The prevention of obesity has become a major public health target. Focus has shifted from considering obesity, nutrition and health as mainly private issues to recognizing the responsibility of society in creating environments and conditions that may support and promote healthy eating habits and an active lifestyle(Reference Dorfman and Wallack1–3). Consequently, settings like worksites and schools have received increasing attention as important health-promoting platforms(3, 4), and many private and public companies are showing interest in investing in health-promoting initiatives to protect and develop their human resources(Reference Heinen and Darling5).
Health-promoting initiatives aimed at increasing the availability and accessibility of healthy food at the worksite have been shown in some environments to provide an efficient and sustainable approach to improving employees’ dietary habits (e.g. free available fruit and serving more fruit and vegetables at worksite canteens)(Reference Alinia, Lassen and Tetens6–Reference Thorsen, Lassen and Tetens8). Moreover, these kinds of environmental-level initiatives may not be restricted to a self-selected subset of motivated individuals who choose to participate in worksite health promotion programmes but rather has the potential to impact the entire worksite population(Reference French9).
Health-promoting initiatives across the settings of the worksite and the home environment provide new opportunities for enabling and promoting a healthy eating pattern. A previous study found that involving the family to create a supportive home environment in addition to a supportive worksite environment resulted in an increase in total fruit and vegetable intake among workers in the worksite-plus-family group compared with the control group(Reference Sorensen, Stoddard and Peterson10).
A direct way of reaching families is by providing employees with healthy ready-to-heat meals offered by the worksite to bring home to their family or elsewhere, at subsidized or market prices. This relatively novel concept, here referred to as healthy Canteen Take Away (CTA), seems to be growing in popularity among employees in Western countries(Reference Wanjek11). Also, Heinen and Darling mention healthy dinners-to-go offered in the employees’ café as an opportunity for employers to impact and help employees and their families to reduce and manage health risk factors such as obesity(Reference Heinen and Darling5). In Denmark we have gathered some information from twenty companies that have already some experience in operating CTA concepts. Different concepts were developed and adapted to the specific character of each worksite. The worksite canteens produce, pack and offer the CTA meals once, twice or several times per week. The employees order their take-away meals usually the day before and pay for and collect them at the canteen on their way home the next day.
To our knowledge the effectiveness of the CTA health-promoting concept in improving employees’ dietary habits has not been described in the scientific literature. Possible nutritional health benefits may include improvement of the nutritional quality of the diet of employees and their families directly by the provision of healthy meals that may substitute less healthy meals, including other types of ready meals and fast food(Reference Jeffery, Baxter and McGuire12), and in the long term influence the norms for healthy eating(Reference Winston, Johnson and Wilson13). Lack of time to prepare healthy meals as well as habit are frequently reported barriers for not following nutritional recommendations, especially for those who intend to eat healthily but do not do so(Reference Groth, Sorensen and Biltoft-Jensen14). From the perspective of the employee another important argument for supplying CTA meals is as a way of tackling the work–family conflict(Reference Roos, Sarlio-Lahteenkorva and Lallukka15). CTA meals may reduce the time needed for shopping, cooking and cleaning at home, thereby helping employees in their daily life to balance work demands with personal and family commitments(Reference Carnethon, Whitsel and Franklin16).
The present study is a part of a bigger public–private partnership project aimed at identifying and disseminating solutions to practical challenges associated with setting up a take-away service in worksite canteens. Three research partners and twelve private and public partners are behind this project, including governmental institutions, labour unions and private companies. The aim of the present study was to investigate the effectiveness of a CTA concept in promoting healthy eating habits among employees by evaluating the nutritional quality of employees’ evening meal intake and their total daily intake on weekdays when they received CTA compared to weekdays not receiving CTA.
Methods
Recruitment and study design
A financial worksite that offered CTA twice weekly (Wednesdays and Fridays) to 750 employees was invited to participate in the present study. About one-third of employees used the scheme regularly. The worksite had offered take-away meals from an in-house canteen to the employees and their families for almost 5 years, being one of the first companies to offer CTA in Denmark. The employees may buy the CTA meals at a price equivalent to the cost of the raw materials and labour. The worksite also offers lunch meals in the canteen as well as free fruit. The worksite was encouraged to supply CTA meals during the study period that would follow generally recognized nutrition recommendations. The simple food-based Healthy Meal Index was provided to the canteen as a tool to evaluate the nutritional quality of the CTA meals. The Healthy Meal Index, as described in a previous paper, focuses on the contents of fruit and vegetables, fat content and quality and contents of wholegrain products and potatoes(Reference Lassen, Biltoft-Jensen and Hansen17). Furthermore, the worksite supported the participation of employees in the study by paying for time off work in order to receive instruction and to complete the dietary interviews. No health promotion activities were directly connected with the CTA project.
