- Email: [email protected]
Fat-Related Dietary Behaviors of Adult Puerto Ricans, with and without Diabetes, in New York City
RESEARCH Current Research
Fat-Related Dietary Behaviors of Adult Puerto Ricans, with and without Diabetes, in New York City THOMAS A. MELNIK, DrPH; MAUREEN M. SPENCE, MS, RD; AKIKO S. HOSLER, PhD
ABSTRACT Objective To assess the fat-related dietary behaviors of adult Puerto Ricans with and without diagnosed diabetes, living in New York City. Design A random-digit-dialing telephone survey was conducted following Behavioral Risk Factor Surveillance System procedures. Dietary behavior was assessed using a brief Fat-Related Diet Habits Questionnaire, in which higher scores indicated higher fat intake. Subjects/setting A total of 1,304 adult Puerto Ricans living in New York City were interviewed. Diabetes status was assessed using standard Behavioral Risk Factor Surveillance System questions. Statistical analyses performed Weighted analyses using SUDAAN software for complex surveys were done, and t tests were used to assess differences in mean fat-related dietary score by sociodemographic and health characteristics. Age-adjusted least-squared means were used to compare scores between those with and without diabetes. Linear regression was used to model characteristics associated with fat-related dietary score. Results Fat-related dietary score was lower among those with diabetes and varied by population and health characteristics. Age-adjusted scores were significantly lower for those with diabetes who were younger, less educated, obese, or physically active. In the regression model, family history, weight, and exercise interacted with diabetes T. A. Melnik is a research scientist and director, Chronic Disease and Risk Factor Surveillance Program, and A. S. Hosler is a research scientist and director, Diabetes Surveillance and Evaluation, Bureau of Chronic Disease Epidemiology and Surveillance; and M. M. Spence is a health program administrator and coordinator, Diabetes Prevention and Control Program, Bureau of Chronic Disease Services, all with the New York State Department of Health, Albany. A. S. Hosler is also with the Department of Epidemiology and Biostatistics, the University at Albany, Albany, NY. Address correspondence to: Thomas A. Melnik, DrPH, New York State Department of Health, Bureau of Chronic Disease Epidemiology and Surveillance, Corning Tower, Room 565, Empire State Plaza, Albany, NY 12237-0679. E-mail: [email protected] Copyright © 2006 by the American Dietetic Association. 0002-8223/06/10609-0004$32.00/0 doi: 10.1016/j.jada.2006.06.007
© 2006 by the American Dietetic Association
status. Those with diabetes were significantly more likely to modify meat consumption practices (eg, remove skin or trim fat) to reduce fat compared with those without diabetes. Conclusions New York City Puerto Ricans with diabetes are somewhat more likely to engage in behaviors to reduce fat compared with those without diabetes. Targeted, culturally sensitive nutrition education and counseling emphasizing lower-fat food choices and other fat-reducing behaviors can help reduce risk and control diabetes. Education messages should be tailored to the individual’s diabetes status and other health and sociodemographic characteristics. J Am Diet Assoc. 2006;106:1419-1425.
T
he Hispanic population, the largest and fastest growing minority group in the United States, includes individuals from diverse ethnic origins (1,2). Puerto Ricans are the largest Hispanic group in New York state and represent approximately 37% of New York City’s Hispanic population (3). Diabetes is a leading cause of mortality, morbidity, and disability that disproportionately affects Hispanics in the United States (4). Diagnosed diabetes prevalence among adult US Hispanics (8.0%) is twice that of non-Hispanic whites (4.0%) (5). It has recently been reported that the prevalence of diagnosed diabetes among New York City Puerto Rican adults is 11.3% and is significantly related to age, obesity, family history of diabetes, and low educational attainment (6). Preventing and controlling diabetes among diverse ethnic populations requires a better understanding of preventable risk factors, including reducing fat in the diet. An evidenced-based technical report and recent position statement of the American Diabetes Association concluded that reduced-fat diets maintained over the longterm contribute to weight loss and improvement in dyslipidemia in diabetes (7,8). The impact of reduced dietary fat intake in diabetes prevention seems related primarily to energy balance, but dietary fat may also be an important determinant of diabetes risk independent of energy intake (7). Because high-fat diets generally include a high intake of saturated fats (9), fat reduction is likely to have a concomitant impact on lowering saturated fat, a primary dietary goal for individuals with diabetes. Structured programs emphasizing lifestyle changes including reduced fat intake are recommended to prevent and treat diabetes, taking into account individual circumstances and cultural and ethnic preferences (8).
Journal of the AMERICAN DIETETIC ASSOCIATION
1419
This study was conducted to assess the fat-related dietary behaviors of adult Puerto Ricans living in New York City, compare dietary behaviors of those with and without diagnosed diabetes, and examine the relationship between fat-related diet behaviors and other factors in the population. METHODS Study Population Respondents were selected using a random-digit-dialing telephone survey methodology adhering closely to procedures established by the Centers for Disease Control and Prevention to conduct the annual state-based Behavioral Risk Factor Surveillance System (BRFSS) (10,11). The sampled population included New York City residents aged 18 years and older who identified themselves as being of Puerto Rican ethnicity. Because Puerto Ricans represent only approximately 10% of the New York City population, a dual-frame sampling design was utilized to improve sampling efficiency. In one frame, telephone exchanges with higher concentrations of Puerto Rican residents were sampled at higher rates. The other frame sampled New York City telephone numbers with associated surnames likely to be Hispanic. Selected households from each frame were screened for the presence of one or more Puerto Rican adults, and one was then randomly selected for the interview. Respondents were identified as having diabetes if they answered affirmatively to the standard BRFSS question, “Have you ever been told by a doctor that you have diabetes?” All respondents reporting diagnosed diabetes were retained in the sample. Respondents without diabetes were sampled at a rate to yield approximately equal numbers of respondents with and without diabetes in the final sample. Female respondents determined to have diabetes only during pregnancy were subsequently classified as nondiabetic for the purpose of the study. Interviews conducted with both English- and Spanishspeaking interviewers were conducted July 1999 through June 2000. A variation of the standard BRFSS introductory script was used to inform respondents of the purpose of the survey, their rights, confidentiality, and the availability of a toll-free number to call for further information. Tacit consent was obtained based on the respondent’s willingness to proceed with the interview according to standard BRFSS protocol. This level of consent was acceptable because of the negligible risk to adult respondents and their ability to not answer specific questions or end the interview at any time. The New York State Department of Health’s Institutional Review Board reviewed and approved the study protocol (study number 98-2-07). A total of 1,304 interviews were completed. The response rate (percent of eligible units completing interviews) was 45.1% and the cooperation rate (percent of eligible units actually contacted and interviews completed) was 62.8%. Questionnaire and Dietary Assessment The study questionnaire, available in both English and Spanish languages, consisted primarily of standard BRFSS questions to determine respondent age, sex, edu-
1420
September 2006 Volume 106 Number 9
Avoid fat as a flavoring Eat bread/rolls without butter or margarine Eat vegetables without butter or margarine Eat potatoes without butter, margarine, or sour cream Avoid fried foods Eat fried chicken Eat fried potatoes like French fries/hash browns Modify meats to be low in fat Take skin off chicken Trim visible fat from red meat Substitute fat-modified food products Drink 1% or skim milk Use low-calorie (diet) salad dressings Use low-fat/nonfat mayonnaise Replace high-fat foods with fruits and vegetables Eat fruit for dessert Eat vegetable or fruit for snacks Figure. Subscale factors used in a brief Fat-Related Diet Habits Questionnaire. cational attainment, health care coverage, smoking status, and exercise. Other questionnaire topics included and reported here were: birth place, principal language spoken in the household, and family history of diabetes. Self-reported height and weight were used to calculate the body mass index (BMI), defined as kg/m2. Fat-related dietary behavior was assessed using a brief version of the Fat-Related Diet Habits Questionnaire. Details on the development and validation of the FatRelated Diet Habits Questionnaire have been published (12,13). The Fat-Related Diet Habits Questionnaire is a behavioral measure of dietary patterns related to selecting and preparing low-fat diets. The Fat-Related Diet Habits Questionnaire scoring system correlates positively with fat intake: higher scores correspond with higher fat intake. The brief Fat-Related Diet Habits Questionnaire has been used in random-digit-dialing telephone surveys and found useful in investigating fatrelated dietary behavior (14-17). The Fat-Related Diet Habits Questionnaire and its brief version have previously been applied in studies involving diverse populations, including Hispanics (18-21). Respondents in this study were asked to report their consumption of 12 items during the 3 months prior to the interview. The items were then combined into five subscale factors as follows: avoid fat as a flavoring (eating foods without butter, margarine, or sour cream); avoid fried foods (fried chicken, fried potatoes); modify meats to be low in fat (take skin off chicken and trim visible fat from meat); substitute fat-modified products (reduced-fat milk, low-calorie dressings and mayonnaise); and replace high-fat foods (eat fruits and vegetables for dessert or snacks). The five subscales and their respective individual items are shown in the Figure. Responses to the items were coded on a 4-point scale (usually or always, often, sometimes, rarely, or never), and were scored 1 through 4 to correlate positively with fat intake for respective items on the questionnaire. Subscale scores were calculated as the mean of the subscale items. A summary fat-related dietary score was calculated as the mean of the five subscales.
