Abstract
Large health disparities exist in the U.S. across ethnic and socioeconomic status groups. Using nationally representative data, we tested whether American patients with diet-related chronic diseases had higher diet quality than nonpatients. We also tested whether nutrition knowledge and beliefs (NKB) and food label (FL) use were associated with the observed differences. The 1994–1996 Continuing Survey of Food Intake by Individuals, and the Diet and Health Knowledge Survey were examined for 4356 U.S. adults. Dietary intakes were assessed using 2 nonconsecutive 24-h recalls and diet quality was assessed by using the USDA 2005 Healthy Eating Index (HEI). Patients’ mean HEI was higher than that of nonpatients (mean ± SE: 53.6 ± 0.5 vs. 51.8 ± 0.4; P < 0.001). Among patients, blacks were 92% more likely to report low diet quality (HEI < 20th percentile) than whites. The positive association between chronic diseases and HEI was observed only for patients with good NKB [OR = 1.80 (95% CI = 1.34, 2.43)]. The diabetes-HEI association was stronger among FL users [OR = 2.24 (95% CI = 1.08, 4.63)] than non-FL users [OR = 1.33 (95% CI = 0.65, 2.73)]. Hypertensive patients’ and nonpatients’ diet quality did not significantly differ; linear regression models showed no difference in their HEI (β ± SE: 0.6 ± 0.6; P > 0.05) or sodium intake (−18.6 ± 91.4 g/d; P > 0.05) between them. In conclusion, U.S. adults with diet-related chronic diseases reported somewhat higher diet quality than nonpatients, especially among those patients with good NKB and use of FL. Efforts are needed to promote healthy eating among Americans with diet-related chronic diseases; nutrition education and promotion of FL use may help.
Introduction
Unhealthful dietary practices, including high intakes of energy, sodium, and sugar and low intakes of fruits and vegetables (FV),7 increase the risks of chronic diseases such as obesity, hypertension, cardiovascular disease, and some types of cancer (1–3). Many Americans have been diagnosed with diet-related chronic diseases (4, 5). For example, >50 million Americans have hypertension (4) and 24 million children and adults have diabetes (5). Improvements in dietary intakes can help patients prevent and/or control many chronic conditions. Given the enormous public health costs of diet-related chronic diseases, improving dietary behaviors is urgently warranted (6).
People with chronic diseases may be more aware of nutrition information and make use of information on food labels (FL) more often than people without such conditions (6). The recent NHANES data showed that patients with chronic diseases who were advised by a health professional were more likely to read FL than those not receiving such advice, and patients who read FL consumed less energy, saturated fat, and sugar and more fiber than non-FL users (7). Few studies to date have examined whether U.S. patients with diet-related chronic diseases alter their diet after diagnosis and what factors may affect their likelihood of doing so.
National data show large disparities in U.S. chronic disease prevalence across racial/ethnic and socioeconomic status (SES) groups (8, 9). It is unclear whether the association between chronic diseases and diet quality varies across these groups. Dietary behavior may be driven by individuals’ nutrition knowledge and beliefs (NKB). Previously, we found that Americans who were less knowledgeable about nutrition had poorer diets (10), which is consistent with other findings (11).
Using nationally representative data, we examined whether U.S. adults with diet-related chronic diseases reported higher diet quality than other Americans and whether the association varied by their race/ethnicity, SES, NKB, and FL use. Because it is well documented that low diet quality is a risk factor for various chronic diseases, adults who initially have worse dietary intakes than average Americans would be more likely to develop the targeted chronic diseases. Thus, if individuals diagnosed with chronic diseases (patients) are found to have similar to or healthier intakes than those without these diseases (nonpatients) based on cross-sectional survey data, this would suggest (though not prove) that they may have improved their diet after diagnosis. If the patients are not found having a healthier diet than the average American, this would suggest the need for greater efforts to promote their eating more healthfully, e.g. for managing their diet-related health conditions.
Methods
Study design and data.
Data from the USDA Continuing Survey of Food Intakes by Individuals (CSFII) and the Diet and Health Knowledge Survey (DHKS) 1994–1996 were used to ascertain dietary intakes (12). Note that although these are cross-sectional data, they could allow us to assess whether Americans who have been diagnosed with diet-related chronic diseases tend to adopt an improved diet (e.g. due to self-education, awareness, family support, or their health professionals’ advice).
The CSFII included a nationally representative, multi-stage stratified sample of 16,103 noninstitutionalized individuals aged 0–90 y residing in the US. Information about dietary intakes by 1 or 2 nonconsecutive, multiple-pass, 24-h recalls collected 3–10 d apart were used. Information on demographics, SES, and lifestyle factors were also collected. A sample of CSFII participants (one adult aged ≥20 y/household) completed the DHKS to answer questions about knowledge and attitudes toward dietary guidance and health.
Study population.
Among CSFII participants, 9872 were aged ≥20 y and had data on d 1 of recall. Of these, 5765 participated in the DHKS. We further excluded those aged >65 y (n = 1319) to include relatively healthy individuals without special dietary needs and those with only one 24-h dietary recall (n = 90). A final sample of 4356 individuals (2219 men and 2137 women) was included.
Chronic diseases.
Participants were asked to report whether they had chronic disease(s) diagnosed by a physician during the survey period (1994–1996), including diabetes, high blood pressure, heart disease, cancer, osteoporosis, high blood cholesterol, and stroke. Participants who reported having 1 or more of the 7 chronic diseases were categorized as patients and those without these diseases were called nonpatients. We focused on diabetes, hypertension, heart disease, and hypercholesterolemia, which are common highly diet-related chronic diseases.
Dietary intakes and diet quality.
Averaged dietary intakes of foods and nutrients were analyzed based on the two 24-h recalls, including total energy, fat (as percent of total energy intake), cholesterol, sodium, calcium, fiber, sugar, and FV.
To assess the overall diet quality, we applied the USDA’s new 2005 Healthy Eating Index (HEI) (13–16). The USDA developed it to help assess how well American diets comply with the 2005 Dietary Guidelines for Americans. It consists of 12 components. Diets are scored on a density basis [amounts are expressed per 1000 kcal (1 kcal = 4.18 kJ) or as a percent of energy]. Six components (total fruit; whole fruit; total vegetable; dark green and orange vegetable and legume; total grain; and whole grain) are assigned 0–5 points; 5 components (milk, meat and beans, oil, saturated fat, and sodium) are worth 0–10 points; the last component of solid fat, alcohol, and added sugar is allocated 0–20 points. HEI scores were calculated proportionally.
HEI scores ranged from 0 to 100 (best) (17). Because a very small percentage of people had a HEI ≥80 (1.2%), a “good” diet as recommended by the USDA (18), we defined high diet quality as being HEI ≥80th percentile (HEI = 64) and low diet quality as HEI <20th percentile (HEI = 41).
Note that in general, 1 serving of FV consists of either of the following: 1 medium-sized fruit (such as apples, oranges, bananas, pears), 1/2 cup (1 cup equivalent = 0.000237 m3) of raw, cooked, canned, or frozen fruits or vegetables; 3/4 cup (6 oz.) of 100% fruit or vegetable juice; 1/2 cup fruit, cut up; 1/2 cup cooked or canned legumes (such as beans and peas); 1 cup of raw, leafy vegetables (such as lettuce and spinach); or 1/4 cup dried fruit (such as raisins, apricots, mango).
NKB.
DHKS participants were asked to rate on a Likert-type scale the degree of importance they were assigned to 11 specific dietary behavior/choices (Supplemental Text 1): 1) use salt or sodium only in moderation; 2) choose a diet low in saturated fat; 3) choose a diet with plenty of FV; 4) use sugars only in moderation; 5) choose a diet with adequate fiber; 6) eat a variety of foods; 7) maintain a healthy weight; 8) choose a diet low in fat; 9) choose a diet low in cholesterol; 10) choose a diet with plenty of breads, cereals, rice, and pasta; and 11) eat at least 2 servings of dairy products daily.
We calculated an overall NKB score (range: 11–44) to summarize participants' answers to these questions. First, answers of “not at all important,” “not too important,” “somewhat important,” or “very important” were assigned scores of 1, 2, 3, and 4, respectively; then the total score was summed. Higher scores indicated better NKB. A NKB score ≥ 80th percentile (score = 42) was considered a good NKB.