All employees working in the worksite were given the opportunity to participate in the present study by an announcement in the worksite's internal newsletter describing the study. Participants received CTA meals free of charge throughout the study period for themselves as well as for their families. Eligible for inclusion were healthy men and healthy non-pregnant women aged 18 years and older expecting to be present at the worksite throughout the 7-week study period (middle of October to the beginning of December 2008). The study was performed in accordance with the ethical standards of the Helsinki Declaration of 1975, as revised in 2008. Written, informed consent was obtained from all participants.
A total of thirty-four employees responded to the announcement. Out of these twenty-eight enrolled in the study. Reasons for not enrolling were either time pressure (four employees) or illness (two employees). Moreover, data from one participant were excluded from the analysis due to deliberate weight loss during the study. Thus, the number of participants included in analysis was twenty-seven.
During the study period the project team randomly selected two portions of each CTA meal provided and weighed each component separately to get a picture of the meals provided by the worksite (ten different CTA meals).
Dietary assessment and background information
Two different dietary assessment instruments were applied for both CTA days and days not receiving CTA (non-CTA days). Four face-to-face 24 h dietary recalls were applied on non-consecutive days over the 7-week study period (different weeks) to measure total food intake during the previous weekday in order to get total dietary intake on CTA and on non-CTA days. Two of the 24 h dietary recalls were scheduled on Thursdays to get intakes during the previous day receiving CTA and two were equally scheduled on Tuesdays, Wednesdays and Fridays to measure the food intake during the previous day with no CTA provided. Due to employees having weekends off it was not possible to perform interviews on Saturdays to measure food intakes on Fridays. The 24 h dietary recall questionnaire was a modified form of the dietary record questionnaire from the Danish National Dietary Survey 2000–2002(Reference Biltoft-Jensen, Matthiessen and Rasmussen18). For canteen lunch and take-away meals, recipes were provided by the canteen manager in order to obtain more precise estimates of intake.
Furthermore, a digital photographic method recording evening meal intake at weekdays for three consecutive weeks was applied to obtain more precise nutrition information on CTA meals in relation to non-CTA meals. Participants were randomly assigned to start the photographic recording from the beginning of either the second or the fourth week of the 7-week study period. Participants were provided with digital cameras and were instructed orally and given detailed written information on how to capture images of their evening meals served on plates (both CTA and non-CTA meals) for all weekdays in three consecutive weeks. The cameras were Nikon Coolpix S210 with electronic VR image stabilization and Motion Detection for sharp, steady results. Briefly, images should be taken before eating and again after finishing eating, including possible leftovers, using the auto function while seated at the table pointing the camera at a 45° angle towards the plates. A ruler was provided to be placed beside the plate as an internal reference in all images. The participants were asked to keep different meal components separate on the plate, and it was emphasized that images should be taken of all foods, including extra food portions, if necessary, on additional plates. Additionally, participants were provided with a notebook to record the recipes and ingredients given in either grams or common household measures like cups, spoons, slices, etc. The participants were repeatedly reminded by emails to pick up their CTA meals and to record their evening meal intake using the digital photographic method at weekdays throughout the 3-week recording period. An average of one evening meal for each participant (out of fifteen recording days) was not recorded, resulting in a compliance rate of 94 %.
Two trained image analysts working within the area of food and nutrition estimated the weights of individual foods within the meals. Beverages were not included as food and drinks have differing effects on satiety and energy intake and in order to enhance comparability with other studies(Reference Johnson, Wilks and Lindroos19, Reference Rolls20). The validity of the digital photographic method was tested prior to the survey in another study against the weighed record method of nineteen participants’ usual evening meals for five consecutive days. Correlation coefficients between the two methods for intake of major food groups and nutrients, including energy content and macronutrient distribution, were between 0·83 and 0·97. Comparable means and acceptable limits of agreement (mean difference ± 2 sd) were found with regard to macronutrient distribution, energy density and energy-adjusted foods(Reference Lassen, Poulsen and Ernst21).
The software program General Intake Estimation Systems version 0·995f (2008-08.04; Danish Food Institute, Technical University of Denmark, Soeborg, Denmark) and the Danish Food Composition Databank(Reference Saxholt, Christensen and Møller22) were used to calculate food and nutrient intakes for both dietary assessment methods and for the CTA meals provided by the worksite. At the beginning of the study period height and body weight (participants without shoes and wearing light indoor clothing) were measured using a Soehnle (Backnang, Germany) Verona Quattrotronic digital scale (model 63686) to the nearest 0·1 kg, and a Soehnle 5001 Ultrasonic Height Measure to the nearest centimetre, respectively. BMR was estimated according to Schofield's equations(23) and the ratio of estimated energy intake (EI) to estimated BMR (EI:BMR) was calculated.
