Division of Epidemiology, School of Public Health, University of Minnesota, Minneapolis, MN 55454-1015, USA

Correspondence to: S A French, Division of Epidemiology, University of Minnesota, 1300 South Second Street, Suite 300, Minneapolis, MN 55454-1015, USA.french@epi.umn.edu

OBJECTIVE: To examine demographic, behavioral and dietary correlates of frequency of fast food restaurant use in a community-based sample of 891 adult women.

DESIGN: A survey was administered at baseline and 3 y later as part of a randomized, prospective intervention trial on weight gain prevention.

SUBJECTS: Women (n=891) aged 20-45 y who enrolled in the Pound of Prevention study.

MEASUREMENTS: Frequency of fast food restaurant use, dietary intake, demographic and behavioral measures were self-reported. Dietary intake was measured using the 60-item Block Food Frequency Questionnaire. Body weight and height were directly measured.

RESULTS: Twenty-one percent of the sample reported eating 3 fast food meals per week. Frequency of fast food restaurant use was associated with higher total energy intake, higher percentage fat energy, more frequent consumption of hamburgers, French fries and soft drinks, and less frequent consumption of fiber and fruit. Frequency of fast food restaurant use was higher among younger women, those with lower income, non-White ethnicity, greater body weight, lower dietary restraint, fewer low-fat eating behaviors, and greater television viewing. Over 3 y, increases in frequency of fast food restaurant use were associated with increases in body weight, total energy intake, percentage fat intake, intake of hamburgers, French fries and soft drinks, and with decreases in physical activity, dietary restraint and low-fat eating behaviors. Intake of several other foods, including fruits and vegetables, did not differ by frequency of fast food restaurant use.

CONCLUSION: Frequency of fast food restaurant use is associated with higher energy and fat intake and greater body weight, and could be an important risk factor for excess weight gain in the population.

International Journal of Obesity (2000) 24, 1353-1359

obesity; fast food; dietary intake; weight gain

Introduction

High-fat diets are a public health issue

Although dietary fat intake, as a percentage of total energy, has declined in the US population in recent years, from 40% in 1977/1978 to 33% in 1994, levels continue to be higher than the 30% recommended.^1,2,3 The prevalence of high-fat diets in the US may have contributed to the epidemic of obesity that currently affects the US population.^4,5 Currently, over 30%, or 58 million, Americans are overweight.⁶ In the most recent decades, the prevalence of obesity has sharply risen from 24.3% in 1960 to 33.3% in 1991.^6,7,8,9 Dietary practices contributing to excessive fat intake and overweight need to be identified so that public health interventions targeted at modifiable dietary behaviors may be implemented. Among the potential contributing dietary behaviors is intake of fast food fare.

Fast food is a growing component of the American diet

'Fast food' has been defined as food purchased in self-service or carry-out eating places without wait service.^10,11 Between 1970 and 1980, the number of fast food outlets increased from 30,000 to 140,000, and fast food sales increased by 300%.¹² Although fast food restaurants have diversified to include a much broader range of foods, outlets focusing on hamburgers and French fries continue to be the industry leaders in terms of sales volume.^11,12,13

Since the early 1970s the frequency of fast food use has dramatically increased. In the early 1970s, about 20% of the household food dollar was spent on food away from home. By 1995, an estimated 40% of the US household food dollar was spent on food away from home.¹³ Although not all of the away-from-home food purchases are from fast food restaurants, fast food restaurants comprise an increasing share. For example, in 1953, fast food comprised 4% of total sales of food away from home, compared to 34% of away-from-home food sales in 1997.^11,13

An early review of the impact of fast food on the nutritional quality of the diet concluded that the frequency of use of fast food restaurants was too low to impact overall dietary quality.¹⁴ Twenty years later, this may not be the case. Data from the 1994-1996 USDA Continuing Survey of Food Intakes by Individuals show that 56% of US adults report eating away from home on any given day; of these, about 33% ate at a fast food outlet.¹¹ Overall, fast foods accounted for 14% of total energy intake in 1995.¹⁰ Thus, current data suggest that fast food may play a significant role in overall energy intake and dietary quality.