FL use.
Participants were asked 5 questions about FL use when buying foods, e.g. “Now think about food labels. When you buy foods, do you use: the list of ingredients: often, sometimes, rarely, or never?” FL users for ingredients were defined as those who answered “often” or “sometimes,” whereas answering “rarely,” “never,” or “never seen” were categorized as non-FL users. Similar questions were asked about the short phrases on the FL such as “low-fat;” the nutrition panel that lists the amount of energy, protein, fat, and such in a serving of the food; the serving size; and statements on the label that describe health benefits of nutrients or foods. Participants who answered all 5 questions with answers “often” or “sometimes” were categorized as FL users and others were categorized as non-FL users.
Overweight and obesity.
BMI (kg/m2) was calculated based on self-reported weight and height in the CSFII. BMI cutoffs of ≥25 and ≥30 were used to classify overweight and obesity, respectively. Those with BMI < 25 were categorized as normal weight.
Sociodemographic characteristics.
Participants were categorized as non-Hispanic (NH) whites, NH blacks, Hispanics, and others based on self-reported information. Education level was grouped into 3 categories: <high school (<12 y), high school, and >high school. Household income levels were assessed using poverty income ratio (PIR), and were categorized as 0–130% (poor: food stamp eligible), 131–350% (middle), and ≥350% (high income).
Other covariates.
All models were adjusted for survey year, 1990 Census geographic regions (Northeast, Midwest, South, and West), and degree of urbanization of the geographical area (metropolitan statistical area-central city, suburban, and rural).
Statistical analysis.
The primary exposure variables were chronic diseases diagnosed by doctors, whereas the outcome variables were dietary intakes, including diet quality. χ2 tests tested the differences in the percentages of chronic diseases across sociodemographics. The Wald F tests for continuous variables (e.g. HEI) and χ2 tests for categorical variables (e.g. high HEI) were used to examine the distributions of these variables by sociodemographics and disease status.
Multivariable linear and logistic regression models were conducted to examine the associations of chronic diseases with dietary intakes as continuous or categorical variables, respectively. OR and 95% CI were calculated. Covariates, including age, gender, race/ethnicity, income, education, survey year, region, urbanization, and overweight/obesity, were adjusted for in all models.
Stratified analyses were conducted by sex, race/ethnicity, income, education, NKB, and FL use to test whether the associations varied by these characteristics. Interaction terms were included and tested in separate models. Moreover, we examined whether the association varied by the number and type of chronic diseases.
All analyses were conducted using survey-related commands in SAS (version 9.2; SAS Institute) to take complex sampling design into account to produce nationally representative estimates and to correct estimates of SE. Significance was set at P < 0.05.
Results
The overall prevalence of chronic diseases was 28% (Supplemental Table 1). Older, NH black, lower SES, and obese Americans had a higher prevalence than average.
Americans with high diet quality (HEI ≥80th percentile) were more likely to be older, women, NH white or Hispanic, better educated, and FL users and have a higher income, higher NKB scores, and normal weight (Table 1). Compared with nonpatients, patients had higher HEI scores, lower energy intake, and higher FV intake. Only 1.7% of patients and 1.1% of nonpatients overall, however, had a HEI ≥80. Patients and nonpatients did not differ in the percentage consuming ≥5 servings/d of FV or the percentage consuming ≥2 servings/d of fruits or ≥3 servings/d of vegetables. Similar results were found for patients reporting diabetes and high blood cholesterol. There was no significant difference in HEI or FV intake between people with and without heart disease or hypertension, although these patients reported lower energy intake.
TABLE 1.
Diet quality [Healthy Eating Index (HEI)] and dietary intakes [energy and fruits and vegetables (FV)] among U.S. adults aged 20–65 y by sociodemographic characteristics and chronic diseases1
| Characteristic | n | HEI | Energy, MJ/d | FV, g/d | HEI ≥ 80 | HEI <20th | HEI ≥80th | FV ≥55 | F ≥2 and V ≥35 |
| All | 4356 | 52.3 ± 0.4 | 8.72 ± 0.11 | 372 ± 6.9 | 1.2 (0.2) | 19.9 (1.1) | 20.8 (0.9) | 42.1 (1.2) | 15.2 (0.8) |
| Age, y | |||||||||
| 20–34 | 1165 | 50.4 ± 0.6*** | 9.46 ± 0.17*** | 359 ± 9.8*** | 0.4 (0.2)*** | 24.2 (2.3)*** | 16.3 (1.6)*** | 42.1 (1.7) | 14.1 (1.3) |
| 35–49 | 1507 | 52.2 ± 0.5 | 8.65 ± 0.23 | 364 ± 11.7 | 1.2 (0.4) | 20.7 (1.7) | 20.5 (1.4) | 41.4 (2.1) | 14.8 (1.2) |
| 50–65 | 1684 | 55.2 ± 0.4 | 7.73 ± 0.09 | 403 ± 9.8 | 2.6 (0.4) | 12.3 (1.0) | 27.9 (1.5) | 43.1 (1.7) | 17.4 (1.0) |
| Sex | |||||||||
| Men | 2219 | 50.7 ± 0.4*** | 10.5 ± 0.19*** | 399 ± 8.4*** | 1.1 (0.2) | 23.8 (1.7)*** | 16.9 (1.1)*** | 48.6 (1.5)*** | 17.3 (1.2)** |
| Women | 2137 | 53.8 ± 0.5 | 6.98 ± 0.06 | 347 ± 8.5 | 1.4 (0.3) | 16.1 (1.2) | 24.6 (1.4) | 35.8 (1.5) | 13.2 (1.0) |
| Race/ethnicity | |||||||||
| NH white | 3285 | 52.3 ± 0.4*** | 8.71 ± 0.09*** | 368 ± 7.4** | 1.2 (0.2) | 19.6 (1.3)*** | 19.8 (1.1)*** | 42.0 (1.3)* | 15.2 (0.8) |
| NH black | 512 | 47.6 ± 0.6 | 8.97 ± 0.71 | 348 ± 17.7 | 1.3 (0.6) | 30.1 (2.8) | 11.6 (2.3) | 36.1 (3.5) | 10.9 (1.6) |
| Hispanic | 411 | 54.4 ± 1.0 | 8.40 ± 0.24 | 388 ± 22.9 | 1.5 (0.8) | 17.0 (3.0) | 26.1 (3.3) | 43.6 (3.3) | 16.7 (2.5) |
| Other | 148 | 59.3 ± 1.7 | 8.88 ± 0.55 | 461 ± 24.9 | 0.8 (0.5) | 6.3 (1.4) | 45.6 (7.6) | 53.8 (5.9) | 21.6 (6.0) |
| Education | |||||||||
| <High school | 731 | 49.4 ± 0.6*** | 8.46 ± 0.59*** | 306 ± 19.4*** | 1.5 (0.6) | 30.5 (2.7)*** | 18.2 (2.0)*** | 33.4 (3.1)*** | 8.6 (1.3)*** |
| High school | 1552 | 49.9 ± 0.5 | 8.56 ± 0.14 | 326 ± 7.1 | 0.7 (0.3) | 25.6 (1.9) | 17.0 (1.3) | 34.6 (1.8) | 11.6 (1.1) |
| >High school | 2073 | 54.5 ± 0.4 | 8.87 ± 0.10 | 417 ± 8.3 | 1.5 (0.2) | 13.7 (1.0) | 23.8 (1.2) | 48.9 (1.5) | 19.0 (1.0) |