Background information such as gender, age, education and occupation was assessed using a questionnaire based on the questionnaire from the Danish National Dietary Survey 2000–2002(Reference Andersen, Christensen and Groth24).
Statistical analysis
Outcome variables from both the 24 h dietary recalls and the digital photographic method included fruit and vegetable intake (g/d and g/10 MJ, excluding potatoes and a maximum of 100 g of fruit juice daily in accordance with the Danish dietary guidelines(Reference Astrup, Andersen and Stender25)), dietary fibre intake (g/10 MJ), total energy intake (kJ) and energy density excluding beverages (kJ/100 g), and finally macronutrient intakes including total fat, saturated fat, protein, carbohydrate and added sugar (percentage of total energy intake, %E). First, all outcome variables were described univariately in terms of unadjusted mean differences between CTA and non-CTA meals. A multivariate analysis was then performed by means of a mixed-effects model, using SAS Enterprise Guide 4·0 (SAS Institute Inc., Cary, NC, USA). Evening meal type, i.e. CTA or non-CTA meals, was included as a fixed effect as well as gender, age, BMI and education (two levels: primary school/high school or university or equivalent). To adjust for dependency in repeated measures within subjects, random effects were added for employee. The interaction between time in the study and meal type was tested and found insignificant. Homogeneity of variance and normality of the residuals were examined using graphical methods. Given the multitude of statistical tests, a P value of <0·01 was taken in order to reduce the probability of false-positive findings.
Results
Participants’ characteristics
Half of the participants (52 %) were women. The mean age was 40 (sd 6, range 27–52) years and mean BMI was 24 (sd 2) kg/m2. Forty-one per cent of the participants were overweight whereas none were obese. Sixty-three per cent of the participants had completed a medium-term or long-term higher education. The majority of the participants lived with a partner and children (70 %). The rest lived alone (11 %), with a partner or another adult (11 %) or with children only (7 %). Unfortunately we have no detailed information on the general employee profile of the worksite. However, according to personal communication with the canteen and personnel managers, the participants in the present study did not differ considerably from other employees at the worksite.
Evening meal intake
The nutritional composition of the CTA meals provided by the worksite during the study is shown in Table 1. On average the provided meals contained 3·0 (sd 0·5) MJ with an average energy density of 520 (sd 134) kJ/100 g.
Table 1 Nutritional composition of the CTA meals provided during the study
CTA, Canteen Take Away; %E, percentage of total energy intake.
*Excluding potatoes.
Employees’ nutritional intakes from CTA and non-CTA meals assessed by the digital photographic method are shown in Table 2. The CTA meals consumed contained on average 2·6 (sd 0·8) MJ, of which 30·3 (sd 11·0) % came from fat, and on average 200 (sd 71) g vegetables, corresponding to the average amount of vegetables found in the CTA meals provided by the company (234 (sd 69) g). The average energy density of the consumed meals was 462 (sd 60) kJ/100 g.
Table 2 Employees’ evening meal intake on days receiving CTA and on days not receiving CTA (beverages not included) assessed by the digital photographic method
CTA, Canteen Take Away; %E, percentage of total energy intake.
*Adjusted difference from a multivariate analysis performed by means of a mixed-effects model. Adjusted for gender, age, BMI and education.
†Excluding potatoes.
‡Including not more than 100 g fruit juice/person per d.
The results from the mixed-effects model analysis showed that the CTA meals contained on average 88 (95 % CI 71, 104) g more fruit and vegetables and 18 (95 % CI 15, 20) g more dietary fibre per 10 MJ compared to the non-CTA meals (P < 0·01). Fat content was on average 7·8 (95 % CI −10·3, −5·4) %E lower for the CTA meals compared to the non-CTA meals, whereas energy density on average was 187 (95 % CI −225, −149) kJ/100 g lower compared to the non-CTA meals (P < 0·001). A gender-related effect on energy intake was seen (0·7 MJ lower for women compared to men) as well as an age- and BMI-related effect on dietary fibre intake, with decreasing intake with increasing age and increasing BMI. Otherwise, there were no significant effects of gender, education, BMI or age.