A number of steps were taken to improve survey administration and response rates. The questionnaire was pretested with 10 Puerto Rican adults at two New York City sites where health services are provided and was found to work well in this population. The introduction to the questionnaire clearly identified the survey as targeted toward the health concerns of Puerto Ricans in New York City. In addition, public service announcements about the survey were broadcast on local radio stations with programming for Hispanic audiences. Statistical Analysis Sample weights were calculated to account for the probabilities of selection. The sample was adjusted to the age and sex distribution of the New York City Puerto Rican population using the most recently available source of data specific to this population at the time of the study (22). All analyses were conducted using SUDAAN software for the calculation of variance estimates and conducting statistical tests for complex survey designs (23). Descriptive analysis was conducted to determine the sociodemographic and health-related characteristics of the population. t tests were used to assess statistical differences in mean fat-related dietary score by the sociodemographic and health-related characteristics of the sample. Because of the older mean age of respondents with diabetes, age-adjusted least squared means from linear regression analysis were used to compare mean fat-related dietary score between those with and without diabetes. Linear regression analysis was also used to model the characteristics associated with fat-related dietary score in the population using standard modeling procedures (24). The independent variables were dichotomized and those contributing to the model were retained as main effects. Variables considered in the model included diabetes status (1⫽diabetes, 2⫽non-diabetes), sex (1⫽male, 2⫽female), educational attainment (1⫽ⱕhigh school, 2⫽⬎high school), weight status (1⫽not obese [BMI ⬍30], 2⫽obese [BMI ⱖ30]), exercise (1⫽active, 2⫽sedentary [defined as no non–work-related physical activity or exercise in the past month]), and family history of diabetes (1⫽yes, 2⫽no). Age was also included in the model as a continuous main-effect variable. Two-way interactions of the main effects with diabetes status were tested, and those found to contribute to the association with fatrelated dietary score were retained in the final model. A total of 1,088 with complete information on these variables were included in the final model. RESULTS Table 1 shows the sociodemographic and health characteristics of the sample. Puerto Ricans were relatively young and low in educational attainment. Approximately half reported that their birth place was Puerto Rico. More than half spoke English at home. Rates of overweight, obesity, smoking, and inactivity were high. Twenty percent reported not having health coverage. Nearly 40% had a family history of diabetes and 11.3% had diagnosed diabetes. Mean fat-related dietary score by sociodemographic and health characteristics is shown in Table 1 and the
Table 1. Mean fat-related dietary score by sociodemographic and health-related characteristics of adult New York City Puerto Ricans, 2000 Fat-related Dietary Score Characteristic
na
%b
MeanⴞSEcd
Total Age (y) 18-44 45-64 ⱖ65 Sex Male Female Educational attainment ⬍High school High school ⬎High school Birth place United States Puerto Rico Household language English Spanish Weight statuse Not overweight or obese Overweight Obese Smoking status Nonsmoker Smoker Exercise status Active Sedentary Health care coverage Medicaid Medicare Other None Diabetes family history Yes No Diagnosed diabetes Yes No
1,304
100.0
2.29⫾0.03
556 482 245
63.9 27.3 8.8
2.37⫾0.04y 2.23⫾0.08y 1.92⫾0.08z
447 857
44.2 55.8
2.38⫾0.06y 2.21⫾0.04z
536 313 364
44.3 21.9 33.8
2.33⫾0.06y 2.34⫾0.07y 2.25⫾0.05y
468 790
52.9 46.1
2.38⫾0.05y 2.19⫾0.05z
628 661
60.6 39.4
2.33⫾0.05y 2.21⫾0.05y
393 472 346
39.2 41.5 19.3
2.28⫾0.07y 2.25⫾0.05y 2.35⫾0.06y
966 327
70.8 29.2
2.26⫾0.03y 2.36⫾0.08y
641 656
53.8 46.2
2.25⫾0.02y 2.32⫾0.06z
372 220 484 208
28.0 8.9 42.9 20.2
1.99⫾0.08y 2.29⫾0.06z 2.28⫾0.06z 2.36⫾0.06z
643 613
38.2 61.8
2.18⫾0.06y 2.36⫾0.04z
606 698
11.3 88.7
2.07⫾0.03y 2.31⫾0.04z
a Unweighted sample size. The numbers within groups may not sum to the total sample of 1,304 because missing values and unknown responses were excluded from the analysis. b Weighted percent. c Mean based on fat-related dietary score items coded on a 4-point scale (usually or always, often, sometimes, rarely, or never), and scored 1 through 4 to correlate positively with fat intake for respective items in the questionnaire. d SE⫽standard error. e Not overweight or obese: Body mass index (BMI) ⬍25; overweight: BMI ⱖ25 and ⬍30; obese: BMI ⱖ30. yz Values with different superscripts within a grouping are significantly different from others in that group (P⬍0.05).
September 2006 ● Journal of the AMERICAN DIETETIC ASSOCIATION
1421
Table 2. Age-adjusted meana fat-related dietary score for adult New York City Puerto Ricans with and without diagnosed diabetes by socio-demographic and health-related characteristics, 2000 Age-adjusted Fat-related Dietary Score Characteristic
Diabetes
Nondiabetes
Independent variables P valueb
c
Total Age (y)d 18-44 45-64 ⱖ65 Sex Male Female Educational attainment ⱕHigh school ⬎High school Birth place United States Puerto Rico Household language English Spanish Weight statuse Nonobese Obese Smoking status Nonsmoker Smoker Exercise status Active Sedentary Health care coverage Medicaid Other or none Diabetes family history Yes No
4™™™™ mean⫾SE ™™™™3 2.17⫾0.05 2.31⫾0.04
0.07
2.27⫾0.05 2.10⫾0.04 1.84⫾0.06
2.38⫾0.04 2.26⫾0.10 1.98⫾0.16
0.05 0.18 0.44
2.18⫾0.06 2.16⫾0.06
2.39⫾0.07 2.24⫾0.05
0.07 0.24
2.19⫾0.05 2.18⫾0.09
2.38⫾0.06 2.24⫾0.05
0.03 0.64
2.29⫾0.06 2.12⫾0.06
2.35⫾0.06 2.22⫾0.07
0.54 0.10
2.23⫾0.06 2.11⫾0.06
2.32⫾0.06 2.29⫾0.07
0.36 0.08
2.20⫾0.05 2.19⫾0.04
2.27⫾0.04 2.45⫾0.06
0.33 ⬍0.01
2.12⫾0.07 2.30⫾0.07
2.30⫾0.04 2.33⫾0.09
0.06 0.74
2.05⫾0.06 2.30⫾0.05
2.27⫾0.03 2.35⫾0.07
⬍0.01 0.58
2.18⫾0.04 2.18⫾0.06
2.31⫾0.09 2.29⫾0.05
0.24 0.14
2.16⫾0.06 2.22⫾0.05
2.21⫾0.08 2.37⫾0.04
0.68 0.06
a Mean based on fat-related dietary score items coded on a 4-point scale (usually or always, often, sometimes, rarely or never), and scored 1 through 4 to correlate positively with fat intake for respective items in the questionnaire. b Comparing those with and without diabetes across rows. c SE⫽standard error. d Age-specific rates were not adjusted. e Nonobese: Body mass index (BMI) ⬍30; obese: BMI ⱖ30.
age-adjusted rates by diabetes status in Table 2. Unadjusted mean scores were significantly lower among women and among those who were aged 65 years and older, born in Puerto Rico, physically active, on Medicaid, have a family history of diabetes, and have diagnosed diabetes (Table 1). After age adjustment, the difference in fat-related dietary score between those with and without diabetes was somewhat attenuated (Table 2). Compared with respondents without diabetes, the age-adjusted mean dietary score for those with diabetes was significantly less (P⬍0.05) for those with a high school education or less, obesity, and those living an active lifestyle.