| Income2 | |||||||||
| Low | 1079 | 49.5 ± 0.7*** | 8.80 ± 0.54* | 322 ± 14.4*** | 1.1 (0.4) | 26.7 (2.5)*** | 15.4 (2.1)*** | 33.2 (2.9)*** | 8.7 (1.3)*** |
| Middle | 1543 | 51.0 ± 0.5 | 8.69 ± 0.12 | 357 ± 8.3 | 0.9 (0.3) | 22.7 (1.4) | 17.8 (1.1) | 39.9 (1.4) | 13.7 (1.0) |
| High | 1734 | 54.4 ± 0.5 | 8.71 ± 0.12 | 401 ± 9.8 | 1.5 (0.3) | 15.2 (1.5) | 25.1 (1.2) | 46.8 (1.8) | 18.6 (1.3) |
| NKB | |||||||||
| <80th percentile | 3451 | 51.3 ± 0.4*** | 8.90 ± 0.14*** | 364 ± 7.2*** | 0.7 (0.2)*** | 21.8 (1.4)*** | 18.0 (0.9)*** | 41.7 (1.4) | 14.6 (0.9) |
| ≥80th percentile | 896 | 56.2 ± 0.7 | 7.99 ± 0.18 | 403 ± 11.7 | 3.3 (0.7) | 12.5 (1.6) | 31.8 (2.3) | 43.9 (1.9) | 17.7 (1.6) |
| FL use3 | |||||||||
| Non-FL use | 3143 | 50.9 ± 0.4*** | 9.07 ± 0.15*** | 359 ± 7.7*** | 0.9 (0.2)* | 23.3 (1.4)*** | 17.9 (1.0)*** | 41.1 (1.3) | 13.8 (0.9)** |
| FL use | 1211 | 55.8 ± 0.5 | 7.81 ± 0.12 | 406 ± 11.2 | 2.2 (0.5) | 11.2 (1.1) | 28.3 (1.8) | 44.7 (1.9) | 18.9 (1.5) |
| Weight status | |||||||||
| BMI <25 | 1803 | 53.0 ± 0.6*** | 8.49 ± 0.12*** | 378 ± 10.3*** | 1.2 (0.2) | 18.4 (1.7) | 22.7 (1.5)* | 41.2 (1.8)** | 16.7 (1.3)** |
| BMI 25–29.9 | 1540 | 52.2 ± 0.5 | 9.33 ± 0.24 | 391 ± 11.2 | 1.2 (0.3) | 20.1 (1.7) | 20.2 (1.4) | 45.8 (1.8) | 15.6 (1.0) |
| BMI ≥30 | 916 | 50.5 ± 0.6 | 8.31 ± 0.17 | 326 ± 11.6 | 1.4 (0.5) | 23.4 (1.6) | 16.6 (2.1) | 37.5 (2.1) | 11.4 (1.2) |
| Chronic diseases4 | |||||||||
| No | 2830 | 51.8 ± 0.4*** | 8.95 ± 0.14*** | 367 ± 8.1* | 1.1 (0.2) | 21.4 (1.3)** | 19.4 (1.1)** | 42.1 (1.5) | 15.1 (1.0) |
| Yes | 1526 | 53.6 ± 0.5 | 8.11 ± 0.14 | 386 ± 7.9 | 1.7 (0.3) | 16.0 (1.4) | 24.5 (1.5) | 41.9 (1.5) | 15.4 (1.1) |
| Diabetes | |||||||||
| No | 4100 | 52.1 ± 0.4*** | 8.77 ± 0.11*** | 371 ± 7.3* | 1.2 (0.2) | 20.3 (1.1)*** | 20.4 (0.9)* | 42.4 (1.3) | 15.2 (0.8) |
| Yes | 252 | 56.1 ± 1.0 | 7.33 ± 0.26 | 401 ± 25.7 | 1.6 (0.7) | 8.4 (1.9) | 30.1 (4.7) | 35.4 (4.1) | 15.6 (2.6) |
| Heart disease | |||||||||
| No | 4046 | 52.3 ± 0.4 | 8.75 ± 0.31*** | 372 ± 7.1 | 1.3 (0.2)* | 20.0 (1.1) | 20.5 (0.9) | 42.1 (1.3) | 15.3 (0.8) |
| Yes | 310 | 53.1 ± 0.8 | 7.95 ± 0.27 | 380 ± 19.5 | 0.4 (0.2) | 16.6 (3.0) | 27.7 (4.0) | 41.6 (4.2) | 13.2 (2.2) |
| Hypercholesterolemia | |||||||||
| No | 3735 | 52.0 ± 0.4*** | 8.79 ± 0.12*** | 368 ± 7.1** | 1.1 (0.2)* | 20.8 (1.2)*** | 20.2 (1.0) | 41.9 (1.2) | 15.0 (0.9) |
| Yes | 615 | 54.8 ± 0.7 | 8.16 ± 0.20 | 407 ± 16.9 | 2.2 (0.6) | 12.6 (1.7) | 25.2 (2.6) | 43.3 (3.0) | 16.4 (1.9) |
| Hypertension | |||||||||
| No | 3460 | 52.2 ± 0.4 | 8.82 ± 0.12*** | 370 ± 7.5 | 1.2 (0.2) | 20.4 (1.2) | 20.3 (1.0) | 42.0 (1.3) | 15.5 (0.9) |
| Yes | 890 | 52.8 ± 0.6 | 8.17 ± 0.19 | 386 ± 10.8 | 1.5 (0.5) | 16.9 (1.7) | 23.7 (2.0) | 42.6 (2.1) | 13.8 (1.4) |
Values are mean ± SEM for continuous variables and percentage (SE) for categorical variables. * < 0.05, **P < 0.01, ***P < 0.001 for between-group differences.
Household income levels were assessed using poverty income ratio (PIR) and were categorized as 0–130% (low income: food stamp eligible), 131–350% (middle income), and ≥350% (high income).
FL use was defined as using all of FL information for ingredient, short phrase, nutrition panel, serving size, and health benefits of nutrients/foods.
Selected health condition(s), including 1 or more of these chronic diseases: diabetes, heart disease, hypercholesterolemia, hypertension, or cancer.
Unit: servings/d. In general, 1 serving of FV consists of 1 of the following: 1 medium-sized fruit (such as apples, oranges, bananas, pears); 1/2 cup (1 cup equivalent = 0.000237 m3) of raw, cooked, canned, or frozen FV; 3/4 cup (6 oz.) of 100% fruit or vegetable juice; 1/2 cup fruit, cut up; 1/2 cup cooked or canned legumes (such as beans and peas); 1 cup of raw, leafy vegetables (such as lettuce and spinach); or 1/4 cup dried fruit (such as raisins, apricots, mango).
Linear regression models with adjustment for potential confounders showed that compared with nonpatients, patients had higher HEI (Table 2). Overall, patients had relatively lower energy intake (P > 0.05), but this was only significant for diabetes patients (P < 0.05). Patients with hypercholesterolemia had a lower fat intake and higher fiber intake than those without hypercholesterolemia. Cholesterol and sodium intake did not differ between patients and nonpatients. No significant differences were found in HEI, sodium, or other dietary intakes by hypertensive status (all P > 0.05). Patients with diabetes and heart disease consumed less sugar. In general, FV intakes were higher among patients (P < 0.01), but this difference was not significant when each disease was examined separately (all P > 0.05).
TABLE 2.
Linear regression models for the associations between chronic diseases and diet quality and average daily dietary intakes among U.S. adults aged 20–65 y1
| Outcome/disease | Model 1: Chronic diseases2 | Model 2: Diabetes | Model 3: Heart disease | Model 4: Hypercholesterolemia | Model 5: Hypertension |
| HEI | 1.1 (0.5)* | 3.9 (1.0)*** | 0.3 (0.9) | 1.8 (0.7)* | 0.6 (0.6) |
| Energy, kJ/d | −283 (193) | −884 (342)* | −346 (324) | −107 (162) | −291 (205) |
| Total fat intake, % energy | 0.0 (0.4) | 1.3 (0.8) | −0.2 (0.6) | −1.0 (0.5)* | 0.3 (0.4) |
| Cholesterol intake, mg/d | −9.4 (10.1) | −4.7 (19.0) | 1.3 (11.9) | −10.3 (12.3) | −1.3 (10.9) |
| Sodium intake, mg/d | −10.8 (83.6) | −169 (144) | −82.2 (127) | 84.3 (100) | −18.6 (91.4) |
| Calcium intake, mg/d | 1.6 (20.9) | −24.3 (41.9) | −17.5 (31.5) | 27.1 (21.7) | −38.9 (20.5) |
| Fiber intake, g/d | 0.4 (0.4) | −0.7 (0.7) | 0.2 (0.7) | 1.3 (0.6)* | 0.1 (0.5) |
| Sugar intake, g/d | −2.0 (1.9) | −12.7 (2.3)*** | −5.1 (1.9)* | −0.2 (2.1) | −3.3 (2.4) |
| FV intake, g/d | 23.1 (8.3)** | 37.4 (28.6) | 5.1 (18.1) | 34.3 (17.4) | 21.7 (11.9) |
Values are β ± SEE, which showed the adjusted differences in intakes between patients vs. nonpatients. * < 0.05, **P < 0.01, ***P < 0.001. Separate models were fit for each dietary outcome. Each model adjusted for survey year, sex, age, education, poverty income ratio (PIR), race/ethnicity, BMI, region, and urbanization.
