A multi-component intervention for modifying food selections in a worksite cafeteria

A multi-component intervention for modifying food selections in a worksite cafeteria

REPORTS & REVIEWS A Multi-Component Intervention for Modifying Food Selections in a W orksite Cafeteria JONI A. MAYER/ PH.D., TIMOTHY P. BROWN,2 M.A...

486KB Sizes 18 Downloads 29 Views

REPORTS & REVIEWS

A Multi-Component Intervention for Modifying Food Selections in a W orksite Cafeteria JONI A. MAYER/ PH.D., TIMOTHY P. BROWN,2 M.A.T., JOAN M. HEINS,3 R.D., AND DONALD B. BISHOP,4 PH.D.

IGraduate School of Public Health, San Diego State University, San Diego, California 92182; 2Department of Psychology, Washington University, St. Louis, Missouri 63130; 3Department ofInternal Medicine, Washington University School of Medicine, St. Louis, Missouri 63110; and 4Minnesota Department of Health, Minneapolis, Minnesota 55440 The modification of food selections at supermarkets, cafeterias, and vending machines has received an increasing amount of attention (Note 1). Given the high rate of dining out by Americans (1) and the variety of selections usually available, cafeterias may be optimal settings for influencing point-of-purchase dietary behavior. Worksite-based cafeteria interventions, in particular, have the advantages of 1) reaching a relatively stable population and 2) being able to interface with other components of employee weight-loss or fitness programs (2). In the present study, we combined three strategies (food labels, a nutrition awareness game, and incentive raffles) which individually had been successful in previous studies (3-9). We implemented two strategies (food labels, and a nutrition awareness game) simultaneously with the goal of reducing the mean number of calories per tray relative to baseline and postintervention levels. A third strategy involved offering employees who selected specific foods incentives through raffles. This procedure was used to increase the purchase rates of the targeted items. The setting of the study was the cafeteria in the local office of a Fortune 500 company. Employees of this office (n = 265) had a mean age of 46 years and included both blue-collar and white-collar workers (Note 2). Approximately 67% were women. The results from a health screening conducted prior to the study indicated that at least 26% of the employees were obese (i.e., ~120% ideal weight). The cafeteria, which served lunch to approximately 130 employees daily, offered a wide variety of foods, including hot entrees, vegetables, salad and sandwich bar items, grill items, and desserts.

0022-3182/87/1906-0277$02.00/0

© 1987 SOCIEfY FOR NUTRITION EDUCATION VOLUME 19 NUMBER 6 1987

OBSERVATION PROCEDURES AND MEASURES Throughout the study, which lasted a total of 12 weeks, two trained observers recorded customers' food selections. Because of the configuration of the food counters and cash register, the observers were, by necessity, visible to customers. The observers recorded food selections at the same time each day (i.e., 11:30 A.M. to 12:45 P.M.) on Tuesdays, Wednesdays, and Thursdays. We developed a checklist on which the observers noted the specific food items (e.g., green beans) in each food category (e.g., vegetables). To assess changes over time in "typical" lunch customers (versus new customers buying ancillary items in order to participate in the nutrition awareness game), observations included only customers who bought the main portion of their lunch from the cafeteria. Approximately 22% of the observation days included simultaneous recordings by the two raters. Cash register readings indicated the number of lunch customers per day and the daily sales throughout the course of the study. EXPERIMENTAL DESIGN AND PROCEDURES The study, which used a withdrawal design (10 and Note 2), included three consecutive phases: baseline 1, intervention, and baseline 2 (i.e., withdrawal of the intervention). Prior to baseline 1, caloric values of all food items were computed by a registered dietitian using a book of food nutrient values compiled by the cafeteria company (11), USDA Handbook No . 456 (12), and Bowes and Church's Food Values o/Portions Commonly Used (13). Beginning with baseline 1 and throughout the study, food choices were monitored, as described above. Each study phase lasted for 4 weeks. The 4JOURNAL OF NUTRITION EDUCATION

