Physical micro-environment interventions for healthier eating in the workplace: protocol for a stepped wedge randomised controlled pilot trial
- 796 Downloads
An estimated one third of energy is consumed in the workplace. The workplace is therefore an important context in which to reduce energy consumption to tackle the high rates of overweight and obesity in the general population. Altering environmental cues for food selection and consumption—physical micro-environment or ‘choice architecture’ interventions—has the potential to reduce energy intake. The first aim of this pilot trial is to estimate the potential impact upon energy purchased of three such environmental cues (size of portions, packages and tableware; availability of healthier vs. less healthy options; and energy labelling) in workplace cafeterias. A second aim of this pilot trial is to examine the feasibility of recruiting eligible worksites, and identify barriers to the feasibility and acceptability of implementing the interventions in preparation for a larger trial.
Eighteen worksite cafeterias in England will be assigned to one of three intervention groups to assess the impact on energy purchased of altering (a) portion, package and tableware size (n = 6); (b) availability of healthier options (n = 6); and (c) energy (calorie) labelling (n = 6). Using a stepped wedge design, sites will implement allocated interventions at different time periods, as randomised.
This pilot trial will examine the feasibility of recruiting eligible worksites, and the feasibility and acceptability of implementing the interventions in preparation for a larger trial. In addition, a series of linear mixed models will be used to estimate the impact of each intervention on total energy (calories) purchased per time frame of analysis (daily or weekly) controlling for the total sales/transactions adjusted for calendar time and with random effects for worksite. These analyses will allow an estimate of an effect size of each of the three proposed interventions, which will form the basis of the sample size calculations necessary for a larger trial.
KeywordsPhysical micro-environment interventions Choice architecture Nudging Stepped wedge trial Randomised controlled trial Healthier eating Workplace interventions Size Availability Labelling
Reducing excess consumption of food and drink is core to tackling the high prevalence of overweight and obesity in the population . It is estimated that about one third of our daily energy intake is consumed while at work , with the majority of food consumed not brought from home . The National Institute for Health and Care Excellence (NICE) estimates that a company employing 1000 people could lose more than £126,000 a year in reduced productivity solely due to obesity . Whilst a systematic review is underway , there is currently limited evidence of the impact of interventions within workplaces to improve employees’ diets.
In a recent review that included six studies examining dietary interventions in the workplace, four reported small increases in fruit and vegetable consumption following implementation of a range of interventions . However, outcomes were mainly based on self-report and incomplete reporting of all the studies made judgements of risk of bias unclear. In addition, the range of interventions assessed was based largely on providing information (with most studies implementing different kinds of nutritional education campaigns via posters, leaflets, and group workshops delivered at workplaces), commonly regarded as insufficient to tackle obesity . Evidence is now accumulating to suggest that altering cues in small-scale or micro physical environments may provide a more effective way of changing behaviour, often without awareness and in ways that do not place demands on our limited cognitive resources [8, 9]. This is sometimes referred to as ‘nudging’ or ‘choice architecture’ . In keeping with this, a recent report on obesity identified the potential value of a number of such choice architecture interventions .
We propose a pilot study to estimate the potential impact of three sets of physical micro-environment interventions to reduce energy purchased in workplace cafeterias in preparation for a future larger trial. A further aim is to examine the feasibility of recruiting eligible worksites, and identify barriers to the feasibility and acceptability of implementing the interventions in preparation for a larger trial.
Portion, package and tableware size
One promising target for interventions in physical micro-environments is the size of food portions, their packages and the tableware used to consume them. The most robust review to date of the impact of portion, package and tableware size upon consumption [11, 12], suggests that the scale of the ‘portion size effect’ (describing the observation that people consistently consume more when offered larger portions, packages or tableware) is such that, removing larger portions, packages and tableware from the diet could reduce the daily energy intake of UK adults by up to 16% (279 calories).
Availability of healthier vs. less healthy foods
A second intervention is to alter the availability of foods and drinks by increasing the range of healthier foods/drink options and/or decreasing the range of less healthy options. A recent review of interventions in vending machines found that sales of healthier items were increased in five of the six identified studies that increased their availability, with no loss of overall sales volume . Indeed, availability and portion size together represent two of the top three interventions suggested in the McKinsey Global Institute report  as having the highest likely impact across the population. A Cochrane review of the impact of availability interventions in all settings is currently underway .
A third intervention that can be used in workplace settings is to provide labels displaying energy (calorie) content. Recent review evidence shows that calorie labelling at point of consumption can reduce average daily energy intake from food and drinks .