Selected health condition(s), including 1 or more of these chronic diseases: diabetes, heart disease, hypercholesterolemia, hypertension, or cancer.
Stratified analyses show that NH blacks with and without any disease had higher intakes of cholesterol and fat than did NH whites (Table 3). Hispanics had a lower sodium intake than NH whites with hypertension (β ± SE: −343 ± 136 mg/d; P < 0.05) and those without hypertension (β ± SE: −780 ± 194 mg/d; P < 0.001).
TABLE 3.
Linear regression models for the associations of chronic diseases with diet among U.S. adults aged 20–65 y, stratified by race/ethnicity1
| Chronic diseases2 |
Diabetes |
Heart disease |
Hypercholesterolemia |
Hypertension |
||||||
| Characteristic | Without | With | Without | With | Without | With | Without | With | Without | With |
| HEI | ||||||||||
| NH black | −2.9 (0.8)*** | −3.4 (1.4)* | −3.6 (0.9)*** | 0.5 (2.3) | −3.1 (0.8)*** | −3.2 (2.3) | −3.1 (0.7)*** | −2.7 (2.1) | −3.2 (0.8)*** | −2.7 (1.8) |
| Hispanic | 3.9 (1.0)*** | 3.9 (2.0) | 3.5 (0.9)*** | 3.5 (3.4) | 3.7 (0.9)*** | 7.4 (5.1) | 3.7 (1.0)*** | 6.0 (2.0)** | 3.9 (0.9)*** | 3.5 (3.0) |
| Other | 6.5 (1.7)*** | 4.9 (2.8) | 6.3 (1.4)*** | 5.6 (4.5) | 6.3 (1.4)*** | 16.6 (6.2)* | 6.4 (1.6)*** | 6.3 (5.1) | 6.5 (1.5)*** | 4.6 (2.2)* |
| Cholesterol intake | ||||||||||
| NH black | 57.5 (29.9) | 92.2 (19.0)*** | 59.7 (22.9)* | 143 (40.7)*** | 60.8 (21.1)*** | 158 (34.7)*** | 60.8 (21.7)** | 110 (32.4)** | 68.5 (27.8)* | 71.6 (24.4)** |
| Hispanic | 6.1 (13.8) | 15.0 (28.8) | 11.7 (12.3) | −37.9 (38.1) | 7.2 (13.4) | −7.1 (47.2) | 5.2 (11.9) | 51.4 (49.3) | 10.3 (13.8) | −3.5 (37.8) |
| Other | 35.5 (28.9) | −11.5 (25.9) | 30.2 (26.1) | −75.9 (46.3) | 29.5 (25.4) | −264 (49.3)*** | 29.5 (25.6) | −9.3 (48.0) | 33.6 (28.2) | −37.5 (27.9) |
| Total fat intake | ||||||||||
| NH black | 1.9 (0.7)* | 1.2 (0.8) | 1.8 (0.6)** | −0.9 (1.3) | 1.7 (0.7)* | 1.3 (1.2) | 1.6 (0.6)* | 1.6 (1.4) | 2.2 (0.7)** | −0.1 (1.1) |
| Hispanic | −0.2 (0.6) | −1.7 (1.2) | −0.3 (0.7) | −7.3 (2.6)** | −0.6 (0.7) | 1.8 (3.2) | −0.4 (0.7) | −3.5 (1.8) | −0.3 (0.7) | −2.8 (1.5) |
| Other | −3.4 (1.0)*** | −4.1 (1.6)* | −3.7 (0.9)*** | −4.1 (2.5) | −3.6 (0.9)*** | −9.2 (2.1)*** | −3.4 (0.9)*** | −6.6 (2.2)** | −3.6 (1.0)** | −4.1 (1.7)* |
Values are β (SEE). Each model adjusted for survey year, sex, age, education, poverty income ratio (PIR), race/ethnicity, BMI, region, and urbanization. NH whites served as the reference. * < 0.05, **P < 0.01, ***P < 0.001.
Selected health condition(s), including 1 or more of these chronic diseases: diabetes, heart disease, hypercholesterolemia, hypertension, or cancer.
We fit multivariable logistic regression models to examine the association between chronic diseases and diet quality (Table 4). Patients were 30% more likely to have high HEI than nonpatients. Diabetes patients were 65% more likely to report a high diet quality than nondiabetes patients. The significant associations of overall chronic diseases and diabetes with high HEI were observed among men but not among women. The disease-diet association was slightly stronger among people with middle SES, but the interaction terms were not significant. The significant association between chronic diseases and high HEI was observed for people with good NKB (≥80th percentile) but not for those with an NKB <80th percentile. Hypercholesterolemia patients were 68% more likely to have high HEI among those with better NKB but only 3% more likely if with an NKB <80th percentile (P-interaction = 0.18). A significant association for high HEI was observed for diabetes among FL users but not in nonusers. We also conducted similar models for FV (FV ≥ 5 vs. FV <5 servings/d) and low diet quality (HEI <20th percentile) and found similar results (data not shown). Our further combined analysis showed no association among people with poor NKB and no FL use; the association was significant among people with good NKB but no FL use. Similar results were found for hypertension (P-interaction = 0.006). A significant association between hypercholesterolemia was only observed among people with good NKB and FL use.
TABLE 4.
Logistic regression models for the association between chronic diseases and high diet quality (HEI ≥ 80th vs. HEI < 80th percentile) among U.S. adults aged 20–65 y1
| Characteristic | Model 1: Chronic diseases4 | Model 2: Diabetes | Model 3: Heart disease | Model 4: Hypercholesterolemia | Model 5: Hypertension |
| Total, for whole sample1 | 1.30 (1.07, 1.58)* | 1.65 (1.02, 2.64)* | 1.42 (0.89, 2.28) | 1.19 (0.88, 1.61) | 1.22 (0.96, 1.55) |
| Stratified by sex2 | |||||
| Men | 1.57 (1.10, 2.24)* | 2.73 (1.38, 5.39)* | 1.88 (0.98, 3.58) | 1.27 (0.84, 1.92) | 1.25 (0.80, 1,95) |
| Women | 1.14 (0.88, 1.47) | 1.08 (0.56, 2.09) | 1.18 (0.63, 2.23) | 1.19 (0.85, 1.69) | 1.25 (0.92, 1.71) |
| Percent change in OR | 37.7 | 152.8 | 59.3 | 6.7 | 0.0 |
| P-interaction | 0.166 | 0.074 | 0.296 | 0.813 | 0.989 |
| Stratified by race/ethnicity2 | |||||
| NH white | 1.36 (1.06, 1.73)* | 1.55 (0.92, 2.63) | 1.44 (0.84 2.49) | 1.26 (0.93, 1.72) | 1.18 (0.92, 1.51) |
| NH black | 1.47 (0.64, 3.40) | 4.24 (1.58, 11.4)** | 1.15 (0.33, 3.97) | 1.02 (0.39, 2.66) | 1.64 (0.72, 3.72) |
| Hispanic | 1.79 (0.78, 4.10) | 2.00 (0.57, 7.05) | 5.04 (1.45, 17.5)* | 2.00 (0.79, 5.06) | 2.25 (0.84, 6.06) |
| Other | 0.67 (0.22, 2.02) | 1.67 (0.33, 8.41) | 3.65 (0.53, 25.2) | 1.55 (0.18, 13.6) | 0.36 (0.11, 1.21) |
| Percent change in OR (white and black) | 8.1 | 173.5 | 338.3 | 23.5 | 39.0 |