1422
September 2006 Volume 106 Number 9
Table 3. Linear regression model of characteristics associated with fat-related dietary score for adult New York City Puerto Ricans, 2000
Intercept Main effects Age (y) Sex (male vs female) Educational attainment (ⱕhigh school vs ⬎high school) Diabetes (yes vs no) Diabetes family history (yes vs no) Weight status (nonobese vs obese) Exercise status (active vs sedentary) Interaction effects Diabetes⫻family history Diabetes⫻weight status Diabetes⫻exercise status
 coefficient
P value
2.80
⬍0.01
⫺.01 .11
⬍0.01 0.20
.10 ⫺.27
0.14 ⬍0.01
⫺.19
0.07
⫺.21
0.02
⫺.06
0.45
.19 .26 ⫺.23
0.03 0.02 0.01
In the linear regression model, fat-related dietary score was inversely associated with age, and diabetes status had significant positive interactions with family history of diabetes and weight status (Table 3). As such, those without diabetes who were obese or had no family history of diabetes had higher scores. A significant negative interaction was observed between diabetes status and exercise status: those with diabetes and reporting active lifestyles had lower scores. The lowest mean subscale score was modifying meat, and those with diabetes were significantly more likely to modify meat compared to those without diabetes (Table 4). Significant differences in meat modifying behaviors between those with and without diabetes were observed for both individual subscale items: taking skin off chicken and trimming visible fat from red meat (data not shown). Avoiding fried foods also contributed to a lower-thanaverage score, although the difference between those with and without diabetes was not statistically significant in that regard. The highest mean subscale score was substituting fat-modified products, indicating that Puerto Rican adults living in New York City were least likely to make substitutions to lower fat intake, regardless of their diabetes status. Replacement of high-fat foods with fruits and vegetables also contributed to a higher-than-average score. Compared with those with diabetes, those without diabetes did not differ with respect to avoiding fat or fried foods, or to substitution or replacement to reduce fat consumption. DISCUSSION The results of this study show that dietary behaviors related to fat consumption in adult Puerto Ricans in New York City are associated with the sociodemographic and health characteristics of the population. The findings fur-
Table 4. Meana fat-related dietary subscale factor scores for adult New York City Puerto Ricans with and without diagnosed diabetes, 2000 Unadjusted Fat-related Dietary Score
Age-adjusted Fat-related Dietary Score
Subscale factor
Total
Diabetes
Non-diabetes
Total Avoid fat as a flavoring Avoid fried foods Modify meats to be low in fat Substitute fat-modified food products Replace high-fat foods with fruit and vegetables
4™™™™™™™™™™™™™™™™™ mean⫾SE c ™™™™™™™™™™™™™™™™™3 2.29⫾0.03 2.17⫾0.05 2.31⫾0.04 2.30⫾0.07 2.34⫾0.10 2.30⫾0.07 1.98⫾0.06 1.86⫾0.10 2.00⫾0.07 1.80⫾0.09 1.49⫾0.09 1.84⫾0.10 2.87⫾0.11 2.79⫾0.08 2.89⫾0.12 2.56⫾0.10 2.38⫾0.14 2.59⫾0.12
P valueb 0.07 0.73 0.33 0.01 0.46 0.27
a Mean based on fat-related dietary score items coded on a 4-point scale (usually or always, often, sometimes, rarely or never), and scored 1 through 4 to correlate positively with fat intake for respective items in the questionnaire. b Comparing those with and without diabetes across rows. c SE⫽standard error.
ther show that characteristics associated with fat-related dietary behaviors are dependent on diabetes status. Those with diabetes were somewhat more likely to engage in behaviors to lower fat compared to those without diabetes. Following adjustment for other factors, New York City Puerto Ricans who are diagnosed with diabetes and have a family history of diabetes or who are obese are at greater risk for increased fat consumption, whereas those without diabetes and active are at lower risk. Higher mean fat-related dietary scores (indicating higher fat intake) were observed as a result of less substitution with fat-modified food products and less replacement with high-fat foods, regardless of diabetes status. The lowest mean fat-related dietary scores (indicating lower fat intake) were observed for modifying meats to be low in fat, but those with diabetes had significantly lower scores in this regard compared with others. This study used a dual-frame telephone survey sampling approach to obtain population-based diet and health-related information from an urban minority population. The brief Fat-Related Diet Habits Questionnaire assessment method was selected because of its demonstrated usefulness in telephone surveys and its previous application in diverse population groups. It has shown to be useful in this study to make relative comparisons of fat-related dietary behaviors in New York City adult Puerto Ricans with respect to population and health characteristics, and to compare behaviors between those with and without diagnosed diabetes. Because the Fat-Related Diet Habits Questionnaire provides valid information related to dietary fat reduction behaviors, the information from this study can be applied directly to educational activities and interventions targeting specific behaviors to avoid, modify, replace, and substitute fatty food items in the diet of New York City Puerto Rican adults to prevent and control their risk for diabetes. These sampling and dietary methods may be adapted for use in other geographic areas and in other hard-to-reach population groups. Comparison with other studies using the Fat-Related Diet Habits Questionnaire is difficult because of differ-
ences in the study design, application of the Fat-Related Diet Habits Questionnaire, populations studied, and differences in reporting subscale scores. For example, in previous reports among northeast US Puerto Ricans and Hispanics from Rhode Island and Massachusetts, a longer 27-item Fat-Related Diet Habits Questionnaire was used (19,20). In a more recent study applying the brief Fat-Related Diet Habits Questionnaire among Mexicans from Washington state, subscale items specific to this population group were added to the instrument (21). Nonetheless, the findings with respect to the subscale scores show a consistent pattern among Hispanics. The lowest fat-related scores were reported with respect to modifying meat and avoidance of frying, and higher scores for replacement and substitution (19-21). However, compared with other Puerto Ricans living in the northeast United States, there is the suggestion that Puerto Ricans in New York City are more likely to avoid fat as a flavoring and to avoid fried foods (19). Lower fat-related scores (indicating lower fat intake) were associated with several sociodemographic characteristics of Puerto Ricans living in New York City, but only the inverse association with age remained significant in the regression model. Improvements in dietary score with respect to female sex, increasing age, and educational attainment have previously been reported (17). Loria and colleagues, using data from the Hispanic Health and Nutrition Examination Survey, reported generally lower fat and saturated fat consumption (expressed as percent of energy intake) with increasing age among mainland US Puerto Ricans (25). Polednak reported that mean fat intake (g/day) was associated with female sex, inversely associated with age, and not related to educational attainment among northeastern US Hispanics (26). There is considerable interest in the impact of immigration and acculturation on diet (27). Language as an indirect measure of acculturation was not significantly associated with fat-related dietary score. This is consistent with recent findings that Fat-Related Diet Habits Questionnaire scores among more acculturated Hispanics were not significantly different compared with those of
September 2006 ● Journal of the AMERICAN DIETETIC ASSOCIATION
1423
Hispanics who are less acculturated (21). Puerto Ricans born in Puerto Rico had significantly lower scores compared to US-born, but this was no longer observed after age adjustment, indicating that the differences were due to age and not birthplace. The findings of this study show that adult New York City Puerto Ricans with diabetes are more likely to engage in behaviors to reduce fat intake compared with those without diabetes. The statistically significant difference in mean fat-related dietary score for those with and without diabetes attenuated after age adjustment, indicating that the dietary behaviors are partially dependent on age differences in the two groups. Nonetheless, substantial differences in mean dietary score remained between those with and without diabetes after age adjustment, and significant differences between the two groups with respect to other health-related characteristics were observed. In the regression model, the association with dietary score remained for family history, weight, and exercise, with the direction of the effect being dependent on diabetes status. These findings point to a reduction in dietary risk behaviors among those with diabetes adopting an active lifestyle. Dietary risk increased for those without diabetes who are obese, or have no family history of diabetes. This suggests that New York City Puerto Ricans with diabetes are taking steps to control their disease, while the dietary behaviors together with other known risk factors are placing those without diabetes at increased risk. Of the five subscale factors, modifying meat to be low in fat and avoiding fried foods contributed to reduced fat scores among adult Puerto Ricans living in New York City, whereas they were less likely to substitute fatmodified food products and replace high-fat foods with fruits and vegetables to reduce fat consumption, on average. With the exception of modifying meats, those with and without diabetes did not differ with respect to subscale factor behaviors. Lower subscale scores for modifying meat and avoiding fried foods were also reported in previous studies of both Hispanics and non-Hispanics (19-21). Puerto Ricans living in New York City, in particular, would benefit from adopting behaviors to substitute and replace high-fat foods with low-fat products and food items. This would include increasing the consumption of fruits and vegetables for dessert and snacks, and the use of low-fat dressing, spreads, and dairy products. Previous studies of Hispanics in New York state and New York City, for example, show that whole milk is a major contributor to fat in the diet (28,29). Although reduced-fat milk represents a small percentage of the volume of milk available to inner–New York City Hispanics, availability is not a major problem (30). Thus, there is a need for greater consumer demand for reduced-fat milk in this population. Limitations to this study include the self-reported nature of the information obtained from telephone surveys and the possible bias in the measurements reported. However, an assessment of BRFSS measures concluded that most were at least moderately reliable and valid and many were highly reliable and valid (31). Second, estimates obtained from telephone surveys may also be biased due to relatively low response rates and to telephone noncoverage. However, bias associated with telephone