However, recent Cochrane reviews of the three interventions used in this pilot study (11, 14, 15) have not identified any randomised studies in worksite cafeterias. Prior to designing and conducting a larger trial, several key design uncertainties need to be addressed. These include the feasibility of recruiting eligible worksites, the suitability of workplaces for implementing one of the three interventions, the practicalities associated with implementing the interventions including any challenges with implementation or acceptability, and worksites’ ability to provide the required data and comply with the study protocol. The pilot study will also enable us to estimate required sample sizes needed for a larger trial, by providing estimates of likely effect sizes for each of the three interventions and identifying key characteristics and issues with the data that are likely to be encountered.
reducing the size of portions, packages and tableware;
increasing the availability of lower energy (calorie) foods and drinks, relative to higher energy foods and drinks;
labelling foods and drinks with their energy (calorie) content.
The impact upon the total energy (calories) of food and drink purchased in workplace cafeterias of:
smaller portions and packages of foods and drinks, and smaller tableware;
removing some of the higher energy items available and replacing these with lower energy options;
providing energy (calorie) labels on foods and drinks.
The extent to which the impact of the interventions varies with the socio-demographic profiles of the participating workplaces.
These will take place in worksite cafeterias. The precise nature of the interventions will vary between sites, depending upon the food, drink and tableware currently available. For example, if sugar-sweetened beverages are already those at the smallest available size (e.g. a 150 ml can or carton of drink) these will not be a focus for reduction in size. The target products for each intervention will be documented and included as covariates in the analyses.
What is the impact on energy purchased of reducing the size of portions and/or packages of foods and drinks and/or tableware size in worksite cafeterias?The intervention will be applied across a combination of the following:
- (a)Portion sizes of served foods: reduce the size of portions of higher energy food and drink items served in worksite cafeterias by approximately 10% to 15%, including
Main meals (e.g. lasagne)
Sides (e.g. chips or fries)
Desserts, cakes, cookies and biscuits
This might involve producing 16 portions of lasagne from a tray that was previously cut into 14 portions, or 14 slices of cake from a cake that was previously cut into 12 slices.
- (b)Pre-packaged products: replace currently available packaged food and drinks in cafeterias with the next smaller available package size (see Fig. 1 for an example).
Tableware: the size of available glasses, plates, bowls and/or cutlery used to serve and consume food and drink items will be reduced to the next smaller available size. For example, a plate 30 cm in diameter may be replaced by a plate 27 cm in diameter.
For all sites, the standard retail price of the packages and portions will be used. In the case of reducing portion size of items which are not pre-packaged (such as main meals or cakes), prices will be reduced proportionally.
What is the impact on energy purchased of reducing the number of less healthy foods and drinks available in worksite cafeterias and replacing these with the equivalent number of healthier food and drinks?This intervention will involve keeping constant the number of options offered but altering the ratio of healthier to less healthy options by reducing the less healthy foods and drinks (products or units of the same product) available and increasing the healthier foods and drinks available (see Fig. 2 for an example).
The size of the change will depend upon the baseline ratio of healthier to less healthy products. For example, in worksite cafeterias where 75% of available options are classified as less healthy the intervention would likely involve reducing this to 50% of options being less healthy.The terms ‘healthier’ and ‘less healthy’ are used to reflect that these products are either side of a point in a continuum of healthiness, rather than ‘healthy’ and ‘unhealthy’ which might suggest taking products from either extreme (for more details see Table 1). Using only products that are at the extremes of the healthiness continuum would limit the range of items across which the intervention could be applied.Table 1
Example definition of ‘healthier’ vs. ‘less healthy’ options for the availability intervention
Cooked main meal (excluding breakfast)
Complete main meals
500+ Cals a
e.g. beef lasagne, chilli con carne and rice
e.g. salads, vegetable tagine and couscous
Main with side to add
300+ Cals a
e.g. hamburgers, battered fish
e.g. vegetable quiche, chicken breast filled with ricotta
Potatoes: Roast, saute, wedges, chips, mash; Garlic bread; Onion rings; Onion bhaji
Potatoes: Boiled, baked. Rice, couscous
Sandwiches or equivalents
Sandwiches, wraps, panini, baguettes, bagels, pasties
350+ Cals a
e.g. full-fat, baked or vegetable crisps
e.g. popped crisps, popcorn, coconut curls
e.g. standard chocolate bars, biscuits
e.g. selected cereal bars (Nakd), popcorn bars, rice cakes, lower calorie (small) chocolate bars: Milky Way, Freddo, Fudge
Cold drinks (100% fruit juice and smoothies excluded, up to 250 ml bottles)
e.g. full-sugar soda, flavoured water, juice-based drinks, squashes, energy drinks
e.g. water, diet soda, lower calorie drinks (Vit Hit etc)
Broth or tomato based
Dessert pot/ice cream
e.g. mousse, trifle, ice cream
e.g. muffins, cupcakes, sponge sandwich
e.g. teacake, fruit loaf
e.g. crumbles, pies, fruit tarts
e.g. low-calorie jelly, sorbet
What is the impact on energy purchased of adding labels showing energy (calorie) content on items purchased from worksite cafeterias?This intervention comprises labels that state the name of a food or drink item together with the energy content written as ‘XXX Calories’. The latter is to be written in the same font type and size as the product name and/or price information, legible and prominent to the customer, rounded to the nearest 5 or 10 calories, and clearly denoting the portion size to which it refers (see Fig. 3 for an example). In keeping with EU regulations, kilocalorie (kcal) and kilojoule (kj) content will be displayed beneath the ‘XXX Calories’.