| Percent change in OR (all) | 167.2 | 173.5 | 25.2 | 96.1 | 525.0 |
| P-interaction | 0.380 | 0.789 | 0.579 | 0.980 | 0.010 |
| Stratified by food stamp program2 | |||||
| Nonparticipants | 1.24 (1.00, 1.53)* | 1.62 (0.92, 2.86) | 1.28 (0.71, 2.33) | 1.19 (0.88, 1.61) | 1.17 (0.91, 1.51) |
| Participants | 7.64 (3.09, 18.9)*** | 3.17 (0.66, 15.3) | 2.28 (0.50, 10.5) | 1.06 (0.36, 3.11) | 2.42 (0.97, 6.04) |
| Percent change in OR | 516.1 | 95.7 | 78.1 | 12.3 | 106.8 |
| P-interaction | 0.045 | 0.402 | 0.205 | 0.500 | 0.147 |
| Stratified by FL use23 | |||||
| Non FL use | 1.23 (0.94, 1.61) | 1.33 (0.65, 2.73) | 1.36 (0.70, 2.64) | 1.00 (0.74, 1.35) | 1.39 (0.99, 1.93) |
| FL use | 1.35 (0.92, 1.96) | 2.24 (1.08, 4.63)* | 1.54 (0.87, 2.73) | 1.46 (0.93, 2.28) | 1.04 (0.66, 1.64) |
| Percent change in OR | 9.8 | 68.4 | 13.2 | 46.0 | 33.7 |
| P-interaction | 0.939 | 0.634 | 0.862 | 0.256 | 0.207 |
| Stratified by NKB2 | |||||
| Score <80th percentile | 1.13 (0.89, 1.42) | 1.61 (0.87, 2.97) | 1.31 (0.72, 2.38) | 1.03 (0.76, 1.40) | 1.13 (0.84, 1.53) |
| Score ≥80th percentile | 1.80 (1.34, 2.43) | 1.58 (0.85, 2.95) | 1.43 (0.74, 2.74) | 1.68 (1.09, 2.60)* | 1.62 (0.99, 2.66)* |
| Percent change in OR | 59.3 | 1.9 | 9.2 | 63.1 | 43.4 |
| P-interaction | 0.243 | 0.727 | 0.970 | 0.181 | 0.574 |
| Stratified by NKB and FL use2 | |||||
| Poor NKB and non-FL use | 1.07 (0.78, 1.46) | 1.03 (0.43, 2.46) | 1.31 (0.59, 2.91) | 0.89 (0.64, 1.25) | 1.11 (0.73, 1.68) |
| Poor NKB and FL use | 1.21 (0.76, 1.94) | 2.90 (1.26, 6.65) | 1.20 (0.62, 2.36) | 1.23 (0.69, 2.18) | 1.22 (0.72, 2.06) |
| Good NKB and non-FL use | 2.16 (1.25, 3.74) | 2.20 (0.78, 6.19) | 1.04 (0.37, 2.96) | 1.24 (0.67, 2.30) | 3.52 (1.66, 7.48) |
| Good NKB and FL use | 1.64 (0.95, 2.81) | 1.07 (0.46, 2.51) | 1.91 (0.68, 5.34) | 2.30 (1.25, 4.23) | 0.82 (0.37, 1.80) |
| Percent change in OR | 102.7 | 180.4 | 83.8 | 157.0 | 331.9 |
| P-interaction | 0.131 | 0.089 | 0.995 | 0.218 | 0.006 |
Values are OR (95% CI). Each model adjusted for survey year, sex, age, education, poverty income ratio (PIR), race/ethnicity, BMI, region, and urbanization. * < 0.05, **P < 0.01, ***P < 0.001.
Adjusted for all above covariates, except for the variables stratified by (e.g. sex, education).
FL use was defined as using all of label information for ingredient, short phrase, nutrition panel, serving size, and health benefits of foods.
Selected chronic disease(s), including at least 1 of diseases: diabetes, heart disease, hypercholesterolemia, hypertension, or cancer.
We found a dose-response relationship between the number of chronic diseases and high HEI (Supplemental Table 2). Patients with multiple diseases were more likely to have high diet quality. For example, people with 3 chronic diseases were 81% more likely to have high HEI and 46% less likely to have low HEI. Compared with people with only hypertension (the most common chronic disease reported), those diagnosed with diabetes were 87% less likely to report low diet quality. There was no significant difference in diet quality between hypertensive patients and those without chronic diseases.
We also compared the predictors of high and low HEI between patients and nonpatients (Table 5). Among patients, NH blacks were 92% more likely to have low HEI than were NH whites. Good NKB was significantly associated with high HEI among both patients and nonpatients. Patients with better NKB were 57% less likely to have a low HEI than those with worse NKB. FL users were 43% less likely to have a low HEI than non-FL users, again among both patients and nonpatients. Similar results were observed for hypertension and diabetes. Diabetes and heart disease patients with higher income were less likely to have a low HEI.
TABLE 5.
Logistic regression models for the predictors of high diet quality (HEI ≥ 80th vs. HEI < 80th percentile) and low diet quality (HEI < 20th vs. HEI ≥20th percentile) among U.S. adults aged 20–65 y with and without chronic disease(s)1
| High diet quality, ≥ 80th vs. HEI < 80th percentile |
Low diet quality, HEI < 20th vs. HEI ≥ 20th percentile |
|||
| Characteristic | Without disease(s): Model 1 | With disease(s): Model 2 | Without disease(s): Model 3 | With disease(s): Model 4 |
| Chronic diseases2 | ||||
| Women (ref: men) | 1.75 (1.36, 2.24)*** | 1.06 (0.78, 1.45) | 0.63 (0.46, 0.85)** | 0.74 (0.52, 1.06) |
| Age (ref: 20–34 y) | ||||
| 35–49 | 1.59 (1.14, 2.23)** | 0.76 (0.43, 1.37) | 0.95 (0.66, 1.36) | 0.55 (0.30, 1.00)* |
| 50–65 | 2.33 (1.62, 3.36)*** | 1.35 (0.75, 2.42) | 0.47 (0.31, 0.71)*** | 0.28 (0.16, 0.51)*** |
| Education (ref: <high school) | ||||
| High school | 0.95 (0.55, 1.64) | 0.81 (0.54, 1.21) | 0.72 (0.48, 1.08) | 0.72 (0.44, 1.19) |
| >High school | 1.60 (0.95, 2.67) | 0.88 (0.54, 1.43) | 0.31 (0.21, 0.46)*** | 0.46 (0.27, 0.80)** |
| Household income (ref: low)3 | ||||
| Middle | 0.95 (0.63, 1.42) | 1.67 (1.01, 2.76)* | 0.98 (0.71, 1.36) | 0.92 (0.52, 1.61) |
| High | 1.24 (0.80, 1.91) | 2.03 (1.21, 3.41)** | 0.81 (0.49, 1.33) | 0.73 (0.40, 1.32) |
| Race/ethnicity (ref: NH white) | ||||
| NH black | 0.52 (0.27, 1.02) | 0.74 (0.36, 1.54) | 1.34 (0.85, 2.10) | 1.92 (1.09, 3.36)* |
| Hispanic | 1.69 (1.07, 2.67)* | 1.74 (0.78, 3.87) | 0.51 (0.31, 0.84)*** | 0.96 (0.37, 2.45) |
| Other | 3.83 (2.00, 7.32)*** | 1.62 (0.74, 3.54) | 0.17 (0.08, 0.37)*** | 1.33 (0.39, 4.52) |
| Good NKB vs. other4 | 1.59 (1.18, 2.12)** | 2.05 (1.51, 2.79)*** | 0.76 (0.50, 1.16) | 0.43 (0.26, 0.71)** |