1424
September 2006 Volume 106 Number 9
noncoverage is likely to be minimal (32). In addition, the response (45.1%) and cooperation rates (62.8%) from this survey of inner-city minorities were higher than the statewide 2000 New York State BRFSS (32.9% and 34.7%, respectively) (33,34). Moreover, the Centers for Disease Control and Prevention has reported that bias in BRFSS data was not associated with response rates (35). CONCLUSIONS Fat-related dietary behaviors in New York City Puerto Rican adults are associated with diabetes status, sociodemographics, and other health characteristics. Culturally sensitive dietary treatments, education, and interventions to reduce fat intake are needed to prevent and control diabetes in this population. The emphasis should be placed on behaviors to substitute and replace high-fat foods with low-fat food items, including reduced-fat dairy products. These efforts should be targeted to New York City Puerto Ricans least likely to engage in dietary fatreducing behaviors, including men and those who are younger, sedentary, and obese. Nutrition and other health professionals can assist Puerto Ricans and others to adopt dietary behaviors to reduce fat consumption for diabetes prevention and control. The Diabetes Prevention Program’s Lifestyle Change Program, for example, provides guidance to reduce total fat and energy intake using basic dietary management concepts (36). The program offers educational materials to help participants identify low-fat food options, choose low-fat foods and flavorings, and modify recipes to reduce total fat. This research was supported by funding from the Association Schools of Public Health (S0742-18/19). References 1. Hanis CL, Hewett-Emmett D, Bertin TK, Schull WJ. Origins of US Hispanics: Implications for diabetes. Diabetes Care. 1991;14:618-627. 2. del Pinal JH. Hispanic Americans in the United States: Young, dynamic, and diverse. Stat Bulletin Metropolitan Insurance Co. 1996;77:2-13. 3. US Census Bureau. American FactFinder. Table PCT11. Hispanic or Latino by Specific Origin. Available at: http://factfinder.census.gov/servlet/ BasicFactsServlet. Accessed October 30, 2002. 4. Stern MP, Mitchell BD. Diabetes in Hispanic Americans. In: Harris MI, Cowie CC, Stern MP, Boyko EJ, Reiber GE, Bennet PH, eds. Diabetes in America. 2nd ed. Washington, DC: US Department of Health and Human Services, National Institutes of Health; 1995. DHHS publication no. (NIH) 95-1468. 5. Centers for Disease Control and Prevention. Selfreported prevalence of diabetes among Hispanics— United States, 1994-1997. MMWR. 1999;48:8-12. 6. Melnik TA, Hosler AS, Sekhobo JP, Duffy TP, Tierney EF, Englegau MM, Geiss LS. Diabetes prevalence among adult New York City Puerto Ricans, 2000. Am J Public Health. 2004;94:434-437. 7. Franz MJ, Bantle JP, Beebe CA, Brunzell JD, Chiasson J-L, Garg A, Holzmeister LA, Hoogwerf B, Mayer-Davis E, Mooradian AD, Purnell JQ, Wheeler
8. 9.
10.
11.
12.
13.
14. 15. 16.
17. 18.
19.
20.
21.
M. Evidenced-based nutrition principles and recommendations for the treatment and prevention of diabetes and related complications. Diabetes Care. 2002; 25:148-198. American Diabetes Association. Nutrition principles and recommendations in diabetes (Position Statement). Diabetes Care. 2004;27(suppl 1):S36-S46. US Department of Health and Human Services, US Department of Agriculture. Dietary Guidelines for Americans 2005. 6th ed. Washington, DC: US Government Printing Office; 2005. Available at: http://www. healthierus.gov/dietaryguidelines. Accessed March 2, 2005. Remington PL, Smith MY, Williamson DF, Anda RF, Gentry EM, Hogelin GC. Design, characteristics, and usefulness of state-based behavioral risk factor surveillance: 1981-87. Public Health Rep. 1988;103:366375. Behavioral Risk Factor Surveillance System User’s Guide. Atlanta, GA: US Department of Health and Human Services, Centers for Disease Control and Prevention; 1999. Kristal AR, Shattuck Al, Henry HJ. Patterns of dietary behavior associated with selecting diets low in fat: Reliability and validity of a behavioral approach to dietary assessment. J Am Diet Assoc. 1990;90:214220. Shannon J, Kristal AR, Curry SJ, Beresford SA. Application of a behavioral approach to measuring dietary change: The fat- and fiber-related diet behavior questionnaire. Cancer Epidemiol Biomarkers Prev. 1997;6:355-361. Neuhouser ML, Kristal AR, Patterson RE. Use of nutrition labels is associated with lower fat intake. J Am Diet Assoc. 1999;99:45-50. Levy L, Patterson RE, Kristal AR, Li SS. How well do consumers understand percentage daily value on food labels? Am J Health Promot. 2000;14:157-160. Patterson RE, Satia JA, Kristal AR, Neuhouser ML, Drewnowski AD. Is there a consumer backlash against the diet and health message? J Am Diet Assoc. 2001;101:43-46. Kristal AR, Hedderson M, Patterson RE, Neuhouser ML. Predictors of self-initiated, healthful dietary change. J Am Diet Assoc. 2001;101:762-766. Kristal AR, Shattuck AL, Patterson RE. Differences in fat-related dietary patterns between black, Hispanic, and white women: Results from the Women’s Health Trial Feasibility Study in Minority Populations. Public Health Nutr. 1999;2:253-262. Gans KM, Burkholder GJ, Upegui DI, Risica PM, Lasater TM, Fortunet R. Comparison of baseline fatrelated eating behaviors of Puerto Rican, Dominican, Colombian, and Guatemalan participants who joined a cholesterol education project. J Nutr Educ Behav. 2002;34:202-210. Gans KM, Burkholder GJ, Risica PM, Lasater TM. Baseline fat-related dietary behaviors of white, Hispanic, and black participants in a cholesterol screening and education project in New England. J Am Diet Assoc. 2003;103:699-706. Neuhouser ML, Thompson B, Coronado GD, Solomon
22.
23. 24. 25.
26. 27.
28. 29.
30.
31.
32.
33. 34. 35.
36.
CC. Higher fat intake and lower fruit and vegetable intakes are associated with greater acculturation among Mexicans living in Washington State. J Am Diet Assoc. 2004;104:51-57. Ruggles S, Sobek M, Fitch CA, Hall PK, Ronnander C. Integrated Public Use Microdata Series: Version 2.0. Minneapolis, MN: Historical Census Projects, University of Minnesota; 1997. Shah BV, Barnwell BG, Bieler GS. SUDAAN User’s Manual, Release 7.5. Research Triangle Park, NC: Research Triangle Institute; 1997. Kleinbaum DG, Kupper LL, Morgenstern H. Epidemiologic Research: Principles and Quantitative Methods. Lifetime Learning Publication. Belmont, CA; 1982. Loria CM, Bush TL, Carroll MD, Looker AC, McDowell MA, Johnson CL, Sempos CT. Macronutrient intake among adult Hispanics: A comparison of Mexican Americans, Cuban Americans, and mainland Puerto Ricans. Am J Public Health. 1995;85:684-689. Polednak AP. Use of selected high-fat foods by Hispanic adults in the northeastern US. Ethn Health. 1997;2:71-76. Satia-Abouta J, Patterson RE, Neuhouser ML, Elder J. Dietary acculturation: Application to nutrition research and dietetics. J Am Diet Assoc. 2002;102:11051118. Basch CE, Shea S, Zybert P. Food sources, dietary behavior, and the saturated fat intake of Latino children. Am J Public Health. 1992;82:810-815. Shea S, Melnik TA, Stein AD, Zansky SM, Maylahn C, Basch CE. Age, sex, educational attainment and race/ethnicity in relation to consumption of specific foods contributing to the atherogenic potential of diet. Prev Med. 1993;22:203-218. Weschler H, Basch CE, Zybert P, Lantigua R, Shea S. The availability of low-fat milk in an inner-city Latino community: Implications for nutrition education. Am J Public Health. 1995;85:1690-1692. Nelson DE, Holtzman D, Bolen J, Stanwyck CA, Mack KA. Reliability and validity of measures from the Behavioral Risk Factor Surveillance System (BRFSS). Soz Praventivmed. 2001;46(suppl 1):S3-S42. Ford ES. Characteristics of survey participants with and without a telephone: Findings from the Third National Health and Nutrition Examination Survey. J Clin Epidemiol. 1998;51:55-60. ORC Macro. 1999-2000 Diabetes Surveillance of Puerto Rican Adult Population in New York City. Technical Report. New York, NY: ORC Macro; 2002. Centers for Disease Control and Prevention. 2000 BRFSS Summary Data Quality Report. Atlanta, GA: Centers for Disease Control and Prevention; 2001. Centers for Disease Control and Prevention. Public health surveillance for behavioral risk factors in a changing environment: Recommendations from the Behavioral Risk Factor Surveillance Team. MMWR. 2003;52(No. RR-9):6. The Diabetes Prevention Program (DPP) Research Group. The Diabetes Prevention Program (DPP). Description of lifestyle intervention. Diabetes Care. 2002;25:2165-2171.