Fast food is high in fat and energy

Concern has arisen over the nutritional quality of fast food fare, especially since such a large proportion of the population is exposed to fast food. Several studies have shown that fast food meals generally provide adequate nutrient density for protein, carbohydrates and several vitamins, but are generally low in calcium, vitamins C and A and fiber, and are high in calories, total fat, saturated fat, and cholesterol.^{10,14,15,16,17} For example, the 1995 CSFII reported that meals eaten out comprised 27% of eating occasions, but 34% of total energy intake.¹⁰ To stay within recommended guidelines, lower-fat food choices would be necessary during other eating occasions that day.

Previous studies have focused on the fat and energy derived from fast foods, but have not assessed the broader impact fast food intake might have on total dietary intake. This is an important issue if greater consumption of fast food is associated with poor food choices on other eating occasions. A recent study of 129 women found that, based on data from 7 day food diaries, women who ate out six or more times per week consumed more energy, fat and sodium, but equal amounts of fiber and calcium, compared to women who ate out five or fewer times per week.¹⁸ However, analyses of fat and micronutrients did not adjust for total energy intake. Consumption of specific foods was not examined by frequency of eating out.

Limited data are available on demographic factors and almost no data are available on behavioral factors associated with fast food use. Demographic factors positively associated with fast food use include age, being single or married without children, White ethnicity, and higher income.^10,19,20,21 People in these demographic groups may be at greater than average risk of excessive fat and energy intake, given their greater intake of fast foods. Behavioral information would be useful to better characterize the overall eating and physical activity patterns of those who heavily rely on fast food and would inform the development of more specific and targeted nutrition interventions.

Our previous report showed that frequency of fast food restaurant use was positively correlated with total energy and fat intake in women both prospectively and cross-sectionally.²² The present study expands on these analyses by examining intake of a broader range of specific foods and nutrients and by examining associations over time between changes in dietary intake and changes in frequency of fast food restaurant use. These data will provide a broader, more comprehensive picture of the impact that fast food restaurant use may have on total dietary quality. In addition, demographic and behavioral variables related to fast food restaurant use will provide descriptive information that may be useful for intervention targeting and development.

Methods

Subjects and procedure

Data for the present report were obtained as part of the Pound of Prevention study (POP).²³ Recruitment, intervention methods and study outcomes are fully described elsewhere.²³ POP was a 3 y intervention trial to examine methods to prevent weight gain among young adults of any body weight. Participants were volunteers recruited from the community using newspaper ads, radio public service announcements and direct mail. In-person recruitment of low-income women took place at community health clinics administering the Supplemental Food and Nutrition Program (WIC). Nine-hundred and ninety-eight women (including 404 low-income women, defined as annual household income £$25,000) and 228 men were recruited and enrolled in the study. To be eligible, participants had to be between 20 and 45 y, not currently pregnant or pregnant within the past year, free from serious disease, and willing to participate for 3 y.

Participants were randomized to a mail-based intervention or to a no-contact control group. Intervention consisted of monthly mailed newsletters with return postcards and periodic opportunities to take part in additional eating and exercise programs. Intervention continued for 3 y with annual clinic visits to assess body weight, dietary intake, and eating and exercise behaviors. Data from women only are reported in the present study due to the smaller number and greater demographic homogeneity among male participants. Women who did not become pregnant (n=106 became pregnant) and who had data at the third annual follow-up were included in the present analysis (n=891/998 women; 89.3%). The women averaged 35 y of age, and weighed 72.8 kg (160.6 lb; s.d.=35.7), with an average BMI of 27.0 kg/m² (s.d.=6.0). Forty-five percent were currently married, 61% had one or more children, 86% were White race/ethnicity, 46% had college degrees, and 60% reported family yearly income of $25,000.

Measures

The following measures were completed at baseline and at each yearly clinic visit.

Fast food restaurant use. Frequency of fast food restaurant use was estimated with the question 'About how many meals per week do you eat from fast food restaurants?' Fast food restaurants were not defined for participants.