| FL use5 | 1.44 (1.07, 1.93)* | 1.36 (0.93, 2.00) | 0.50 (0.38, 0.66)*** | 0.57 (0.38, 0.84)** |
| Diabetes | ||||
| Women (ref: men) | 1.61 (1.31, 1.98)*** | 0.45 (0.16, 1.21) | 0.66 (0.52, 0.83)** | 0.58 (0.15, 2.28) |
| Age (ref: 20–34 y) | ||||
| 35–49 | 1.42 (1.06, 1.91)* | 1.51 (0.18, 12.98) | 0.87 (0.65, 1.17) | 1.74 (0.19, 15.65) |
| 50–65 | 2.25 (1.66, 3.07)*** | 2.82 (0.27, 29.71) | 0.39 (0.29, 0.51)*** | 6.29 (0.71, 55.84) |
| Education (ref: <high school) | ||||
| High school | 0.90 (0.66, 1.22) | 1.63 (0.58, 4.56) | 0.72 (0.51, 1.01) | 1.70 (0.46, 6.23) |
| >High school | 1.38 (0.99, 1.94) | 0.95 (0.29, 3.16) | 0.33 (0.24, 0.46)*** | 0.56 (0.09, 3.63) |
| Income (ref: low) | ||||
| Middle | 1.04 (0.73, 1.48) | 2.80 (0.84, 9.35) | 1.02 (0.78, 1.35) | 0.33 (0.11, 0.97)* |
| High | 1.35 (0.95, 1.90) | 1.88 (0.52, 6.75) | 0.86 (0.57, 1.30) | 0.06 (0.01, 0.35)** |
| Race/ethnicity (ref: NH white) | ||||
| NH black | 0.55 (0.29, 1.05) | 0.97 (0.37, 2.56) | 1.51 (1.00, 2.28)* | 1.29 (0.22, 7.52) |
| Hispanic | 1.56 (1.07, 2.29)* | 3.01 (0.77, 11.7) | 0.61 (0.40, 0.94)* | 0.95 (0.09, 9.84) |
| Other | 3.21 (1.81, 5.68)*** | 10.4 (1.0, 106.5)* | 0.27 (0.17, 0.41)*** | 16.3 (1.9, 143.6)* |
| Good NKB vs. other4 | 1.75 (1.36, 2.25)*** | 1.88 (0.69, 5.10) | 0.66 (0.46, 0.95)* | 1.26 (0.36, 4.46) |
| FL use5 | 1.39 (1.09, 1.77)** | 1.68 (0.69, 4.08) | 0.50 (0.40, 0.62)*** | 0.93 (0.28, 3.07) |
| Heart disease | ||||
| Women (ref: men) | 1.57 (1.28, 1.93)*** | 0.97 (0.43, 2.18) | 0.66 (0.51, 0.84)* | 0.71 (0.26, 1.94) |
| Age (ref: 20–34 y) | ||||
| 35–49 | 1.47 (1.11, 1.94)* | 0.30 (0.12, 0.80)* | 0.85 (0.63, 1.15) | 0.46 (0.13, 1.66) |
| 50–65 | 2.41 (1.80, 3.23)*** | 0.77 (0.33, 1.82) | 0.38 (0.28, 0.50)** | 0.15 (0.05, 0.41)*** |
| Education (ref: <high school) | ||||
| High school | 0.85 (0.61, 1.17) | 1.41 (0.51, 3.91) | 0.71 (0.51, 1.00) | 0.84 (0.32, 2.18) |
| >High school | 1.31 (0.94, 1.82) | 0.99 (0.30, 3.25) | 0.33 (0.24, 0.46)** | 0.70 (0.24, 2.02) |
| Income (ref: low) | ||||
| Middle | 1.03 (0.73, 1.45) | 2.32 (0.60, 8.88) | 1.02 (0.76, 1.36) | 0.47 (0.19, 1.12) |
| High | 1.31 (0.94, 1.84) | 3.37 (0.84, 13.6) | 0.86 (0.57, 1.29) | 0.30 (0.09, 0.97)* |
| Race/ethnicity (ref: NH white) | ||||
| NH black | 0.62 (0.35, 1.12) | 0.32 (0.06, 1.65) | 1.52 (1.02, 2.25) | 0.45 (0.15, 1.37) |
| Hispanic | 1.59 (1.10, 2.31)* | 4.73 (0.85, 26.23) | 0.60 (0.39, 0.92)* | 1.30 (0.05, 33.12) |
| Other | 3.27 (1.84, 5.81)*** | 18.2 (2.39, 138.8)** | 0.30 (0.20, 0.46)** | NA6 |
| Good NKB vs. other4 | 1.72 (1.32, 2.23)*** | 1.91 (0.88, 4.17) | 0.66 (0.46, 0.95)* | 0.70 (0.31, 1.57) |
| FL use5 | 1.43 (1.11, 1.84)** | 2.56 (1.17, 5.58)* | 0.50 (0.40, 0.62)** | 1.10 (0.40, 3.03) |
| Hypercholesterolemia | ||||
| Women (ref: men) | 1.59 (1.28, 1.96)*** | 1.15 (0.69, 1.91) | 0.66 (0.51, 0.86)* | 0.58 (0.28, 1.18) |
| Age (ref: 20–34 y) | ||||
| 35–49 | 1.43 (1.08, 1.91)* | 1.18 (0.41, 3.36) | 0.94 (0.70, 1.28) | 0.22 (0.09, 0.50)*** |
| 50–65 | 2.16 (1.57, 2.96)*** | 3.02 (1.11, 8.25)* | 0.43 (0.32, 0.59)** | 0.12 (0.05, 0.29)*** |
| Education (ref: <high school) | ||||
| High school | 1.01 (0.72, 1.43) | 0.45 (0.20, 1.01) | 0.73 (0.52, 1.03) | 0.42 (0.16, 1.09) |
| >High school | 1.48 (1.02, 2.14)* | 0.69 (0.32, 1.49) | 0.33 (0.24, 0.45)*** | 0.32 (0.13, 0.78)** |
| Income (ref: low) | ||||
| Middle | 1.04 (0.71, 1.52) | 2.49 (1.09, 5.69)* | 1.04 (0.79, 1.36) | 0.31 (0.12, 0.83)** |
| High | 1.38 (0.94, 2.03) | 2.55 (1.24, 5.24)* | 0.85 (0.56, 1.30) | 0.42 (0.14, 1.31) |
| Race/ethnicity (ref: NH white) | ||||
| NH black | 0.59 (0.34, 1.03) | 0.48 (0.14, 1.66) | 1.46 (0.98, 2.19) | 1.48 (0.48, 4.58) |
| Hispanic | 1.70 (1.14, 2.52)** | 2.05 (0.82, 5.10) | 0.57 (0.37, 0.88)** | 0.45 (0.07, 2.98) |
| Other | 3.29 (1.80, 5.99)*** | 3.50 (0.82, 14.9) | 0.22 (0.12, 0.40)*** | 1.42 (0.33, 6.10) |
| Good NKB vs. other4 | 1.69 (1.29, 2.20)*** | 2.63 (1.58, 4.36)*** | 0.70 (0.49, 1.01) | 0.42 (0.17, 1.08) |
| FL use5 | 1.38 (1.06, 1.80)* | 1.74 (1.05, 2.90)* | 0.49 (0.39, 0.62)** | 0.91 (0.45, 1.82) |
| Hypertension | ||||
| Women (ref: men) | 1.57 (1.23, 2.00)*** | 1.43 (0.97, 2.12) | 0.63 (0.48, 0.84)** | 0.67 (0.43, 1.05) |
| Age (ref: 20–34 y) | ||||
| 35–49 | 1.51 (1.12, 2.05)** | 0.72 (0.32, 1.65) | 0.86 (0.62, 1.18) | 0.91 (0.39, 2.12) |
| 50–65 | 2.50 (1.84, 3.39)*** | 1.12 (0.47, 2.65) | 0.38 (0.27, 0.53)*** | 0.44 (0.21, 0.94)* |
| Education (ref: <high school) | ||||
| High school | 0.98 (0.65, 1.48) | 0.73 (0.44, 1.20) | 0.68 (0.48, 0.98)* | 0.95 (0.51, 1.77) |
| >High school | 1.51 (0.99, 2.31) | 0.93 (0.51, 1.71) | 0.31 (0.21, 0.44)*** | 0.55 (0.28, 1.09) |
| Income (ref: low) | ||||
| Middle | 1.11 (0.75, 1.65) | 1.29 (0.67, 2.46) | 0.89 (0.67, 1.17) | 1.46 (0.76, 2.82) |
| High | 1.40 (0.94, 2.08) | 1.58 (0.78, 3.17) | 0.81 (0.52, 1.26) | 0.71 (0.32, 1.59) |
| Race/ethnicity (ref: NH white) | ||||
| NH black | 0.47 (0.23, 0.92)* | 1.03 (0.50, 2.09) | 1.41 (0.95, 2.11) | 1.78 (0.95, 3.34) |
| Hispanic | 1.57 (1.03, 2.38)* | 2.91 (1.07, 7.89)* | 0.53 (0.32, 0.86)* | 1.00 (0.37, 2.68) |
| Other | 3.69 (2.01, 6.79)*** | 1.01 (0.44, 2.35) | 0.28 (0.18, 0.42)*** | 0.46 (0.06, 3.29) |
| Good NKB vs. other4 | 1.67 (1.30, 2.14)*** | 2.20 (1.33, 3.63)** | 0.71 (0.49, 1.04) | 0.43 (0.22, 0.88)* |
| FL use5 | 1.56 (1.19, 2.06)** | 0.95 (0.56, 1.63) | 0.49 (0.39, 0.61)* | 0.53 (0.31, 0.93)* |
Values are OR (95% CI). Variables mutually adjusted for. Further adjusted variables included BMI, survey year, region, and urbanization. * < 0.05, **P < 0.01, ***P < 0.001.