Within each of the three sets of six worksite cafeterias, the time at which the intervention is introduced will be randomly determined to control for time trends while maximising sample size. Worksite cafeterias will be allocated to a phase of the stepped wedge design by means of random permutations using random variates of the uniform distribution . The randomisation will be performed by a statistician, with the assistance of computer software.
Number of employees (>350)
Mix of office-based and other site types (to recruit sites that are likely to represent different ranges of socio-economic status)
Ability to provide weekly data on sales of individual items and their energy content
Feasibility assessment measures
Feasibility of recruiting and retaining eligible worksites: measured by recording recruitment rates, and the number of worksite cafeterias dropping-out of the pilot trial (i) post-recruitment, (ii) during the baseline or (iii) during the intervention period.
Feasibility of implementing the assigned intervention: determined after initial visits to worksite cafeterias by the research team, discussions regarding suitability with worksite managers and catering teams, and by examining the sales data supplied by the sites. Qualitative interviews with worksite managers will provide an additional measure of potential challenges with implementation of the intervention.
Acceptability of the interventions: measured by surveying patrons of the worksite cafeterias regarding acceptability of, and other perceptions concerning, the intervention. Qualitative interviews with worksite managers will supplement the survey data, providing insight into the acceptability of study and assessment procedures.
Compliance with the study protocol: assessed via compliance visits coinciding with the initial period of intervention implementation for each randomised worksite.
Intervention impact measures
Total energy (calories) purchased per time frame of analysis controlling for the total sales/transactions.
Total energy (calories) purchased per time frame of analysis from (a) intervention items, and (b) non-intervention items controlling for the total sales/transactions.
Various other measures will be collected that will allow us to estimate the primary and secondary outcomes with greater precision. Covariates examined for all three interventions include: worksite demographic characteristics (age, gender, and dominant occupational status); day of week (if the distribution of the data allows for analysis per day); and weather conditions.
Size: mean energy (calorie) per product pre-intervention; mean reduction (%) in energy (calorie) per product; mean price pre-intervention; mean reduction (%) in price; mean size of tableware pre-intervention; mean reduction (%) in size of tableware.
Availability: proportion of food/drink intervention items that are healthier pre-intervention; proportion of food/drink items that are targeted in the intervention; mean energy (calories) per item pre-intervention; reduction (%) in proportion of intervention items that are healthier from pre-intervention; reduction (%) in energy (calories) per item from pre-intervention; Mean price pre-intervention; mean change (%) in price.
Labelling: mean energy (calories) per item pre-intervention; mean energy (calories) per item during intervention.
Feasibility assessment analyses
Feasibility and acceptability outcomes, including recruitment and attrition rates, will be reported using descriptive statistics. Qualitative assessments gathered via semi-structured interviews with worksite cafeteria staff will be coded and summarised in narrative form.
Intervention impact analyses
One of the main aims of this pilot trial is to estimate the effect size of each of the three interventions, to inform the sample size calculations necessary for a future, larger trial. This will be achieved by conducting quantitative analyses using linear mixed models, which will be reported in full in the scientific paper reporting the results of the pilot trial. By estimating intervention parameters, while taking into account the within-site data dependence and controlling for other parameters, mixed models will help define the effect size of each intervention.
The potential impact of each intervention will be estimated in separate linear mixed models examining the impact on total energy (calories) purchased per time frame of analysis controlling for the total sales/transactions, adjusted for calendar time and with random effects for worksite.
If heteroscedasticity due to different worksite sizes (number of employees) is observed and/or if model assumptions are not respected, outcome change of scale would be considered, as well as alternative analyses considering, for example, weighting worksites according to their variance . If mixed model assumptions are not met, generalised estimating equations (GEE) will be preferred.
Initial analyses will be conducted to confirm the selected time frame level of weekly (as opposed to daily) sales. A weekly time frame will be selected if the conditional distribution of the daily outcome is incompatible with the model assumptions (after transformation), and if time-dependence is still present after controlling for total sales/transactions.
Other analyses will be conducted to inform the design of the main study. These analyses include attempting to estimate the intra-cluster correlation (ICC) using aggregated data, as parallel cluster randomised trials are known to be more efficient than stepped wedge design trials when the ICC is small .