Daily dietary intake
For the assessment of under-reporting of energy intake, the EI:BMR ratio was determined for each participant. Average EI:BMR was 1·44 (sd 0·3) with three of the participants (11 %) below the Goldberg cut-off value of 1·06 (4 d of dietary data; not shown)(Reference Goldberg, Black and Jebb26). The data presented on total daily intake in Table 3 include the under-reporters, as removal of under-reporters did not change the overall trends observed. The results from the mixed-effects model analysis revealed that participants’ intake of fruit and vegetables was significantly higher on days receiving CTA compared to days not receiving CTA when expressed as both g/d and g/10 MJ (Table 3). The intake of fruit and vegetables was on average 129 (95 % CI 49, 210) g higher on CTA days compared to non-CTA days (P = 0·002). The difference in intake of vegetables alone was 109 (95 % CI 62, 155) g between meal types (P < 0·001). Moreover, energy density and protein content were significantly lower and higher, respectively, on days receiving CTA compared to non-CTA days (an average difference of −77 (95 % CI −132, −21) kJ/100 g on a daily basis, P = 0·01, and an average difference of 2·7 (95 % CI 1·6, 3·8) %E on a daily basis, P < 0·001, respectively). A gender-related effect was seen both for energy intake (2·9 MJ lower for women compared to men) and energy-adjusted vegetable intake (269 g/10 MJ higher for women compared to men). Moreover, a BMI-related effect was seen for fruit intake, with decreasing intake with increasing BMI (expressed both as g and g/10 MJ). Otherwise no significant effects were seen of gender, education, BMI or age.
Table 3 Employees’ total daily intake on days receiving CTA and on days not receiving CTA assessed by 24 h dietary recalls
CTA, Canteen Take Away; %E, percentage of total energy intake.
*Adjusted difference from a multivariate analysis performed by means of a mixed-effects model. Adjusted for gender, age, BMI and education.
†Excluding beverages.
‡Excluding potatoes.
§Including not more than 100 g fruit juice/person per d.
Discussion
The results showed that receiving CTA from the worksite was associated with a higher overall nutritional quality of evening meals compared to non-CTA meals when using energy density as a dietary quality marker(Reference Patterson, Warnberg and Poortvliet27). Average energy density of the consumed CTA meals excluding beverages was 462 (sd 60) kJ/100 g, on average 187 (95 % CI −225, −149) kJ/100 g lower than the consumed non-CTA meals. Also, compared to the non-CTA meals, the CTA meals showed other positive nutritional benefits, including a higher content of both dietary fibre and fruit and vegetables as well as a lower content of fat and saturated fat.
The differences observed in food and nutrient contents between the CTA and the non-CTA meals were generally reflected in dietary intakes across the whole day. The difference in intake of fruit and vegetables was on average 129 g/d between CTA and non-CTA days. Most of the difference in fruit and vegetable intake was accounted for by an increase of vegetable intake by 109 g, equalling about 1 serving. This is promising, as it is considered more challenging to increase vegetable intake compared with fruits, which require little preparation and have a sweet taste that appears to be preferred(Reference Satia, Kristal and Patterson28).
The present study supports findings by Lachat et al. that providing fruit and vegetables from a university canteen led to a higher intake of fruit and vegetables among students both at lunch and on a daily basis(Reference Lachat, Verstraeten and De Meulenaer29). Other studies have shown that an increase in the consumption of fast foods and an increase in the consumption of convenience or ready-prepared foods were associated with less healthy diets(Reference Jabs and Devine30). Bowman and Vinyard compared adults’ mean dietary intake on days having fast food v. dietary intake on non-fast food days and found substantial differences in energy, energy density and macronutrient intakes in favour of the non-fast food days(Reference Bowman and Vinyard31). This seems to be in accordance with consumers’ perception that ready meals and fast food are not seen as appropriate for dinner meals(Reference Prim, Gustafsson and Hall32).
In the present study we have no information on food intake of the participants’ families. The evening meal often has special meaning in a family and is often considered as the most significant meal of the day(Reference Blake, Bisogni and Sobal33). It is likely that the evening meal intake of all participants’ family members was influenced on days receiving CTA. A supportive home environment with access to healthy foods has been found to be an important determinant of healthy eating habits among children and adolescents(Reference Pearson, Biddle and Gorely34, Reference Woodruff, Hanning and McGoldrick35), and Jabs et al. argue that there is a need to develop healthful, affordable, child-acceptable, quickly prepared food that could help parents feel good about the way they feed their families(Reference Jabs, Devine and Bisogni36).