Selected chronic disease(s), including at least 1 of diseases: hypertension, diabetes, heart diseases, hypercholesterolemia, or cancer.
Household income levels were assessed using poverty income ratio (PIR) and were categorized as 0–130% (low income: food stamp eligible), 131–350% (middle income), and ≥350% (high income).
NKB score ≥ 80th percentile was treated as good NKB.
FL use was defined as using all of label information for ingredient, short phrase, nutrition panel, serving size, and health benefits of foods.
NA, not available.
Discussion
Using nationally representative data, we detected several associations of note. First, we found that U.S. adults with diet-related chronic diseases reported higher diet quality, lower energy intake, and higher FV intake than other Americans. Diabetes patients were most likely to have higher HEI and consume less sugar than nondiabetes patients. These patients may be benefitting from the advice commonly offered by health professionals to improve their diet by limiting energy-dense food consumption and consuming more nutrient-dense foods (19).
Second, we found that the differences in diet between patients and nonpatients were relatively small and, although better than nonpatients’, patients’ dietary intakes were far from nutritional recommendations. For example, the average HEI was 52.3 and only 1.2% had a HEI >80, much lower on average than the diet recommended by the USDA (18). There was no significant difference in cholesterol or sodium intake between patients and nonpatients. We found no substantial differences in diet quality between hypertensive patients and nonpatients.
Compared with people with only hypertension, however, those with only diabetes were 87% less likely to report low diet quality. In other words, patients with hypertension were less likely to have high diet quality than diabetes patients. Thus, the diagnosis of diet-related chronic diseases may not be associated with healthier dietary behavior among certain populations (e.g. hypertensive patients). Previous research also indicates that only one-third of hypertensive patients have received their doctors’ advice to reduce salt intake (20). Health professionals should encourage people with chronic diseases to adopt a healthier diet for control of blood pressure and other diet-related chronic health conditions.
Third, we found considerable racial/ethnic differences. Among patients with 1 or more chronic diseases, NH blacks were 92% more likely than NH whites to have low diet quality; NH blacks had higher intakes of fat and cholesterol regardless of their chronic disease status. Reducing dietary and health disparities is a high priority in the US (21). Diet-related health disparities often reflect differences in chronic diseases. NH blacks have higher mortality from diet-related chronic diseases than NH whites (21). NH blacks are less likely to meet USDA guidelines than NH whites and have the lowest FV consumption among U.S. racial/ethnic groups (22). We found that only 36% of NH blacks reported consuming ≥5 servings/d of FV compared with 42% for NH whites and 54% for other groups. The effects of SES on health disparities have been suggested to be stronger than that of race/ethnicity (24), with higher SES being associated with a greater likelihood of meeting dietary guidelines (22). Our study found that patients with higher education and/or incomes were more likely to have high diet quality. Overall, these data support the need for greater efforts to promote healthy eating targeting racial/ethnic minorities, particularly NH blacks and lower SES groups, especially among those with diet-related chronic diseases.
Fourth, the disease-diet association was only significant among people with good NKB. Those with good NKB were more likely to report high diet quality, and this association was somewhat stronger among patients than among nonpatients. Patients with good NKB were 57% less likely to have low diet quality than patients with poor NKB. Previous research has also shown a positive association of NKB with diet and health (25, 26). NKB may affect whether patients adopt desirable dietary changes.
Fifth, we observed a stronger association between diabetes and high HEI among FL users than among non-FL users. FL users reported higher diet quality than non-FL users, both among patients and nonpatients. A significant association between hypercholesterolemia and high diet quality was observed only among people with good NKB and FL use. This indicates that NKB and FL use may modify the disease-diet association.
Sixth, we found that patients did not consume more FV than nonpatients. Only 42% of them had ≥5 FV servings/d and only 15% had ≥2 servings/d of fruits and ≥3 servings/d of vegetables. Low FV consumption is associated with increased risks of diet-related chronic diseases, such as diabetes (27, 28) and cardiovascular disease (29, 30). Consuming adequate FV among patients may help control their health conditions. However, only a small proportion of Americans meet the USDA guidelines for FV intake (22, 31). More vigorous efforts are needed to promote FV consumption. Health professionals can and should play a more active role in encouraging their patients to consume adequate FV.
Our study has several important strengths. First, it is based on nationally representative data collected from a large sample. To our knowledge, these data are the only available datasets that provide comprehensive, nationally representative data on the variables studied. Some of the measures were not collected in other, more recent national surveys, such as the NHANES. Furthermore, recent NHANES data show that Americans’ diet quality has changed little since the mid-1990s, when the CSFII data were gathered (13–15). Second, it included rich data regarding participants’ nutrition knowledge and behaviors. Third, we used HEI-2005 to assess overall diet quality. These findings may help shed light on the underlying causes of health disparities in the US and help guide future intervention efforts.
Our study also has some limitations. First, cross-sectional data cannot ascertain when and how much the patients changed their diet because of the disease diagnosis. Second, the data were collected 10 y ago. However, they are the only available national surveys that provided all the measures we were interested in (e.g. NKB), which were not collected in other, more recent national surveys, such as the NHANES. Third, two 24-h recalls may not be adequate to assess individuals’ usual diet (32). Fourth, the lack of healthcare information would not allow us to adjust for its potential effect, although our analysis controlled for income and education. Fifth, diet is not the only risk factor for the chronic diseases examined. Other factors, such as genetics and environmental factors, also affect chronic disease but were not available in the datasets. Finally, self-reported medical diagnoses and dietary intakes are subject to reporting errors.
In conclusion, Americans diagnosed with diet-related chronic diseases have somewhat higher diet quality than those without these diagnoses. However, many of the differences are small, and their cholesterol and sodium intakes are similar to those of nonpatients. Patients with better nutrition knowledge or FL use are more likely to report high diet quality. The protective effects of nutrition knowledge and FL use are stronger among patients than among nonpatients. Further, patients’ dietary intakes remain far from desirable compared with the Dietary Guidelines for Americans. The importance of nutrition knowledge and FL use should be highlighted in health promotion programs, especially among patients with or at increased risk of chronic, diet-related diseases. Health professionals can and should play a more active role in encouraging their patients to adopt a more healthful diet.
Supplementary Material
Acknowledgments
We thank Dr. May Beydoun for her technical assistance in working on the data sets. X.C. and Y.W. had full access to the data in the study and take full responsibility for the integrity of the data and the accuracy of the data analysis; X.C. and Y.W. analyzed the data and wrote the paper; L.C. and L.S. revised the manuscript; and Y.W. had primary responsibility for final content and related funding and administration issues. All authors read and approved the final manuscript.
Footnotes
Supported by the NIH/The National Institute of Diabetes and Digestive and Kidney Diseases (R01DK81335-01A1).
Supplemental Text 1 and Supplemental Tables 1 and 2 are available from the “Online Supporting Material” link in the online posting of the article and from the same link in the online table of contents at jn.nutrition.org.
Abbreviations used: CSFII, Continuing Survey of Food Intakes by Individuals; DHKS, Diet and Health Knowledge Survey; FL, food label; FV, fruits and vegetables; HEI, Healthy Eating Index; NH, non-Hispanic; NKB, nutrition knowledge and beliefs; PIR, poverty income ratio; SES, socioeconomic status.