Dietary intake. Habitual dietary intake during the past year was estimated using the 60-item Block Food Frequency Questionnaire (FFQ).²⁴ The FFQ has been validated in diverse populations, including women and low-income populations.^25,26,27,28 It should be noted that food frequency dietary assessment instruments underestimate food intake and are therefore most useful for ranking individuals rather than estimating absolute levels of energy or nutrient intakes.^25,26,27 Dietary variables used in the present report include total energy, percentage energy from fat, calcium, fiber and frequency of consumption of specific foods of interest, including those that might be served at fast food restaurants (eg hamburgers, French fries, soft drinks). However, the source of the food is not reported on the FFQ. It is not known, for example, whether French fries reported were obtained in a fast food restaurant or elsewhere. Total fruit servings and total vegetable servings were examined because these food groups are believed to be healthful and have been targeted for increase by public health guidelines. French fries were omitted from the aggregate measure of vegetable servings because they were examined separately as an example of a fast food. Milk consumption was examined due to concerns about low calcium intake among women, the reported low calcium content of fast food, and the possibility that soft drinks might be displacing milk among fast food restaurant users.

Behavioral variables

Physical activity. Physical activity was measured using an instrument adapted from Jacobs et al that has been used in several large epidemiologic studies, including those involving low-income populations.²⁹ The questionnaire describes 13 physical activities (eg running, racket sports, hiking, walking, swimming, snow shoveling, gardening, etc). Respondents check the average frequency per week that they did each of the 13 activities for 20 min or greater duration during the past year. A physical activity score was calculated as the reported frequency per week for each activity multiplied by its estimated intensity level in metabolic equivalents (METS). These products were summed across the 13 activities. Higher scores reflect greater physical activity levels (based on frequency and intensity). As a measure of sedentary behavior, participants reported the average number of hours per week spent watching television.

Restrained eating. Restrained eating was measured using the Cognitive Restraint subscale of Stunkard and Messick's Three Factor Eating Questionnaire (TFEQ-R).³⁰ The TFEQ-R measures dietary restraint, or the extent to which one engages in cognitive and behavioral efforts to limit food intake. Items focus on specific eating behaviors such as eating smaller portions, keeping track of caloric intake and avoiding specific foods. Higher scores indicate greater conscious control over eating or efforts to limit food intake. The TFEQ-R is valid and reliable and identifies those who consume fewer calories in naturalistic settings.^30,31,32

Low-fat eating behaviors. The low fat eating behaviors scale measures the extent to which individuals practice food preparation methods related to a low-fat diet.³³ This 18-item scale assesses five theoretically based dimensions of eating behavior and includes behaviors such as eating bread without butter or margarine, choosing low-fat dairy products, removing skin from chicken, and using low-calorie salad dressing. Higher scores indicate engaging in a greater number or more frequent behaviors to limit fat intake. The measure is valid, reliable and sensitive to interventions targeting low-fat eating behaviors.³³

Smoking. Smoking behavior was self-reported. Participants were classified as current smokers if they reported currently smoking cigarettes.

Body weight and height. Weight was measured in light clothing without shoes on a calibrated balance beam scale. Height was measured at baseline with a wall-mounted ruler. Body mass index was computed (BMI: weight (kg)/height (m)²).

Demographic variables

Demographic variables were self-reported and included age in years, current marital status (married vs not married), educational attainment ( high school, some college, college degree), income (annual household income £$25,000 vs >$25,000), ethnic identification (White vs non-White), employment status (employed full or part time vs not employed), and number of children.

Statistical analysis

All analyses were conducted using SAS statistical software.³⁴ To examine cross-sectional associations between frequency of fast food restaurant use and demographic, behavioral and dietary variables, frequency of fast food restaurant use was categorized into terciles. PROC GLM was used for continuous dependent variables in univariate analyses in which tercile of frequency of fast food restaurant use was the independent variable. Chi-square analyses were used for categorical variables. Total energy intake was included as a covariate in analyses of dietary intake variables.