September 2006 ● Journal of the AMERICAN DIETETIC ASSOCIATION
1425
Fat-Related Dietary Behaviors of Adult Puerto Ricans, with and without Diabetes, in New York City THOMAS A. MELNIK, DrPH; MAUREEN M. SPENCE, MS, RD; AKIKO S. HOSLER, PhD
ABSTRACT Objective To assess the fat-related dietary behaviors of adult Puerto Ricans with and without diagnosed diabetes, living in New York City. Design A random-digit-dialing telephone survey was conducted following Behavioral Risk Factor Surveillance System procedures. Dietary behavior was assessed using a brief Fat-Related Diet Habits Questionnaire, in which higher scores indicated higher fat intake. Subjects/setting A total of 1,304 adult Puerto Ricans living in New York City were interviewed. Diabetes status was assessed using standard Behavioral Risk Factor Surveillance System questions. Statistical analyses performed Weighted analyses using SUDAAN software for complex surveys were done, and t tests were used to assess differences in mean fat-related dietary score by sociodemographic and health characteristics. Age-adjusted least-squared means were used to compare scores between those with and without diabetes. Linear regression was used to model characteristics associated with fat-related dietary score. Results Fat-related dietary score was lower among those with diabetes and varied by population and health characteristics. Age-adjusted scores were significantly lower for those with diabetes who were younger, less educated, obese, or physically active. In the regression model, family history, weight, and exercise interacted with diabetes T. A. Melnik is a research scientist and director, Chronic Disease and Risk Factor Surveillance Program, and A. S. Hosler is a research scientist and director, Diabetes Surveillance and Evaluation, Bureau of Chronic Disease Epidemiology and Surveillance; and M. M. Spence is a health program administrator and coordinator, Diabetes Prevention and Control Program, Bureau of Chronic Disease Services, all with the New York State Department of Health, Albany. A. S. Hosler is also with the Department of Epidemiology and Biostatistics, the University at Albany, Albany, NY. Address correspondence to: Thomas A. Melnik, DrPH, New York State Department of Health, Bureau of Chronic Disease Epidemiology and Surveillance, Corning Tower, Room 565, Empire State Plaza, Albany, NY 12237-0679. E-mail: [email protected] Copyright © 2006 by the American Dietetic Association. 0002-8223/06/10609-0004$32.00/0 doi: 10.1016/j.jada.2006.06.007
© 2006 by the American Dietetic Association
status. Those with diabetes were significantly more likely to modify meat consumption practices (eg, remove skin or trim fat) to reduce fat compared with those without diabetes. Conclusions New York City Puerto Ricans with diabetes are somewhat more likely to engage in behaviors to reduce fat compared with those without diabetes. Targeted, culturally sensitive nutrition education and counseling emphasizing lower-fat food choices and other fat-reducing behaviors can help reduce risk and control diabetes. Education messages should be tailored to the individual’s diabetes status and other health and sociodemographic characteristics. J Am Diet Assoc. 2006;106:1419-1425.
T
he Hispanic population, the largest and fastest growing minority group in the United States, includes individuals from diverse ethnic origins (1,2). Puerto Ricans are the largest Hispanic group in New York state and represent approximately 37% of New York City’s Hispanic population (3). Diabetes is a leading cause of mortality, morbidity, and disability that disproportionately affects Hispanics in the United States (4). Diagnosed diabetes prevalence among adult US Hispanics (8.0%) is twice that of non-Hispanic whites (4.0%) (5). It has recently been reported that the prevalence of diagnosed diabetes among New York City Puerto Rican adults is 11.3% and is significantly related to age, obesity, family history of diabetes, and low educational attainment (6). Preventing and controlling diabetes among diverse ethnic populations requires a better understanding of preventable risk factors, including reducing fat in the diet. An evidenced-based technical report and recent position statement of the American Diabetes Association concluded that reduced-fat diets maintained over the longterm contribute to weight loss and improvement in dyslipidemia in diabetes (7,8). The impact of reduced dietary fat intake in diabetes prevention seems related primarily to energy balance, but dietary fat may also be an important determinant of diabetes risk independent of energy intake (7). Because high-fat diets generally include a high intake of saturated fats (9), fat reduction is likely to have a concomitant impact on lowering saturated fat, a primary dietary goal for individuals with diabetes. Structured programs emphasizing lifestyle changes including reduced fat intake are recommended to prevent and treat diabetes, taking into account individual circumstances and cultural and ethnic preferences (8).
Journal of the AMERICAN DIETETIC ASSOCIATION
1419
This study was conducted to assess the fat-related dietary behaviors of adult Puerto Ricans living in New York City, compare dietary behaviors of those with and without diagnosed diabetes, and examine the relationship between fat-related diet behaviors and other factors in the population. METHODS Study Population Respondents were selected using a random-digit-dialing telephone survey methodology adhering closely to procedures established by the Centers for Disease Control and Prevention to conduct the annual state-based Behavioral Risk Factor Surveillance System (BRFSS) (10,11). The sampled population included New York City residents aged 18 years and older who identified themselves as being of Puerto Rican ethnicity. Because Puerto Ricans represent only approximately 10% of the New York City population, a dual-frame sampling design was utilized to improve sampling efficiency. In one frame, telephone exchanges with higher concentrations of Puerto Rican residents were sampled at higher rates. The other frame sampled New York City telephone numbers with associated surnames likely to be Hispanic. Selected households from each frame were screened for the presence of one or more Puerto Rican adults, and one was then randomly selected for the interview. Respondents were identified as having diabetes if they answered affirmatively to the standard BRFSS question, “Have you ever been told by a doctor that you have diabetes?” All respondents reporting diagnosed diabetes were retained in the sample. Respondents without diabetes were sampled at a rate to yield approximately equal numbers of respondents with and without diabetes in the final sample. Female respondents determined to have diabetes only during pregnancy were subsequently classified as nondiabetic for the purpose of the study. Interviews conducted with both English- and Spanishspeaking interviewers were conducted July 1999 through June 2000. A variation of the standard BRFSS introductory script was used to inform respondents of the purpose of the survey, their rights, confidentiality, and the availability of a toll-free number to call for further information. Tacit consent was obtained based on the respondent’s willingness to proceed with the interview according to standard BRFSS protocol. This level of consent was acceptable because of the negligible risk to adult respondents and their ability to not answer specific questions or end the interview at any time. The New York State Department of Health’s Institutional Review Board reviewed and approved the study protocol (study number 98-2-07). A total of 1,304 interviews were completed. The response rate (percent of eligible units completing interviews) was 45.1% and the cooperation rate (percent of eligible units actually contacted and interviews completed) was 62.8%. Questionnaire and Dietary Assessment The study questionnaire, available in both English and Spanish languages, consisted primarily of standard BRFSS questions to determine respondent age, sex, edu-
1420
September 2006 Volume 106 Number 9
Avoid fat as a flavoring Eat bread/rolls without butter or margarine Eat vegetables without butter or margarine Eat potatoes without butter, margarine, or sour cream Avoid fried foods Eat fried chicken Eat fried potatoes like French fries/hash browns Modify meats to be low in fat Take skin off chicken Trim visible fat from red meat Substitute fat-modified food products Drink 1% or skim milk Use low-calorie (diet) salad dressings Use low-fat/nonfat mayonnaise Replace high-fat foods with fruits and vegetables Eat fruit for dessert Eat vegetable or fruit for snacks Figure. Subscale factors used in a brief Fat-Related Diet Habits Questionnaire. cational attainment, health care coverage, smoking status, and exercise. Other questionnaire topics included and reported here were: birth place, principal language spoken in the household, and family history of diabetes. Self-reported height and weight were used to calculate the body mass index (BMI), defined as kg/m2. Fat-related dietary behavior was assessed using a brief version of the Fat-Related Diet Habits Questionnaire. Details on the development and validation of the FatRelated Diet Habits Questionnaire have been published (12,13). The Fat-Related Diet Habits Questionnaire is a behavioral measure of dietary patterns related to selecting and preparing low-fat diets. The Fat-Related Diet Habits Questionnaire scoring system correlates positively with fat intake: higher scores correspond with higher fat intake. The brief Fat-Related Diet Habits Questionnaire has been used in random-digit-dialing telephone surveys and found useful in investigating fatrelated dietary behavior (14-17). The Fat-Related Diet Habits Questionnaire and its brief version have previously been applied in studies involving diverse populations, including Hispanics (18-21). Respondents in this study were asked to report their consumption of 12 items during the 3 months prior to the interview. The items were then combined into five subscale factors as follows: avoid fat as a flavoring (eating foods without butter, margarine, or sour cream); avoid fried foods (fried chicken, fried potatoes); modify meats to be low in fat (take skin off chicken and trim visible fat from meat); substitute fat-modified products (reduced-fat milk, low-calorie dressings and mayonnaise); and replace high-fat foods (eat fruits and vegetables for dessert or snacks). The five subscales and their respective individual items are shown in the Figure. Responses to the items were coded on a 4-point scale (usually or always, often, sometimes, rarely, or never), and were scored 1 through 4 to correlate positively with fat intake for respective items on the questionnaire. Subscale scores were calculated as the mean of the subscale items. A summary fat-related dietary score was calculated as the mean of the five subscales.