Literature Cited
- 1.McCullough ML, Feskanich D, Rimm EB, Giovannucci EL, Ascherio A, Variyam JN, Spiegelman D, Stampfer MJ, Willett WC. Adherence to the Dietary Guidelines for Americans and risk of major chronic disease in men. Am J Clin Nutr. 2000;72:1223–31 [DOI] [PubMed] [Google Scholar]
- 2.McCullough ML, Feskanich D, Stampfer MJ, Rosner BA, Hu FB, Hunter DJ, Variyam JN, Colditz GA, Willett WC. Adherence to the Dietary Guidelines for Americans and risk of major chronic disease in women. Am J Clin Nutr. 2000;72:1214–22 [DOI] [PubMed] [Google Scholar]
- 3.Guo X, Warden BA, Paeratakul S, Bray GA. Healthy Eating Index and obesity. Eur J Clin Nutr. 2004;58:1580–6 [DOI] [PubMed] [Google Scholar]
- 4.Chobanian AV, Bakris GL, Black HR, Cushman WC, Green LA, Izzo JL, Jr, Jones DW, Materson BJ, Oparil S, et al. The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure: the JNC 7 report. JAMA. 2003;289:2560–72 [DOI] [PubMed] [Google Scholar]
- 5.CDC National diabetes fact sheet: general information and national estimates on diabetes in the United States, 2007. Atlanta: US Department of Health and Human Services, CDC; 2008 [Google Scholar]
- 6.Lewis JE, Arheart KL, LeBlanc WG, Fleming LE, Lee DJ, Davila EP, Caban-Martinez AJ, Dietz NA, McCollister KE, et al. Food label use and awareness of nutritional information and recommendations among persons with chronic disease. Am J Clin Nutr. 2009;90:1351–7 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Post RE, Mainous AG III, Diaz VA, Matheson EM, Everett CJ. Use of the nutrition facts label in chronic disease management: results from the National Health and Nutrition Examination Survey. J Am Diet Assoc. 2010;110:628–32 [DOI] [PubMed] [Google Scholar]
- 8.Wang Y, Wang QJ. The prevalence of prehypertension and hypertension among US adults according to the new joint national committee guidelines: new challenges of the old problem. Arch Intern Med. 2004;164:2126–34 [DOI] [PubMed] [Google Scholar]
- 9.Wang Y, Beydoun MA. The obesity epidemic in the United States: gender, age, socioeconomic, racial/ethnic, and geographic characteristics: a systematic review and meta-regression analysis. Epidemiol Rev. 2007;29:6–28 [DOI] [PubMed] [Google Scholar]
- 10.Beydoun MA, Wang Y. Do nutrition knowledge and beliefs modify the association of socio-economic factors and diet quality among US adults? Prev Med. 2008;46:145–53 [DOI] [PubMed] [Google Scholar]
- 11.Robb CA, Huston SJ, Finke MS. The mitigating influence of time preference on the relation between smoking and BMI scores. Int J Obes (Lond). 2008;32:1670–7 [DOI] [PubMed] [Google Scholar]
- 12.Tippett KS, Yasmin SC. Design and operation: the Continuing Survey of Food Intakes by Individuals and the Diet and Health Knowledge Survey, 1994–1996. USDA, Agricultural Research Service, Nationwide Food Surveys Report, No. 96–1; 1998 [Google Scholar]
- 13.Beydoun MA, Gary TL, Caballero BH, Lawrence RS, Cheskin LJ, Wang Y. Ethnic differences in dairy and related nutrient consumption among US adults and their association with obesity, central obesity, and the metabolic syndrome. Am J Clin Nutr. 2008;87:1914–25 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.USDA. Diet quality of Americans in 1994–96 and 2001–02 as measured by the Healthy Eating Index-2005 [cited 2010 May 24]. Available from: http://www.cnpp.usda.gov/Publications/NutritionInsights/Insight37.pdf.
- 15.USDA. Diet quality of low-income and higher income Americans in 2003–04 as measured by the Healthy Eating Index-2005 [cited 2010 May 22]. Available from: http://www.cnpp.usda.gov/Publications/NutritionInsights/Insight42.pdf.
- 16.USDA. Healthy eating index [cited 2010 Apr 3]. Available from: http://www.cnpp.usda.gov/HealthyEatingIndex.htm.
- 17.Gao SK, Beresford SA, Frank LL, Schreiner PJ, Burke GL, Fitzpatrick AL. Modifications to the Healthy Eating Index and its ability to predict obesity: the Multi-Ethnic Study of Atherosclerosis. Am J Clin Nutr. 2008;88:64–9 [DOI] [PubMed] [Google Scholar]
- 18.Basiotis PP, Carlson A, Gerrior SA, Juan WY, Lino M. The Healthy Eating Index: 1999–2000. Washington, DC: USDA, Center for Nutrition Policy and Promotion. CNPP-12; 2002 [Google Scholar]
- 19.Ornish D, Scherwitz LW, Billings JH, Brown SE, Gould KL, Merritt TA, Sparler S, Armstrong WT, Ports TA, et al. Intensive lifestyle changes for reversal of coronary heart disease. JAMA. 1998;280:2001–7 [DOI] [PubMed] [Google Scholar]
- 20.Booth AO, Nowson CA. Patient recall of receiving lifestyle advice for overweight and hypertension from their general practitioner. BMC Fam Pract. 2010;11:8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Satia JA. Diet-related disparities: understanding the problem and accelerating solutions. J Am Diet Assoc. 2009;109:610–5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Casagrande SS, Wang Y, Anderson C, Gary TL. Have Americans increased their fruit and vegetable intake? The trends between 1988 and 2002. Am J Prev Med. 2007;32:257–63 [DOI] [PubMed] [Google Scholar]
- 23.Akil L, Ahmad HA. Effects of socioeconomic factors on obesity rates in four southern states and Colorado. Ethn Dis. 2011;21:58–62 [PMC free article] [PubMed] [Google Scholar]
- 24.Adler NE, Rehkopf DHUS. Disparities in health: descriptions, causes, and mechanisms. Annu Rev Public Health. 2008;29:235–52 [DOI] [PubMed] [Google Scholar]
- 25.Patterson RE, Kristal AR, White E. Do beliefs, knowledge, and perceived norms about diet and cancer predict dietary change? Am J Public Health. 1996;86:1394–400 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Trudeau E, Kristal AR, Li S, Patterson RE. Demographic and psychosocial predictors of fruit and vegetable intakes differ: implications for dietary interventions. J Am Diet Assoc. 1998;98:1412–7 [DOI] [PubMed] [Google Scholar]
- 27.Ford ES, Mokdad AH. Fruit and vegetable consumption and diabetes mellitus incidence among U.S. adults. Prev Med. 2001;32:33–9 [DOI] [PubMed] [Google Scholar]
- 28.Feskens EJ, Virtanen SM, Rasanen L, Tuomilehto J, Stengard J, Pekkanen J, Nissinen A, Kromhout D. Dietary factors determining diabetes and impaired glucose tolerance. A 20-year follow-up of the Finnish and Dutch cohorts of the Seven Countries Study. Diabetes Care. 1995;18:1104–12 [DOI] [PubMed] [Google Scholar]
- 29.Genkinger JM, Platz EA, Hoffman SC, Comstock GW, Helzlsouer KJ. Fruit, vegetable, and antioxidant intake and all-cause, cancer, and cardiovascular disease mortality in a community-dwelling population in Washington County, Maryland. Am J Epidemiol. 2004;160:1223–33 [DOI] [PubMed] [Google Scholar]
- 30.Hung HC, Joshipura KJ, Jiang R, Hu FB, Hunter D, Smith-Warner SA, Colditz GA, Rosner B, Spiegelman D, et al. Fruit and vegetable intake and risk of major chronic disease. J Natl Cancer Inst. 2004;96:1577–84 [DOI] [PubMed] [Google Scholar]
- 31.Serdula MK, Gillespie C, Kettel-Khan L, Farris R, Seymour J, Denny C. Trends in fruit and vegetable consumption among adults in the United States: behavioral risk factor surveillance system, 1994–2000. Am J Public Health. 2004;94:1014–8 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Tooze JA, Vitolins MZ, Smith SL, Arcury TA, Davis CC, Bell RA, DeVellis RF, Quandt SA. High levels of low energy reporting on 24-hour recalls and three questionnaires in an elderly low-socioeconomic status population. J Nutr. 2007;137:1286–93 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.