277

week menu cycle used during baseline 1 was repeated in the subsequent phases; the regular menu items were not altered. The intervention consisted of three components, which are detailed below. 1) Labels. We labeled all food items with respect to caloric content. Index cards (3" x 5 " and 4" x 6") displaying the name of the food item and the number of calories, were placed on the sneeze guards, as close as possible to the corresponding foods. 2) Nutrition awareness game. The "Food for Thought" game (Note 3) consisted of 52 playing cards, which, along with a number and suit, contained specific nutritional messages, colorful illustrations, and recipes. Employees (approximately 65% of those observed per day) took one card each day from a dispenser located at the cash register. At the end of the second, third, and fourth weeks, participants who had collected particular combinations of cards were eligible to enter raffles. Participation in this game was not contingent on food selections. In each raffle, 10 to 20 prizes were awarded. Prizesincluding baseball tickets, free cafeteria meals, gift certificates, and a barbecue grill (grand prize)were donated by the company and local merchants. Informational posters and fliers were circulated one week prior to the initiation of the game. 3) Incentive raffles. During the intervention phase, we examined the effects of an incentive procedure on the purchase of three food items. Specifically, a raffle was held in the cafeteria one day per week during the first 3 weeks of the intervention. Participation in these raffles was independent of participation in the "Food for Thought" game. All customers who chose the targeted food item (a salad, a fruit, and a carton of skim or low-fat milk) received a raffle ticket. On the day of each raffle, we placed a poster announcing the raffle and the targeted food item at the beginning of the food line. For each raffle one winner was chosen and the prize was awarded. Prizes for the incentive raffles included free cafeteria lunches and refrigerator magnets. At the end of the intervention, we removed the calorie labels and discontinued the game and the incentive raffles. We continued behavioral observations during baseline 2. Figure 1 presents the study design and sequence of intervention procedures.

POSTINTERVENTION SURVEY During two one-hour periods during the final week of baseline 2, the first author conducted brief interviews with all cafeteria customers (n = 48) as they 278

JOURNAL OF NUTRITION EDUCATION

Monitoring of Food Trays Baseline 1

Nutrition Game plus Labels

/R1~ ~ 1121314

5161718 WEEKS

Baseline 2 survey"", 9110111112

R1 salad raffle R2 milk raffle R:1 fruit raffle Figure 1. Schedule of study procedures

exited the cafeteria. Specifically, she asked them about their awareness of and participation in each component, whether the procedures influenced their food selections, and which procedures they would like to have continued.

FINDINGS Interrater agreement, which was calculated on 7 observation days and included 287 trays and 674 individual items, was 89%. This percentage was considered sufficiently high, given the nature of the observations (10). The mean number of calories per tray during each experimental phase was 46l.56 (SD = 188.11), 454.22 (SD = 184.63), and 463.69 (SD = 184.66), respectively. These values were not significantly different, as shown by a one-way analysis of variance (general linear models procedure; F(2) = .98, p > .38). The daily percentages of trays containing salads, low-fat milk, and fruits were evaluated to determine the effects of the incentive raffles. The purchase rates of each item increased when the related raffle was introduced and decreased back to baseline levels when the raffle was discontinued, demonstrating a functional relationship between the raffles and the purchase rates. The proportions of trays containing the targeted item on the day of each specific raffle were 46% (43/94) for salads, 17% (16/93) for lowfat milk, and 16% (17/105) for fruit during the salad raffle, low-fat milk raffle, and fruit raffle, respectively. When compared with the mean proportion of trays containing the item on other intervention days, these proportions were considerably higher. Specifically, on the salad-raffle day, the proportion of trays containing salads was 70% higher; on the milk-raffle day, the proportion of trays containing low-fat milk was 88% higher; and on the fruit-raffle day , the proportion of trays containing fruit was 78% higher. In the survey, all respondents (n = 48) said they were aware of the calorie labels, and the majority of VOLUME 19 NUMBER 6 1987

