits of diabetics with those of nondiabetics who are at high risk for the disease. P values are based on logistic regression models with given food habit as the dependent variable and diabetes group as the independent variable, adjusted for clinic site, age, sex, race, smoking status, alcohol consumption, and physical activity. N ranges from 801 to 822, depending on missing data for the food habit. Percentages are unadjusted.

Table 1.

Demographic and Health Status of Study Participants (N = 1714)*

	Number	Percentage	Mean	SD	Range
Age (years)			47.4	16.9	18.0, 94.9
Gender
Male	604	35.3
Female	1107	64.7
Education
8th grade or less	200	11.8
High school, no diploma	205	12.1
High school graduate or GED	506	30.0
Some college, no degree	387	22.9
Associate’s degree	144	8.5
Bachelor’s degree	151	8.9
Postgraduate school or degree	95	5.6
Married/living with partner	1052	62.8
Race/ethnicity
African American	275	16.8
White, non-Hispanic	982	60.1
American Indian	124	7.6
Hispanic/Latino	242	14.8
Other	12	0.7
Practice location
A	441	25.7
B	390	22.8
C	621	36.2
D	262	15.3
Current health problems
Heart disease	210	12.7
High blood pressure	700	41.6
Lung disease	97	5.9
Stroke or mini-stroke	72	4.4
Depression	538	32.4
Chronic back pain	478	28.8
Cancer	76	4.6
Diabetes	291	17.5
Osteoarthritis	237	14.4
Rheumatoid arthritis	150	9.1
Fibromyalgia	105	6.4
Self-reported health
Excellent	142	8.3
Very good	392	23.0
Good	575	33.7
Fair	458	26.8
Poor	140	8.2
Drinks alcohol	536	31.7
Engaged in physical activity in past week	981	59.5
Physically active 5 days/week, >30 minutes/day	642	39.0
Current smoker	401	23.6
Body mass index (BMI)			29.6	7.1	15.0, 64.6
Normal or underweight (BMI <25)	414^†	27.3
Overweight (BMI 25–30)	484	31.9
Obese (BMI >30)	621	40.9

* The sample N includes all those who answered the nutritional habit questions. The sample size for individual characteristics varies from 1519 to 1711 because of missing data.
† Includes 30 people classified as underweight (BMI <18.5).

Table 2.

Self-Reported Nutritional Habits of Study Participants (N = 1714)*

Item and Response Choices from Nutritional Habits Survey	Number	Percentage
Number of times fast food eaten per week	239	14.6
0
1 to 2	877	53.4
3 or more	526	32.0
Number of glasses of soda or sweet tea consumed per day	549	32.7
0
1 to 2	639	38.0
3 or more	492	29.3
Number of times high-fat snacks consumed per week	836	50.4
0 to 1
2	438	26.4
3 or more	386	23.3
Number of times desserts or sweets consumed per week	610	36.4
0 to 1
2	433	25.9
3 or more	631	37.7
Amount of margarine, butter, meat fat consumed	729	42.7
None or very little
Some	787	46.1
A lot	192	11.2
Number of servings of fruits or vegetables eaten per day	568	34.4
3 or more
2	562	34.1
1 or less	519	31.5
Number of times lean protein eaten per week	947	56.5
3 or more
1 to 2	674	40.2
0	55	3.3
Overall Nutritional Habit Score (0–14), Mean (SD)^†	6.01	(2.73)

* The sample N includes all those who answered at least 5 of the 7 nutritional habit questions. The sample size for individual items varies from 1642 to 1708 because of missing data.
† Scored as the sum of the 7 items, each 0 to 2, with higher scores reflecting poorer nutritional habits. If 2 or fewer items were missing, values for missing items were imputed as the mean of the nonmissing items. Cronbach’s α for the score is 0.558, and item total correlation range from 0.069 (lean protein consumption) to 0.407 (high-fat snack consumption). The median value is 6, and the most common (modal) value is 7.0, indicating very little skew. Visual inspection of the distribution and Q-Q plot indicate that the assumption of a normal distribution (and the application of parametric statistical tests) is valid.

Table 3.

Relationship between Nutritional Habit Score and Selected Demographic and Health Status Variables (N = 1714)

Variable	Nutritional Habit Score		P Value^*
Variable	Mean	SD	P Value^*
Race/ethnicity
African American	6.27	2.79	.297
American Indian	6.57	2.56	.759
Hispanic/Latino	5.10	2.45	<.001
Other	5.62	2.13	.285
White, non-Hispanic	6.12	2.79	—
Diabetes/pre-diabetes
Not diabetic and low risk	6.12	2.74	<.001
Not diabetic and high risk	6.43	2.68	<.001
Diabetic	5.06	2.64	—^†
Sex
Male	6.36	2.65	<.001
Female	5.82	2.75	—
Age
18 to 44 years old	6.42	2.76	<.001
45+ years old	5.67	2.67
Smoking status
Nonsmoker	5.73	2.68	<.001
Smoker	6.91	2.70	—
Alcohol consumption
Do not drink alcohol	5.87	2.68	.002
Drink alcohol	6.32	2.81	—
Physically active 5 days/week, >30 minutes/day
No	6.09	2.65	.350
Yes	5.94	2.85	—
Self-reported health status
Fair/poor	5.91	2.57	.493
Good/very good/excellent	6.07	2.81	—
BMI
Normal or underweight	5.99	2.90	.632
Overweight	6.25	2.69	.431
Obese	6.11	2.68	—^‡

* Based on ANOVA with nutritional habit score as the dependent variable and the specified demographic or health status characteristic as the independent variable and adjusting for practice location.
† P value for comparison of low risk versus high risk among nondiabetics is .097.
‡ P value for comparison of normal/underweight versus overweight is 0.243.

Table 4.

Multiple Linear Regression for Nutritional Habit Score (N = 1424)

Variable	Estimate	SE	P Value
Intercept	7.682	0.329	<.001
Race/ethnicity			<.001^*
African American	0.363	0.198	.067
American Indian	−0.109	0.342	.750
Hispanic/Latino	−1.284	0.273	<.001
Other	−0.726	0.892	.416
White, non-Hispanic	reference
Diabetes/prediabetes			<.001^†
Not diabetic, low risk	0.538	0.215	.013
Not diabetic, high risk	1.164	0.199	<.001
Diabetic	reference
Male gender	0.571	0.148	.001
Age (per year)	−0.041	0.005	<.001
Practice location			.003^‡
A	0.300	0.205	.143
B	−0.257	0.215	.232
C	0.628	0.283	.027
D	reference
Smoker	0.834	0.168	<.001
Drinks alcohol	−0.262	0.162	.106
Physically active 5 days/week, >30 minutes/day	−0.245	0.147	.097
Self-reported health status as very good or excellent	−0.102	0.159	.521

* Overall type 3 P value for race/ethnicity (4 df); additional pairwise comparisons in which P < .05 include African American versus Hispanic/Latino (P < .001) and American Indian versus Hispanic/Latino (P = .005).
† Overall type 3 P value for diabetes status (2 df); P value for pairwise comparison of high and low risk nondiabetics is <0.001.
‡ Overall type 3 P value for clinic location (3 df). Pairwise comparisons that are significant at P < .05 include A versus B (P = .004) and B versus C (P < .002).