A number of steps were taken to improve survey administration and response rates. The questionnaire was pretested with 10 Puerto Rican adults at two New York City sites where health services are provided and was found to work well in this population. The introduction to the questionnaire clearly identified the survey as targeted toward the health concerns of Puerto Ricans in New York City. In addition, public service announcements about the survey were broadcast on local radio stations with programming for Hispanic audiences. Statistical Analysis Sample weights were calculated to account for the probabilities of selection. The sample was adjusted to the age and sex distribution of the New York City Puerto Rican population using the most recently available source of data specific to this population at the time of the study (22). All analyses were conducted using SUDAAN software for the calculation of variance estimates and conducting statistical tests for complex survey designs (23). Descriptive analysis was conducted to determine the sociodemographic and health-related characteristics of the population. t tests were used to assess statistical differences in mean fat-related dietary score by the sociodemographic and health-related characteristics of the sample. Because of the older mean age of respondents with diabetes, age-adjusted least squared means from linear regression analysis were used to compare mean fat-related dietary score between those with and without diabetes. Linear regression analysis was also used to model the characteristics associated with fat-related dietary score in the population using standard modeling procedures (24). The independent variables were dichotomized and those contributing to the model were retained as main effects. Variables considered in the model included diabetes status (1⫽diabetes, 2⫽non-diabetes), sex (1⫽male, 2⫽female), educational attainment (1⫽ⱕhigh school, 2⫽⬎high school), weight status (1⫽not obese [BMI ⬍30], 2⫽obese [BMI ⱖ30]), exercise (1⫽active, 2⫽sedentary [defined as no non–work-related physical activity or exercise in the past month]), and family history of diabetes (1⫽yes, 2⫽no). Age was also included in the model as a continuous main-effect variable. Two-way interactions of the main effects with diabetes status were tested, and those found to contribute to the association with fatrelated dietary score were retained in the final model. A total of 1,088 with complete information on these variables were included in the final model. RESULTS Table 1 shows the sociodemographic and health characteristics of the sample. Puerto Ricans were relatively young and low in educational attainment. Approximately half reported that their birth place was Puerto Rico. More than half spoke English at home. Rates of overweight, obesity, smoking, and inactivity were high. Twenty percent reported not having health coverage. Nearly 40% had a family history of diabetes and 11.3% had diagnosed diabetes. Mean fat-related dietary score by sociodemographic and health characteristics is shown in Table 1 and the
Table 1. Mean fat-related dietary score by sociodemographic and health-related characteristics of adult New York City Puerto Ricans, 2000 Fat-related Dietary Score Characteristic
na
%b
MeanⴞSEcd
Total Age (y) 18-44 45-64 ⱖ65 Sex Male Female Educational attainment ⬍High school High school ⬎High school Birth place United States Puerto Rico Household language English Spanish Weight statuse Not overweight or obese Overweight Obese Smoking status Nonsmoker Smoker Exercise status Active Sedentary Health care coverage Medicaid Medicare Other None Diabetes family history Yes No Diagnosed diabetes Yes No
1,304
100.0
2.29⫾0.03
556 482 245
63.9 27.3 8.8
2.37⫾0.04y 2.23⫾0.08y 1.92⫾0.08z
447 857
44.2 55.8
2.38⫾0.06y 2.21⫾0.04z
536 313 364
44.3 21.9 33.8
2.33⫾0.06y 2.34⫾0.07y 2.25⫾0.05y
468 790
52.9 46.1
2.38⫾0.05y 2.19⫾0.05z
628 661
60.6 39.4
2.33⫾0.05y 2.21⫾0.05y
393 472 346
39.2 41.5 19.3
2.28⫾0.07y 2.25⫾0.05y 2.35⫾0.06y
966 327
70.8 29.2
2.26⫾0.03y 2.36⫾0.08y
641 656
53.8 46.2
2.25⫾0.02y 2.32⫾0.06z
372 220 484 208
28.0 8.9 42.9 20.2
1.99⫾0.08y 2.29⫾0.06z 2.28⫾0.06z 2.36⫾0.06z
643 613
38.2 61.8
2.18⫾0.06y 2.36⫾0.04z
606 698
11.3 88.7
2.07⫾0.03y 2.31⫾0.04z
a Unweighted sample size. The numbers within groups may not sum to the total sample of 1,304 because missing values and unknown responses were excluded from the analysis. b Weighted percent. c Mean based on fat-related dietary score items coded on a 4-point scale (usually or always, often, sometimes, rarely, or never), and scored 1 through 4 to correlate positively with fat intake for respective items in the questionnaire. d SE⫽standard error. e Not overweight or obese: Body mass index (BMI) ⬍25; overweight: BMI ⱖ25 and ⬍30; obese: BMI ⱖ30. yz Values with different superscripts within a grouping are significantly different from others in that group (P⬍0.05).
September 2006 ● Journal of the AMERICAN DIETETIC ASSOCIATION
1421
Table 2. Age-adjusted meana fat-related dietary score for adult New York City Puerto Ricans with and without diagnosed diabetes by socio-demographic and health-related characteristics, 2000 Age-adjusted Fat-related Dietary Score Characteristic
Diabetes
Nondiabetes
Independent variables P valueb
c
Total Age (y)d 18-44 45-64 ⱖ65 Sex Male Female Educational attainment ⱕHigh school ⬎High school Birth place United States Puerto Rico Household language English Spanish Weight statuse Nonobese Obese Smoking status Nonsmoker Smoker Exercise status Active Sedentary Health care coverage Medicaid Other or none Diabetes family history Yes No
4™™™™ mean⫾SE ™™™™3 2.17⫾0.05 2.31⫾0.04
0.07
2.27⫾0.05 2.10⫾0.04 1.84⫾0.06
2.38⫾0.04 2.26⫾0.10 1.98⫾0.16
0.05 0.18 0.44
2.18⫾0.06 2.16⫾0.06
2.39⫾0.07 2.24⫾0.05
0.07 0.24
2.19⫾0.05 2.18⫾0.09
2.38⫾0.06 2.24⫾0.05
0.03 0.64
2.29⫾0.06 2.12⫾0.06
2.35⫾0.06 2.22⫾0.07
0.54 0.10
2.23⫾0.06 2.11⫾0.06
2.32⫾0.06 2.29⫾0.07
0.36 0.08
2.20⫾0.05 2.19⫾0.04
2.27⫾0.04 2.45⫾0.06
0.33 ⬍0.01
2.12⫾0.07 2.30⫾0.07
2.30⫾0.04 2.33⫾0.09
0.06 0.74
2.05⫾0.06 2.30⫾0.05
2.27⫾0.03 2.35⫾0.07
⬍0.01 0.58
2.18⫾0.04 2.18⫾0.06
2.31⫾0.09 2.29⫾0.05
0.24 0.14
2.16⫾0.06 2.22⫾0.05
2.21⫾0.08 2.37⫾0.04
0.68 0.06
a Mean based on fat-related dietary score items coded on a 4-point scale (usually or always, often, sometimes, rarely or never), and scored 1 through 4 to correlate positively with fat intake for respective items in the questionnaire. b Comparing those with and without diabetes across rows. c SE⫽standard error. d Age-specific rates were not adjusted. e Nonobese: Body mass index (BMI) ⬍30; obese: BMI ⱖ30.
age-adjusted rates by diabetes status in Table 2. Unadjusted mean scores were significantly lower among women and among those who were aged 65 years and older, born in Puerto Rico, physically active, on Medicaid, have a family history of diabetes, and have diagnosed diabetes (Table 1). After age adjustment, the difference in fat-related dietary score between those with and without diabetes was somewhat attenuated (Table 2). Compared with respondents without diabetes, the age-adjusted mean dietary score for those with diabetes was significantly less (P⬍0.05) for those with a high school education or less, obesity, and those living an active lifestyle.