respondents read the labels on a regular basis and wanted them to be reinstated. Furthermore, nearly one-half of the respondents reported that the labels had influenced their food selections. The nutrition game was also popular, with 67% of respondents reporting that they participated. This self-reported participation rate was comparable with the percentage of customers that was observed taking playing cards (i.e., 65%). In addition, anecdotal data suggested that the game generated involvement and coworker interaction (e.g., participants were frequently observed trading cards), and cafeteria sales increased by 16% from baseline 1 to intervention. For each item (e.g., awareness, participation), the probability of an affirmative response differed significantly between the intervention components, as shown by Cochran Q tests (14, Note 4). IMPLICATIONS The results of the present study suggested that in this setting, a combination of a calorie-labeling procedure and a nutrition awareness game was not effective in significantly decreasing overall calorie consumption. In contrast, although the number of observations was small, we found that incentives that were contingent on specific food selections had a noticeable impact on increasing purchases of the targeted selections. The overall pattern of our findings is consistent with the findings of Cincirpini (9). His results indicated that in a university cafeteria, a combination of labels and incentives produced consistently positive nutritional changes, whereas labeling alone produced negligible results. Similarly, in the context of the labels and nutrition game in this study, the raffle incentives resulted in increased sales of each item that was targeted. The experimental design used in this study helped to control for potential seasonal influences (15) and the replication of the menu cycle across phases provided a stable context in which to evaluate the intervention. However, several methodological weaknesses in this study should be addressed. First, the observers were visible to employees. Although we gave employees a nonspecific explanation about the purpose of the observers, it is probable that during the intervention, employees realized that the observers were monitoring calories. However, in light of the nonsignificant changes from baseline 1 to intervention, the need to account for confounding variables due to the reactivity of the measurement procedure was reduced. Nevertheless, future studies could avoid potential reactivity by using an inventory system or conducting inconspicuous observations. VOLUME 19 NUMBER 6 1987

A second methodological factor that reduces the generalizability of the findings related to incentives was the scheduling of the raffles. More specifically, each raffle focused on the food purchases of one day only. Thus, it is not surprising that the effects were transient and that many customers who were interviewed reported being unaware of the raffles (Note 5). Future studies will be needed to evaluate both the short- and long-term cost-effectiveness (and acceptability by the clientele) of longer periods of incentive strategies. As noted above, the design of the study did not allow for the assessment of the individual components and their incremental effects. It is possible that either the labels or the nutrition game used alone, or the raffles in the absence of the other strategies, would have produced different results. The individual and combined contributions of each component should be evaluated directly in future studies. Because the nutrition information delivered by the game and the labels may have resulted in increased knowledge of the employees, pre- and postassessment of specific knowledge variables (in patrons vs. nonpatrons) may be warranted. Briefly, possible explanations for the failure of the intervention to result in meaningful caloric reductions include the following: 1) Overall, there were relatively few low-calorie items available (with the exception of salads); 2) the intervention was relatively brief compared to previous worksite cafeteria programs (3, 4); and 3) the labeling and game procedures were general, providing descriptive information about all cafeteria items without endorsing the lowercalorie items. Future research should analyze the influence of the various factors in the intervention, while addressing the methodological issues discussed above. 0

ACKNOWLEDGMENTS This research was funded by Grant No. 5-P60-AM20579-09 from the Pilot and Feasibility Studies Program, Diabetes Research and Training Center, Washington University School of Medicine, and Training Grant No. 2T32-HL07456-06 from the National Heart, Lung, and Blood Institute. Portions of this research were reported at the seventh annual meeting of the Society of Behavioral Medicine, San Francisco, California, March 1986. The authors are grateful to Gary Smith, Stephanie Thompson, Don Kramer, Marybell Bego, and Mitzi Hobbs for their assistance in implementing the program. Edwin B. Fisher, Jr. provided helpful comments throughout the JOURNAL OF NUTRITION EDUCATION

279

course of the study. We would also like to thank Denise Morrison for assistance in data reduction and Brenda Sims and Susanne Webb for manuscript preparation.

NOTES 1 Mayer, J. A., P. M. Dubbert, and J. Elder. Promoting nutrition at point-of-purchase: A review. (manuscript under review). 2 This research was approved by the Washington University School of Medicine Human Studies Committee. 3 The "Food for Thought" game was developed by the National Heart, Lung, and Blood Institute, and can be ordered through local chapters of the American Heart Association. 4 This analysis is used to test whether three or more matched sets (e.g., responses from the same subjects) of frequencies or proportions differ significantly among themselves. 5 All customers who chose the targeted item during the food incentive raffles received a raffle ticket. Since their numbers were automatically included in the drawing, no additional responses were required of them. Thus, it is possible that some of them were unaware of the specific details of the raffle (Le., did not notice the poster) or of their participation in the raffle.