For each of the three interventions, all six worksite cafeterias will initially undergo a baseline period (4 weeks) in which their usual sales are recorded. Sites will then be randomised to implement the intervention at one of six, two-weekly intervals.
The current pilot study is designed to finalise the design and conduct of a future, full-scale trial to assess the impact of three sets of physical micro-environment interventions with the potential to reduce energy consumption in the workplace. As well as providing exploratory estimates of the effect sizes of each of the three interventions, this pilot study will address key design uncertainties for the trial including the feasibility of recruiting eligible worksites, the suitability of workplaces for implementing one of the three interventions, the practicalities associated with implementing the interventions including any challenges with implementation, worksites’ ability to provide the required data and to comply with the study design. The pilot study will also allow us to explore the acceptability of the study procedures and assessment methods.
In addition to the strengths described above, this pilot trial has several limitations. First, the outcome measures pertain to purchasing and not consumption of energy. Although purchasing provides an indication of consumption in the workplace, it does not take into account food obtained from other sources or food waste. Second, this study will not directly measure change in the purchasing of individual workers, since the units of randomisation and analysis are the individual worksite cafeterias. Finally, this study measures changes in purchasing over a relatively short time-period; any longer-term impact of the interventions on worksite cafeterias (or the employees using them) will not be assessed. This pilot trial will nevertheless provide an indication of whether a longer follow-up in the larger trial is feasible, as well as the possibility of supplementing the main purchasing measures with complementary self-report measures of employees’ diets over time.
The stepped wedge design has strengths and limitations. A strength of this design is that it combines features of within- and between-subjects designs thus allowing the examination of changes in purchasing within each worksite cafeteria depending on the time-period (baseline vs. intervention). It also allows examination of differences in purchasing between worksite cafeterias at different time-points. Furthermore, in the current context, the stepped wedge design is expected to have more statistical power than a parallel group design. It is also more suitable than a multiple-treatment reversal (ABAB) design, which would alternate intervention and intervention free periods, as the effect of an intervention may extend beyond the strict intervention time frame. A notable limitation of a stepped wedge design is the potential for an underlying temporal trend to confound the intervention effect given more participating units are exposed to the intervention at later than earlier time-periods .
This pilot trial will inform a future larger trial that is expected to extend current knowledge regarding the effectiveness of physical micro-environment interventions (size, availability, and labelling), applied singly and together, to reduce the energy purchased in worksite cafeterias.
The study is funded by the Department of Health Policy Research Programme (Policy Research Unit in Behaviour and Health [PR-UN-0409-10109] and the Institute for Grocery Distribution [RG83425]). RP is supported by a Wellcome Trust Research Fellowship in Society and Ethics [106679/Z/14/Z]. The funders had no role in the study design, data collection, analysis, or interpretation. The research was conducted independently of the funders, and the views expressed in this protocol are those of the authors and not necessarily those of the Department of Health in England or other government departments.
Availability of data and materials
All authors collaborated in designing the study. MV drafted the manuscript, EC, RP, GJH, DLC, SAJ and TMM provided critical revisions to the manuscript. All authors read and approved the final manuscript.
The authors declare that they have no competing interests.
Consent for publication
Ethics approval and consent to participate
Approved by the Psychology Research Ethics Committee of the University of Cambridge (Reference Number: PRE.2016.035).
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
- 2.Boost your health at work. http://www.nhs.uk/Livewell/workplacehealth/Pages/workplaceoverview.aspx. Accessed 21st Sept 2016.
- 3.Dietitian-led wellbeing for employers and workers. http://www.bdaworkready.co.uk/. Accessed 21st Sept 2016.
- 4.National Institute for Health and Care Excellence. Workplace health: management practices. 2015. www.nice.org.uk/guidance/NG13. Accessed 21st Sept 2016.
- 7.McKinsey Global Institute. Overcoming obesity: an initial economic analysis. 2014.Google Scholar
- 11.Hollands GJ, Shemilt I, Marteau TM, Jebb SA, Lewis HB, Wei Y, Higgins JPT, Ogilvie D. Portion, package or tableware size for changing selection and consumption of food, alcohol and tobacco. Cochrane Database Syst Rev. 2015;9:CD011045.Google Scholar
- 14.Hollands GJ, Carter P, Shemilt I, Marteau TM, Jebb SA, Higgins JP, Ogilvie D. Altering the availability or proximity of food, alcohol or tobacco products to change their selection and consumption. Cochrane Database Syst Rev. 2017;3:CD012573.Google Scholar
- 15.Crockett R, Hollands G, Jebb S, Marteau TM. Nutritional labelling for promoting healthier food purchasing and consumption (Protocol) Cochrane Database Syst Rev. 2011;9:CD009315.Google Scholar
- 17.Woodward M. Epidemiology: Study Design and Data Analysis, Chapman & Hall/CRC Texts in Statistical Science. Boca Raton: Taylor & Francis; 2014.Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.