Associations between changes in frequency of fast food restaurant use and changes in dietary intake were examined with PROC GLM in which the follow-up value of the food group was used as the dependent variable, and the baseline value was a covariate. Change in frequency of fast food restaurant use was examined by entering the baseline and follow-up frequency of fast food restaurant use as independent variables and interpreting the coefficient of the follow-up frequency of fast food restaurant use variable only. This approach is equivalent to a change score analysis and is preferred for several reasons described in detail by Cohen and Cohen.³⁵ Longitudinal analyses were adjusted for demographic co-variates and treatment group. P-values were considered significant at P<0.05.

Results

Prevalence of fast food restaurant use and trends over 3 y

Overall, 24% reported that they ate on average zero times per week at fast food restaurants, while 39.2% reported one visit per week, 15.7% reported two visits, and 21.1% reported three or more visits to a fast food restaurant on average per week. This distribution was fairly consistent at each year of follow-up. At follow-up three, 26.1% reported zero visits, 40.7% reported one visit, 14% reported two visits, and 19.2% reported three or more visits, on average, per week, to a fast food restaurant. There were no significant differences between treatment and control groups on baseline frequency of fast food restaurant use (mean=1.5 visits per week vs 1.7 visits per week, respectively, P<0.15). The correlation between number of weekly fast food restaurant visits at baseline and at follow-up three was r=0.57 (P<0.0001). Of those in the lowest tercile of fast food restaurant use at baseline, 62.9% remained in the lowest tercile at follow-up, while 8.7% moved into the highest tercile of fast food restaurant use. Similarly, 61% of those in the highest tercile of fast food restaurant use remained in the highest tercile at the follow-up, while 8.8% moved into the lowest tercile. Twenty-seven percent decreased their fast food restaurant visits by one or more meals per week, while 26.4% increased their fast food restaurant visits by one or more meals per week.

Cross-sectional demographic, behavioral and dietary correlates of fast food restaurant use

Cross-sectional associations between frequency of fast food restaurant use and demographic, behavioral and dietary variables are shown in Table 1. More frequent fast food restaurant use was significantly associated with younger age, being unmarried, lower income, non-White ethnicity, heavier body weight and higher BMI. Those in the highest tercile of frequency of fast food restaurant use were more likely to be employed, but also had the largest percentage of women with low incomes. Those falling in the middle tercile of frequency of fast food restaurant use were least likely to be White and had a greater number of children than those in the lowest or highest tercile of frequency of fast food restaurant use.

Measures of restrained eating and low-fat eating behaviors were inversely related to frequency of fast food restaurant use. Associations were linear, with women in the lowest frequency of fast food restaurant use tercile reporting the highest restrained eating scores and the highest low fat eating behaviors scores. Television viewing was highest in the highest frequency of fast food restaurant use tercile and lowest in the lowest frequency of fast food restaurant use tercile. Smoking and physical activity were unrelated to frequency of fast food restaurant use.

Total energy intake and percentage fat energy were positively associated with frequency of fast food restaurant use (Table 1). By contrast, intake of healthful foods and nutrients was inversely associated with frequency of fast food restaurant use. Those in the highest tercile of fast food restaurant frequency consumed significantly less fiber and fewer servings per day of vegetables and fruits. However, total calcium and milk intake did not differ by fast food restaurant frequency. Fast food restaurant use frequency was positively associated with hamburger, French fries and soft drinks consumption. For example, those in the highest tercile of frequency of fast food restaurant use reported consuming on average 1.8 hamburgers, 2.2 servings of French fries, and 7.6 soft drinks per week, compared to 0.8 hamburgers, 1.3 servings of French fries, and 4.1 soft drinks per week among low users of fast food restaurants. Intake of fried chicken, other chicken, fried fish, other fish, beef, bacon, sausage and eggs did not differ by frequency of fast food restaurant use.

