10 ways to improve food choice interventions in real-life settings

Over the past decades a lot of behavioural interventions to stimulate healthy and/or sustainable food choices and consumption among consumers have been developed and conducted. Most interventions are focused on providing information on healthy and/or sustainable consumption or on offering educational programmes. Recently, more attention in research is focused on unconsciously changing eating behaviour by creating environments that stimulate healthy and/or sustainable choices.

A lot of research is done, but where are we missing insights? This article presents ten ways to improve food choice interventions in real-life settings.

Knowledge gaps

The ten ways to improve food choice interventions are based on ten knowledge gaps that have been identified within the project Food, Value and Impact. In this project we aim to provide insights into existing knowledge about the effects of environmental interventions that consciously or unconsciously lead to healthy and sustainable consumer behaviour in real-life settings. This has, amongst others, led to the identification of ten knowledge gaps in environmental interventions aimed at unconscious processes to promote healthy and/or sustainable food choices.

1. Need for studies with long-term effects

How long-lasting are the effects of interactions? The longer the effects of interventions are sustained over time, the more successful an intervention can be judged to be. The majority of intervention studies have a short-term perspective. Most study designs include only limited measurements of the target behaviour after the intervention. This makes it difficult to determine what the duration is of the effectiveness of an intervention.

For instance, Bucher and colleagues (2016) discuss the need to study long-term effects of placement of healthy food items at check-out lines. Another example is the need to determine the long-term effectiveness of interventions for worksite health promotion (Kahn-Marshall & Gallant, 2012, Soler et al., 2010). The longer the effects of interventions are sustained over time, the more successful an intervention can be judged to be.

2. Need for objective outcome measures

Many intervention studies use self-reported outcome measures. This can provide valuable information to assess to what extent an intervention has affected consumers’ behaviours (such as food choices). However, self-reported outcome measures have some disadvantages. The problem of such measures is that social desirability can affect them, namely that respondents can tend to report outcomes that are more favourable than they actually are, like overestimating their fruit consumption.

Objective data of the actual food consumption itself, like using biomarkers in urine or internal measure in the mouth or stomach (as opposed to purchases) are still hard to establish and costly.

3. Need for secondary outcome measures in experimental designs

The end-goal of interventions aiming to promote healthy food choices is for consumers to live more healthy lives. However, most studies do not actually measure the effects of interventions on consumers’ health. These insights can be reached through secondary outcome measures like the degree of healthcare utilisation, indicators of cardiovascular health, BMI, or productivity and absenteeism (in case the intervention is conducted in a workplace setting).

4. Need to measure spill over-effects

While many interventions have quite a broad scope in what they aim to achieve, the operationalisation of the target variable often is far more limited in scope. What if an intervention succeeds in making a consumer choose a more sustainable and/or healthy food item - for instance a vegetable snack instead of a candy bar - how does this affect a subsequent choice he or she will make? Will the consumer show licensing effects and choose a less sustainable and/or healthy food item (negative) or will the next food item be more likely to be sustainable and/or healthy as well (positive)?

5. Need to distinguish between different target populations

Interventions vary in the extent to which they target specific groups. It is unclear to what extent certain interventions are effective across different target populations. There is a need to conduct interventions across a more diverse range of target populations to get insight into the generalisability of the intervention and to see if it is necessary to tailor interventions to specific characteristics of a certain population in order to be effective.

Examples of different target populations that have been proposed are low SES groups, non-western populations, adolescents, elderly, minority populations and people from rural areas.


6. Need for more multi-level approaches to find interactions between community, social network and individual level

Multi-level approaches would help to assess how environmental factors in multiple levels might interact with each other. The most common multi-level approach when designing interventions, is one that takes into account the macro (community), meso (social network) and micro level (individuals). This way we can figure out how the effects of for instance individual interventions (e.g., someone who is persuaded to eat more healthy) affect the food choices of his or her social network, for instance partners, family and friends, and vice versa.

Ammerman and colleagues (2002), for example, describe a combination of individual-level interventions, population-level prevention strategies and macro-level state and environmental interventions in order to test the effect of health promotion strategies on (chronic) disease rates.

7. Determining which mechanisms cause effects of interventions

Many interventions involve multiple components, especially when conducted in the field. For instance, an intervention can encompass several alterations to the environment to test what the effect is on consumers’ selection of healthy food items.

However, if the intervention is successful, it then becomes very difficult to establish which of the alterations were critical in causing the effects. Was it for instance because of a change in aisle arrangement, colour use or promotional activities?

8. More emphasis on real-life settings

More intervention studies should be conducted in real-life settings, such as restaurants and cafeterias. The outcomes of studies in real-life settings should then be compared to studies in more controlled settings to get a better indication of both the internal and external validity of an intervention.

9. Focus on implementation of the interventions

There is little insight into how interventions are implemented afterwards. For instance, if an intervention shows that a certain change in the environment (such as aisle arrangement in a supermarket) led to more purchases of sustainable food items, did the supermarket actually change the environment afterwards? And if not, why did they choose to not implement the intervention?

10. Measuring cost-effectiveness of the intervention: What does it cost?

Many intervention studies currently do not take into account the (financial) costs of running the intervention. However, the cost-effectiveness of an intervention can play an important role in how successful it is; a cheap, effective intervention may be considered to be more successful than an intervention that is slightly more effective, but much more expensive to run and implement. Organisations will be more inclined to implement an intervention if they have a clear indication of what it costs and when the costs are relatively low, given the effects the intervention can have on its target population.

More information about the research

In the factsheet you will find more background information about the research method and the used literature.