LITERATURE CITED 1 National Restaurant Association. National restaurant association pocket factbook, Washington, DC, 1986, 6 pp. 2 Glanz, K., and T. Seewald-Klein. Nutrition at the worksite: An overview. Journal of Nutrition Education 18 (Supp.):SI-SI2,1986. 3 Zifferblatt, S. M., C. S. Wilbur, and J. L. Pinsky. Changing cafeteria eating habits. Journal of the American Dietetic Association 76:15-20,1980.

4 Schmitz, M. F., and J. E. Fielding. Point-of-choice nutritional labeling: Evaluation in a worksite cafeteria. Journal of Nutrition Education 18(Supp.):S65-S68, 1986. 5 Milich, R., J. Anderson, and M. Mills. Effects of visual presentation of caloric values on food buying by normal and obese person. Perceptual and Motor Skills 42: 15562, 1976. 6 Dubbert, P. M., W. G. Johnson, D. G. Schlundt, and N. W. Montague. The influence of caloric information on cafeteria food choices. Journal of Applied Behavior Analysis 17:85-92, 1984. 7 Mayer, J. A., J. M. Heins, J. M. Vogel, D. C. Morrison, L. D. Lankester, and A. L. Jacobs. Promoting low-fat entree choices in a public cafeteria. Journal of Applied Behavior Analysis 19: 397-402, 1986. 8 Davis-Chervin, D., T. Rogers, and M. Clark. Influencing food selection with point-of-choice nutrition information. Journal of Nutrition Education 17:18-22, 1985. 9 Cincirpini, P. M. Changing food selections in a public cafeteria. Behavior Modification 8:520-39, 1984. 10 Hersen, M., and D. H. Barlow. Single case experimental designs: Strategies for studying behavior change. New York: Pergamon Press, 1976,374 pp. 11 Interstate United. Nutrient analysis of Interstate United Corporation's recipe file. Chicago: Interstate United, 1984,56 pp. 12 United States Department of Agriculture. Nutritive value of American foods: Agriculture handbook no. 456. Washington, DC: U.S. Government Printing Office, 1975,291 pp. 13 Pennington, J. A. T., and H. N. Church. Bowes and Church's food values of portions commonly used (14th ed.). Philadelphia: J.B. Lippincott Company, 1985, 257 pp. 14 Siegal, S. Nonparametric statistics for the behavioral sciences. New York: McGraw-Hill, 1956,312 pp. 15 Zifferblatt, S. M., C. S. Wilbur, and J. L. Pinsky. Understanding food habits. Journal of the American Dietetic Association 76:9-14, 1980.

TEST-WEIGHING FOR HUMAN MILK VOLUME ASSESSMENT Test-weighing has long been the accepted means to assess milk volume intake of breastfed infants. The infant is weighed before and after the feeding with the difference representing the weight of breast milk consumed. This weight is converted to volume by use of a specific gravity value. In an attempt to validate the accuracy of test-weighing, Borschel, Kirksey, and Hannemann used test-weighing to assess milk volume intake of formula-fed infants, since the volume of formula can be directly measured (The American Journal of Clinical Nutrition 43:367-73,1986). They found that direct measurement of formula intake did not significantly differ from estimates of intake by the test-weighing method. In addition, milk intake per feeding for formula-fed infants was significantly higher at 4 and 6 months compared to intakes at earlier ages while intakes of breast fed infants remained unchanged throughout the first 6 months; yet, body weights of formula-fed and breastfed infants were similar throughout the entire period. This finding suggests that the breastfed infants were consuming fewer milliliters of milk per kilogram but still maintaining the same weight as their formula-fed counterparts.

280

JOURNAL OF NUTRITION EDUCATION

VOLUME 19 NUMBER 6 1987