Changes in fast food intake and changes in dietary and behavioral variables

Table 2 shows associations between changes in fast food restaurant use frequency and changes in dietary and behavioral variables over 3 y. Increases in frequency of fast food restaurant use were associated with increases in total energy intake, percentage energy from fat, servings of hamburgers, French fries and soft drinks, and body weight. On average, an increase of one fast food meal per week was associated with an increase of 234.4 kJ/day (56 kcal/ day), an increase of 0.6% in fat energy/day, and a weight gain of 0.72 kg (1.6 lb) over 3 y above the average weight gain over the 3 y period (average weight gain was 1.68 kg (3.7 lb)).²¹

Increases in frequency of fast food restaurant use were associated with decreases in vegetable servings per day, but changes in frequency of fast food restaurant use were unrelated to changes in intake of fruit, fiber and milk. Changes in frequency of fast food restaurant use were associated with changes in several other dietary variables, but associations were weak. Changes in behavioral variables reflecting attention to healthy eating (restraint scores, low-fat eating behaviors score), and physical activity, were inversely associated with changes in frequency of fast food restaurant use. Changes in television viewing time were not associated with changes in frequency of fast food restaurant use. There were no significant differences between treatment and control groups in change in frequency of fast food restaurant use (P<0.15).

Discussion

The purpose of this study was to describe the prevalence of fast food restaurant use and its demographic, dietary and behavioral correlates in a diverse sample of adult women. Frequency of fast food restaurant use was cross-sectionally and prospectively associated with higher fat and energy intake, greater body weight, more frequent intake of hamburgers, French fries and soft drinks, less frequent intake of vegetables, less restrained eating, and fewer low-fat eating behaviors. Frequency of fast food restaurant use was cross-sectionally positively associated with television viewing and prospectively inversely associated with leisure time physical activity.

These data have several potential implications. First, although the mean frequency of fast food restaurant use was low, about 37% of the sample visited fast food restaurants two or more times per week. Increases in the frequency of fast food restaurant use were associated with increases in body weight over a 3 y period. The average weight gain was 1.7 kg over 3 y. Women in the highest tercile of fast food restaurant use gained an additional 0.72 kg more than women in the lowest tercile of fast food restaurant use during the 3 y study period. This finding suggests that frequent fast food restaurant use could be a risk factor for excess weight gain over time.

Second, fruit and vegetable intake was lower among more frequent fast food restaurant users, suggesting either displacement by other foods, or else a lack of choosing fruits and vegetables at other eating occasions. It is questionable whether fruits and vegetables would be selected if offered on the menu at fast food restaurants. Low-fat menu items have not sold well at fast food restaurants, thus the reasons for low intake could be lack of interest or motivation or lack of perceived taste, rather than lack of availability or displacement by other foods.^19,36,37

Third, frequent fast food restaurant use was not associated with lower milk consumption or total calcium intake. Perhaps milk consumption occurs primarily at breakfast, while fast food meals are primarily concentrated at lunch and dinner.³⁸ However, data from at least one national study suggest an inverse association between soft drink consumption and calcium or milk intake, at least in some population subgroups.^39,40

Finally, frequent fast food restaurant use was highest prospectively and cross-sectionally, among those who were less concerned about restricting food intake or engaging in low fat eating behaviors and prospectively, among those who were less physically active. Together, these findings suggest that frequent fast food restaurant users may be less concerned with their eating and exercise behaviors more broadly, including the quality of their diets and their physical activity levels.

The current epidemic of obesity is caused largely by an environment that discourages physical activity and encourages the overconsumption of food energy.⁴ Environmental influences are variables that increase the behaviors that contribute to creating a postive energy balance.⁴ Environmental variables include the availability, convenience and low price of fast foods, the pervasive influence of television advertising for fast foods,^41,42 and the high population exposure to both television and fast food restaurants. In fact, the largest fast food restaurant in the US recently partnered with a well-known television manufacturer to give away free television sets in their restaurants.⁴³ Behaviors include leisure time physical activity, sedentary behaviors such as television viewing, and dietary intake. Thus, fast food use is one of a cluster of inter-related behaviors that may contribute to excess weight gain and obesity. The independent contribution of each behavior and their potential interactions on body weight change over time warrant additional research attention.