1422
September 2006 Volume 106 Number 9
Table 3. Linear regression model of characteristics associated with fat-related dietary score for adult New York City Puerto Ricans, 2000
Intercept Main effects Age (y) Sex (male vs female) Educational attainment (ⱕhigh school vs ⬎high school) Diabetes (yes vs no) Diabetes family history (yes vs no) Weight status (nonobese vs obese) Exercise status (active vs sedentary) Interaction effects Diabetes⫻family history Diabetes⫻weight status Diabetes⫻exercise status
 coefficient
P value
2.80
⬍0.01
⫺.01 .11
⬍0.01 0.20
.10 ⫺.27
0.14 ⬍0.01
⫺.19
0.07
⫺.21
0.02
⫺.06
0.45
.19 .26 ⫺.23
0.03 0.02 0.01
In the linear regression model, fat-related dietary score was inversely associated with age, and diabetes status had significant positive interactions with family history of diabetes and weight status (Table 3). As such, those without diabetes who were obese or had no family history of diabetes had higher scores. A significant negative interaction was observed between diabetes status and exercise status: those with diabetes and reporting active lifestyles had lower scores. The lowest mean subscale score was modifying meat, and those with diabetes were significantly more likely to modify meat compared to those without diabetes (Table 4). Significant differences in meat modifying behaviors between those with and without diabetes were observed for both individual subscale items: taking skin off chicken and trimming visible fat from red meat (data not shown). Avoiding fried foods also contributed to a lower-thanaverage score, although the difference between those with and without diabetes was not statistically significant in that regard. The highest mean subscale score was substituting fat-modified products, indicating that Puerto Rican adults living in New York City were least likely to make substitutions to lower fat intake, regardless of their diabetes status. Replacement of high-fat foods with fruits and vegetables also contributed to a higher-than-average score. Compared with those with diabetes, those without diabetes did not differ with respect to avoiding fat or fried foods, or to substitution or replacement to reduce fat consumption. DISCUSSION The results of this study show that dietary behaviors related to fat consumption in adult Puerto Ricans in New York City are associated with the sociodemographic and health characteristics of the population. The findings fur-
Table 4. Meana fat-related dietary subscale factor scores for adult New York City Puerto Ricans with and without diagnosed diabetes, 2000 Unadjusted Fat-related Dietary Score
Age-adjusted Fat-related Dietary Score
Subscale factor
Total
Diabetes
Non-diabetes
Total Avoid fat as a flavoring Avoid fried foods Modify meats to be low in fat Substitute fat-modified food products Replace high-fat foods with fruit and vegetables
4™™™™™™™™™™™™™™™™™ mean⫾SE c ™™™™™™™™™™™™™™™™™3 2.29⫾0.03 2.17⫾0.05 2.31⫾0.04 2.30⫾0.07 2.34⫾0.10 2.30⫾0.07 1.98⫾0.06 1.86⫾0.10 2.00⫾0.07 1.80⫾0.09 1.49⫾0.09 1.84⫾0.10 2.87⫾0.11 2.79⫾0.08 2.89⫾0.12 2.56⫾0.10 2.38⫾0.14 2.59⫾0.12
P valueb 0.07 0.73 0.33 0.01 0.46 0.27
a Mean based on fat-related dietary score items coded on a 4-point scale (usually or always, often, sometimes, rarely or never), and scored 1 through 4 to correlate positively with fat intake for respective items in the questionnaire. b Comparing those with and without diabetes across rows. c SE⫽standard error.
ther show that characteristics associated with fat-related dietary behaviors are dependent on diabetes status. Those with diabetes were somewhat more likely to engage in behaviors to lower fat compared to those without diabetes. Following adjustment for other factors, New York City Puerto Ricans who are diagnosed with diabetes and have a family history of diabetes or who are obese are at greater risk for increased fat consumption, whereas those without diabetes and active are at lower risk. Higher mean fat-related dietary scores (indicating higher fat intake) were observed as a result of less substitution with fat-modified food products and less replacement with high-fat foods, regardless of diabetes status. The lowest mean fat-related dietary scores (indicating lower fat intake) were observed for modifying meats to be low in fat, but those with diabetes had significantly lower scores in this regard compared with others. This study used a dual-frame telephone survey sampling approach to obtain population-based diet and health-related information from an urban minority population. The brief Fat-Related Diet Habits Questionnaire assessment method was selected because of its demonstrated usefulness in telephone surveys and its previous application in diverse population groups. It has shown to be useful in this study to make relative comparisons of fat-related dietary behaviors in New York City adult Puerto Ricans with respect to population and health characteristics, and to compare behaviors between those with and without diagnosed diabetes. Because the Fat-Related Diet Habits Questionnaire provides valid information related to dietary fat reduction behaviors, the information from this study can be applied directly to educational activities and interventions targeting specific behaviors to avoid, modify, replace, and substitute fatty food items in the diet of New York City Puerto Rican adults to prevent and control their risk for diabetes. These sampling and dietary methods may be adapted for use in other geographic areas and in other hard-to-reach population groups. Comparison with other studies using the Fat-Related Diet Habits Questionnaire is difficult because of differ-
ences in the study design, application of the Fat-Related Diet Habits Questionnaire, populations studied, and differences in reporting subscale scores. For example, in previous reports among northeast US Puerto Ricans and Hispanics from Rhode Island and Massachusetts, a longer 27-item Fat-Related Diet Habits Questionnaire was used (19,20). In a more recent study applying the brief Fat-Related Diet Habits Questionnaire among Mexicans from Washington state, subscale items specific to this population group were added to the instrument (21). Nonetheless, the findings with respect to the subscale scores show a consistent pattern among Hispanics. The lowest fat-related scores were reported with respect to modifying meat and avoidance of frying, and higher scores for replacement and substitution (19-21). However, compared with other Puerto Ricans living in the northeast United States, there is the suggestion that Puerto Ricans in New York City are more likely to avoid fat as a flavoring and to avoid fried foods (19). Lower fat-related scores (indicating lower fat intake) were associated with several sociodemographic characteristics of Puerto Ricans living in New York City, but only the inverse association with age remained significant in the regression model. Improvements in dietary score with respect to female sex, increasing age, and educational attainment have previously been reported (17). Loria and colleagues, using data from the Hispanic Health and Nutrition Examination Survey, reported generally lower fat and saturated fat consumption (expressed as percent of energy intake) with increasing age among mainland US Puerto Ricans (25). Polednak reported that mean fat intake (g/day) was associated with female sex, inversely associated with age, and not related to educational attainment among northeastern US Hispanics (26). There is considerable interest in the impact of immigration and acculturation on diet (27). Language as an indirect measure of acculturation was not significantly associated with fat-related dietary score. This is consistent with recent findings that Fat-Related Diet Habits Questionnaire scores among more acculturated Hispanics were not significantly different compared with those of
September 2006 ● Journal of the AMERICAN DIETETIC ASSOCIATION
1423
Hispanics who are less acculturated (21). Puerto Ricans born in Puerto Rico had significantly lower scores compared to US-born, but this was no longer observed after age adjustment, indicating that the differences were due to age and not birthplace. The findings of this study show that adult New York City Puerto Ricans with diabetes are more likely to engage in behaviors to reduce fat intake compared with those without diabetes. The statistically significant difference in mean fat-related dietary score for those with and without diabetes attenuated after age adjustment, indicating that the dietary behaviors are partially dependent on age differences in the two groups. Nonetheless, substantial differences in mean dietary score remained between those with and without diabetes after age adjustment, and significant differences between the two groups with respect to other health-related characteristics were observed. In the regression model, the association with dietary score remained for family history, weight, and exercise, with the direction of the effect being dependent on diabetes status. These findings point to a reduction in dietary risk behaviors among those with diabetes adopting an active lifestyle. Dietary risk increased for those without diabetes who are obese, or have no family history of diabetes. This suggests that New York City Puerto Ricans with diabetes are taking steps to control their disease, while the dietary behaviors together with other known risk factors are placing those without diabetes at increased risk. Of the five subscale factors, modifying meat to be low in fat and avoiding fried foods contributed to reduced fat scores among adult Puerto Ricans living in New York City, whereas they were less likely to substitute fatmodified food products and replace high-fat foods with fruits and vegetables to reduce fat consumption, on average. With the exception of modifying meats, those with and without diabetes did not differ with respect to subscale factor behaviors. Lower subscale scores for modifying meat and avoiding fried foods were also reported in previous studies of both Hispanics and non-Hispanics (19-21). Puerto Ricans living in New York City, in particular, would benefit from adopting behaviors to substitute and replace high-fat foods with low-fat products and food items. This would include increasing the consumption of fruits and vegetables for dessert and snacks, and the use of low-fat dressing, spreads, and dairy products. Previous studies of Hispanics in New York state and New York City, for example, show that whole milk is a major contributor to fat in the diet (28,29). Although reduced-fat milk represents a small percentage of the volume of milk available to inner–New York City Hispanics, availability is not a major problem (30). Thus, there is a need for greater consumer demand for reduced-fat milk in this population. Limitations to this study include the self-reported nature of the information obtained from telephone surveys and the possible bias in the measurements reported. However, an assessment of BRFSS measures concluded that most were at least moderately reliable and valid and many were highly reliable and valid (31). Second, estimates obtained from telephone surveys may also be biased due to relatively low response rates and to telephone noncoverage. However, bias associated with telephone
1424
September 2006 Volume 106 Number 9
noncoverage is likely to be minimal (32). In addition, the response (45.1%) and cooperation rates (62.8%) from this survey of inner-city minorities were higher than the statewide 2000 New York State BRFSS (32.9% and 34.7%, respectively) (33,34). Moreover, the Centers for Disease Control and Prevention has reported that bias in BRFSS data was not associated with response rates (35). CONCLUSIONS Fat-related dietary behaviors in New York City Puerto Rican adults are associated with diabetes status, sociodemographics, and other health characteristics. Culturally sensitive dietary treatments, education, and interventions to reduce fat intake are needed to prevent and control diabetes in this population. The emphasis should be placed on behaviors to substitute and replace high-fat foods with low-fat food items, including reduced-fat dairy products. These efforts should be targeted to New York City Puerto Ricans least likely to engage in dietary fatreducing behaviors, including men and those who are younger, sedentary, and obese. Nutrition and other health professionals can assist Puerto Ricans and others to adopt dietary behaviors to reduce fat consumption for diabetes prevention and control. The Diabetes Prevention Program’s Lifestyle Change Program, for example, provides guidance to reduce total fat and energy intake using basic dietary management concepts (36). The program offers educational materials to help participants identify low-fat food options, choose low-fat foods and flavorings, and modify recipes to reduce total fat. This research was supported by funding from the Association Schools of Public Health (S0742-18/19). References 1. Hanis CL, Hewett-Emmett D, Bertin TK, Schull WJ. Origins of US Hispanics: Implications for diabetes. Diabetes Care. 1991;14:618-627. 2. del Pinal JH. Hispanic Americans in the United States: Young, dynamic, and diverse. Stat Bulletin Metropolitan Insurance Co. 1996;77:2-13. 3. US Census Bureau. American FactFinder. Table PCT11. Hispanic or Latino by Specific Origin. Available at: http://factfinder.census.gov/servlet/ BasicFactsServlet. Accessed October 30, 2002. 4. Stern MP, Mitchell BD. Diabetes in Hispanic Americans. In: Harris MI, Cowie CC, Stern MP, Boyko EJ, Reiber GE, Bennet PH, eds. Diabetes in America. 2nd ed. Washington, DC: US Department of Health and Human Services, National Institutes of Health; 1995. DHHS publication no. (NIH) 95-1468. 5. Centers for Disease Control and Prevention. Selfreported prevalence of diabetes among Hispanics— United States, 1994-1997. MMWR. 1999;48:8-12. 6. Melnik TA, Hosler AS, Sekhobo JP, Duffy TP, Tierney EF, Englegau MM, Geiss LS. Diabetes prevalence among adult New York City Puerto Ricans, 2000. Am J Public Health. 2004;94:434-437. 7. Franz MJ, Bantle JP, Beebe CA, Brunzell JD, Chiasson J-L, Garg A, Holzmeister LA, Hoogwerf B, Mayer-Davis E, Mooradian AD, Purnell JQ, Wheeler
8. 9.