Providing take-away meals from the worksite could provide an alternative convenient and more nutritious food source(Reference Rydell, Harnack and Oakes37) that consumers also might place more trust in compared to other ready meals from the industry(Reference Prim, Gustafsson and Hall32). However, the success of the CTA as a health-promoting activity relies on the actual quality of the meals including the nutritional quality. To match these expectations food service professionals need empowerment. Therefore, an important aim of the present public–private partnership was to develop and disseminate education and practical tools and conditions enabling the canteens to effectively implement and maintain a CTA service, including providing network opportunities and newsletters. Moreover, the simple Healthy Meal Index(Reference Lassen, Biltoft-Jensen and Hansen17) was developed in order to help canteen personnel monitor and improve the nutritional quality of the meals offered without the use of nutrition calculation programs. Often, the canteen staff has neither the time nor the skills to perform complicated calculations(Reference Guilland38). Another important precondition for workers’ adoption of healthy canteen meals is that they are willing to pay the price of these healthy meals, and hence that their perceived benefits exceed their perceived costs. If employees’ willingness to pay does not match the extra costs of supplying such meals, additional financing could be motivated by external benefits for the employer or for society. Price responsiveness and cost–benefit analyses are being investigated by another research team in the partnership project and will be published later.
The methodology of the present study calls for caution when interpreting the results. The present study is an explorative study testing the effectiveness of a health promotion strategy under real-life conditions. This may help the adaptation and expansion of research to practice(Reference Bowen, Kreuter and Spring39). However, the lack of baseline data is a shortcoming as the provision of CTA may have positive or negative nutritional effects on the meal quality and quantity on the days before or after consumption. Participants served as their own control by comparing nutritional intake on days receiving CTA meals with days not receiving CTA. Also, we have no specific knowledge of the lunch eaten but only of the whole day intake. The knowledge of having CTA later could change the employees’ food choice at lunch. The influence of these effects cannot be estimated. Also, the naturalistic experimental approach is often more difficult to manage and often lacks the control that is present in the laboratory(Reference Meiselman40). Accordingly, participants in the present study could not be randomly selected and were not necessarily representative of the population as a whole. They were higher educated compared to the general population and probably more health conscious and had healthier lifestyles. This means that the ability to discriminate between nutritional intakes on different days receiving CTA or not receiving CTA could have been reduced to some degree. Moreover, extrapolation of the results to individuals with different occupational profiles cannot be done. The effects of the present study should be confirmed by a randomized intervention trial.
The present study is the first study to our knowledge to evaluate the effectiveness of the CTA concept in promoting healthy eating. The results are supported by two different dietary assessment methods: assessing total daily intake using 24 h recalls as well as dinner meal intake separately using a digital photographic method for recording in real time over a prolonged period of time. The limitations of 24 h recalls in capturing habitual intake at the individual level include a large day-to-day variation as well as possible under-reporting due to dependence on memory(Reference Palaniappan, Cue and Payette41). Another limitation of the 24 h recall is that socially desirable answers could be higher due to the presence of an interviewer. The mean EI:BMR of 1·4 in the present study likewise indicates that some under-reporting did occur. To avoid bias caused by either underestimation or different energy requirements, the amount of fruit and vegetables and nutrients were adjusted for energy intake and expressed either per 10 MJ or as percentage of total energy intake. More studies are needed to confirm the initial findings of the present study and to further develop feasible and effective environmental-level strategies for health promotion within and across the settings of everyday life.
In conclusion, the results from the present study suggest that providing healthy take-away dinners from the worksite has potential for promoting healthy dietary habits among employees. The nutritional quality of the employees’ evening meals as well as the overall quality of their diet was significantly enhanced on days receiving CTA compared to days not receiving CTA. This reinforces the importance of availability and convenience as effective tools in promoting healthy eating habits and provides worksites with an important role in enabling and promoting healthy eating patterns even across the contexts of worksite and family.
Acknowledgements
The project is part of the Canteen Take Away study funded by the Danish Council for Strategic Research, Programme Commission for Health, Food and Welfare. The authors declare that they have no competing interests. A.D.L. drafted the initial manuscript and conducted the analyses. L.E. and S.P. carried out the data collection and K.K.A. contributed to the statistical analysis of the data. All authors contributed to the study concept and design, interpretation of data, provided critical feedback on drafts and read and approved the final manuscript. The authors would like to thank the canteen staff and all participants who took part in this survey. They also thank Tue Christensen, MSc, and Karin Hess Ygil, dietitian from DTU National Food Institute, for their assistance with data analysis; Dr Lars Ovesen, from Slagelse Hospital, Department of Gastroenterology, and Vibeke K Knudsen, PhD, and Anne Vibeke Thorsen, PhD, from DTU National Food Institute, for their advice and manuscript review.