Community-based nutrition interventions promoting healthful eating behaviors may need to incorporate specific components addressing fast food consumption and its potential impact of overall dietary intake. Frequent fast food restaurant users might also be targeted for intervention around increasing fruit and vegetable intake and decreasing fat intake at other eating occasions during the day

1 US Department of Health and Human Services. Healthy People 2000: National Health promotion and disease prevention objective. US Department of Health and Human Services, Public Health Service: Washington DC, 1988,

2 McDowell MA, Briefel RR, Alaimo R, Bischof AM, Caughman CR, Carroll MD, Loria CM, Johnson CL. Energy and macronutrient intakes of persons ages 2 months and over in the United States: Third National Health and Nutrition Examination Survey, Phase 1, 1988-91. Advance Data From Vital and Health Statistics, no. 255, National Center for Health Statistics: Hyattsville, MD, 1994,

3 Borrud L, Wilkinson-Enns C, Mickle S. What we eat: USDA surveys food consumption changes. CNI 1997; 18: 4-5.

4 Hill J, Peters J. Environmental contributions to the obesity epidemic. Science 1998; 280: 1371-1374. Article MEDLINE

5 Shah M, Jeffery R. Is obesity due to overeating and inactivity, or to a defective metabolic rate: a review. Ann Behav Med 1991; 13: 73-81.

6 Kruczmarski RJ, Flegal KM, Campbell SM, Johnson CL. Increasing prevalence of overweight among US adults: the National Health and Nutrition Examination Surveys, 1960-1991. JAMA 1994; 272: 205-211. MEDLINE

7 Flegal KM, Harlan WR, Landis JR. Secular trends in body mass index and skinfold thickness with socioeconomic factors in young adult women. Am J Clin Nutr 1988; 48: 535-543. MEDLINE

8 Flegal KM, Harlan WR, Landis JR. Secular trends in body mass index and skinfold thickness with socioeconomic factors in young adult men. Am J Clin Nutr 1988; 48: 544-551. MEDLINE

9 Shah M, Hannan PJ, Jeffery RW. Secular trend in body mass index in the adult population of three communities from the upper mid-western part of the USA: the Minnesota Heart Health Program. Int J Obes 1991; 15: 499-503. MEDLINE

10 Biing-Hwan L, Frazao E. Nutritional quality of foods at and away from home. Food review. US Government Printing Office, Food and Rural Economics Division, Economics Research Service, USDA. 1997,; pp 33-40.

11 USDA continuing survey of food intakes by individuals, 1994-1996. US Department of Agriculture, Economic Research Service: Washington DC, 1997,

12 Sheridan M, McPherrin E. Fast food and the American diet. American Council of Science and Health: New York, 1983,

13 Putnam J, Allshouse J. Food consumption, prices, and expenditures, 1996, SB-928. US Department of Agriculture, Economic Research Service: Washington, DC, April 1996,

14 Shannon BM, Parks SC. Fast foods: a perspective on their nutritional impact. J Am Diet Assoc 1980; 76: 242-247. MEDLINE

15 Guenther PM, Ricart G. Effects of eating at food service establishments on the nutritional quality of women's diets. Top Clin Nutr 1989; 4: 41-45.

16 Haines PS, Popkin BM, Guilkey DK. Methods of patterning eating behaviors of American women. J Nutr Educ 1990; 22: 124-132.

17 Greecher CP, Shannon B. Impact of fast food meals on nutrient intake of two groups. J Am Diet Assoc 1977; 70: 368-372. MEDLINE

18 Eck-Clemens LH, Slawson DL, Klesges RC. The effect of eating out on quality of diet in premenopausal women. J Am Diet Assoc 1999; 99: 442-444. MEDLINE

19 Schwenk NE. Trends in food and alcohol consumption away from home. Family Econ Nutr Rev 1995; 8: 30-40.

20 Ries CP, Kline K, Weaver SO. Impact of commercial eating on nutrient adequacy. J Am Diet Assoc 1987; 87: 463-468. MEDLINE

21 Consumer Reports. Where's the beef? Consumer Reports 1984 July: ; 367-373.

22 Jeffery RW, French SA. Epidemic obesity in the US: are fast foods contributing? Am J Public Health 1998; 88: 277-280. MEDLINE

23 Jeffery RW, French SA. Prevention of weight gain in adults: the Pound of Prevention Study. Am J Public Health 1999; 89: 747-751. MEDLINE