10.
11.
12.
13.
14. 15. 16.
17. 18.
19.
20.
21.
M. Evidenced-based nutrition principles and recommendations for the treatment and prevention of diabetes and related complications. Diabetes Care. 2002; 25:148-198. American Diabetes Association. Nutrition principles and recommendations in diabetes (Position Statement). Diabetes Care. 2004;27(suppl 1):S36-S46. US Department of Health and Human Services, US Department of Agriculture. Dietary Guidelines for Americans 2005. 6th ed. Washington, DC: US Government Printing Office; 2005. Available at: http://www. healthierus.gov/dietaryguidelines. Accessed March 2, 2005. Remington PL, Smith MY, Williamson DF, Anda RF, Gentry EM, Hogelin GC. Design, characteristics, and usefulness of state-based behavioral risk factor surveillance: 1981-87. Public Health Rep. 1988;103:366375. Behavioral Risk Factor Surveillance System User’s Guide. Atlanta, GA: US Department of Health and Human Services, Centers for Disease Control and Prevention; 1999. Kristal AR, Shattuck Al, Henry HJ. Patterns of dietary behavior associated with selecting diets low in fat: Reliability and validity of a behavioral approach to dietary assessment. J Am Diet Assoc. 1990;90:214220. Shannon J, Kristal AR, Curry SJ, Beresford SA. Application of a behavioral approach to measuring dietary change: The fat- and fiber-related diet behavior questionnaire. Cancer Epidemiol Biomarkers Prev. 1997;6:355-361. Neuhouser ML, Kristal AR, Patterson RE. Use of nutrition labels is associated with lower fat intake. J Am Diet Assoc. 1999;99:45-50. Levy L, Patterson RE, Kristal AR, Li SS. How well do consumers understand percentage daily value on food labels? Am J Health Promot. 2000;14:157-160. Patterson RE, Satia JA, Kristal AR, Neuhouser ML, Drewnowski AD. Is there a consumer backlash against the diet and health message? J Am Diet Assoc. 2001;101:43-46. Kristal AR, Hedderson M, Patterson RE, Neuhouser ML. Predictors of self-initiated, healthful dietary change. J Am Diet Assoc. 2001;101:762-766. Kristal AR, Shattuck AL, Patterson RE. Differences in fat-related dietary patterns between black, Hispanic, and white women: Results from the Women’s Health Trial Feasibility Study in Minority Populations. Public Health Nutr. 1999;2:253-262. Gans KM, Burkholder GJ, Upegui DI, Risica PM, Lasater TM, Fortunet R. Comparison of baseline fatrelated eating behaviors of Puerto Rican, Dominican, Colombian, and Guatemalan participants who joined a cholesterol education project. J Nutr Educ Behav. 2002;34:202-210. Gans KM, Burkholder GJ, Risica PM, Lasater TM. Baseline fat-related dietary behaviors of white, Hispanic, and black participants in a cholesterol screening and education project in New England. J Am Diet Assoc. 2003;103:699-706. Neuhouser ML, Thompson B, Coronado GD, Solomon
22.
23. 24. 25.
26. 27.
28. 29.
30.
31.
32.
33. 34. 35.
36.
CC. Higher fat intake and lower fruit and vegetable intakes are associated with greater acculturation among Mexicans living in Washington State. J Am Diet Assoc. 2004;104:51-57. Ruggles S, Sobek M, Fitch CA, Hall PK, Ronnander C. Integrated Public Use Microdata Series: Version 2.0. Minneapolis, MN: Historical Census Projects, University of Minnesota; 1997. Shah BV, Barnwell BG, Bieler GS. SUDAAN User’s Manual, Release 7.5. Research Triangle Park, NC: Research Triangle Institute; 1997. Kleinbaum DG, Kupper LL, Morgenstern H. Epidemiologic Research: Principles and Quantitative Methods. Lifetime Learning Publication. Belmont, CA; 1982. Loria CM, Bush TL, Carroll MD, Looker AC, McDowell MA, Johnson CL, Sempos CT. Macronutrient intake among adult Hispanics: A comparison of Mexican Americans, Cuban Americans, and mainland Puerto Ricans. Am J Public Health. 1995;85:684-689. Polednak AP. Use of selected high-fat foods by Hispanic adults in the northeastern US. Ethn Health. 1997;2:71-76. Satia-Abouta J, Patterson RE, Neuhouser ML, Elder J. Dietary acculturation: Application to nutrition research and dietetics. J Am Diet Assoc. 2002;102:11051118. Basch CE, Shea S, Zybert P. Food sources, dietary behavior, and the saturated fat intake of Latino children. Am J Public Health. 1992;82:810-815. Shea S, Melnik TA, Stein AD, Zansky SM, Maylahn C, Basch CE. Age, sex, educational attainment and race/ethnicity in relation to consumption of specific foods contributing to the atherogenic potential of diet. Prev Med. 1993;22:203-218. Weschler H, Basch CE, Zybert P, Lantigua R, Shea S. The availability of low-fat milk in an inner-city Latino community: Implications for nutrition education. Am J Public Health. 1995;85:1690-1692. Nelson DE, Holtzman D, Bolen J, Stanwyck CA, Mack KA. Reliability and validity of measures from the Behavioral Risk Factor Surveillance System (BRFSS). Soz Praventivmed. 2001;46(suppl 1):S3-S42. Ford ES. Characteristics of survey participants with and without a telephone: Findings from the Third National Health and Nutrition Examination Survey. J Clin Epidemiol. 1998;51:55-60. ORC Macro. 1999-2000 Diabetes Surveillance of Puerto Rican Adult Population in New York City. Technical Report. New York, NY: ORC Macro; 2002. Centers for Disease Control and Prevention. 2000 BRFSS Summary Data Quality Report. Atlanta, GA: Centers for Disease Control and Prevention; 2001. Centers for Disease Control and Prevention. Public health surveillance for behavioral risk factors in a changing environment: Recommendations from the Behavioral Risk Factor Surveillance Team. MMWR. 2003;52(No. RR-9):6. The Diabetes Prevention Program (DPP) Research Group. The Diabetes Prevention Program (DPP). Description of lifestyle intervention. Diabetes Care. 2002;25:2165-2171.
September 2006 ● Journal of the AMERICAN DIETETIC ASSOCIATION
1425