24 Block G, Hartman AM, Dresser CM, Carroll MD, Gannon J, Gardener L. A data-based approach to diet questionnaire design and testing. Am J Epidemiol 1986; 124: 453-469. MEDLINE

25 Mares-Perlman JA, Klein BE, Klein R, Ritter LL, Fisher MR, Freudenheim JL. A diet history questionnaire ranks nutrient intakes in middle-aged and older men and women similarly to multiple food records. J Nutr 1993; 123: 489-501. MEDLINE

26 Block G, Woods M, Potosky A, Clifford C. Validation of a self-administered diet history questionnaire using multiple diet records. J Clin Epidemiol 1990; 43: 1327-1335. MEDLINE

27 Block G, Thompson FE, Hartman AM, Larkin FA, Guire KE. Comparison of two dietary questionnaires validated against multiple dietary records collected during a 1-year period. J Am Diet Assoc 1992; 92: 686-693. MEDLINE

28 Coates RJ, Eley JW, Block G, Gunter EW, Sowell Al, Grossman C, Greenberg RS. An evaluation of a food frequency questionnaire for assesing dietary intake of specific carotenoids and vitamin E among low-income black women. Am J Epidemiol 1991; 134: 658-371. MEDLINE

29 Jacobs DR, Hahn LP, Haskell WL, Pirie P, Sidney S. Validity and reliability of short physical activity history: CARDIA and the Minnesota Heart Health Program. J Cardiopulmon Rehab 1989; 9: 448-459.

30 Stunkard AJ, Messick S. The three factor eating questionnaire to measure dietary restraint, disinhibition, and hunger. J Psychosom Res 1985; 29: 71-83. MEDLINE

31 Laessle RG, Tuschl RJ, Kotthaus BC, Pirke KM. A comparison of the validity of three scales for the assessment of dietary restraint. J Abnorm Psychol 1989; 98: 504-507. MEDLINE

32 Tuschl RJ, Platte P, Laessle RG, Stichler W, Pirke KM. Energy expenditure and everyday eating behavior in healthy young women. Am J Clin Nutr 1990; 52: 81-86. MEDLINE

33 Kristal AR, White E, Shattuck AL, Curry S, Anderson GL, Fowler A, Urban N. Long-term maintenance of a low-fat diet: durability of fat-related dietary habits in the Women's Health Trial. J Am Diet Assoc 1992; 92: 553-559. MEDLINE

34 SAS Institute Inc. SAS/STAT software: changes and enhancements through release 6.12. SAS Institute: Cary, NC, 1997,

35 Cohen J, Cohen P. Applied Multiple Regression/Correlation Analysis for the Behavioral Sciences. Erlbaum: Hillsdale, NJ, 1975,, pp 72-73, 413-423.

36 Consumer Reports. Fast food for fat-watchers. Consumer Rep 1993 September; 574-578.

37 Kramer L. Taco Bell 'lightens up' on Border Lights. Taco Bell amends its Border Lights menu and launches Lighten Up. Nation's Restaurant News 1996; 30: 1.

38 Jekanowski M. Causes and consequences of fast food sales growth. USDA/ERS Food Rev 1999; 22: 11-16.

39 Liquid candy: how soft drinks are harming American's health. Center for Science in the Public Interest: Washington, DC, , 1998,

40 Harnack L, Stang J, Story M. Soft drink consumption among US children and adolescents: nutritional consequences. J Am Diet Assoc 1999; 99: 436-441. MEDLINE

41 Story M, Faulkner P. The prime time diet: a content analysis of eating behavior and food messages in television program content and commercials. Am J Public Health 1990; 80: 738-740. MEDLINE

42 Kotz K, Story M. Food advertisements during children's Saturday morning television programming: are they consistent with dietary recommendations? J Am Diet Assoc 1994; 94: 1296-1300. MEDLINE

43 Zenith teams with McDonald's in Monopoly promo. Brandweek 2000; 41: 90 (Article A61408896).

Table 1 Cross-sectional associations at baseline^a between frequency of fast food restaurant use and demographic, dietary and behavioral variables among 891 women

Table 2 Associations between changes in frequency of fast food restaurant use and changes in dietary intake over 3 y among 891 women^a