Contextual Specificity of (Un)Healthy Food/Drink Intake in Everyday Life: A Study Based on Episodic Memories

1 Introduction

Mexico is among the countries with the highest prevalence of cardiometabolic diseases such as obesity, type 2 diabetes, hypertension and dyslipidaemia (Lozano et al., 2013). The prevention and therapy of these chronic conditions includes primarily modifiable factors like lifestyle habits. In addition to healthy eating guidelines for the population (Secretaría de Salud, 2012), current nutritional policy includes taxation and front-of-pack warning labels for high-energy dense food/drink products (Barrientos-Gutierrez et al., 2017; White & Barquera, 2020). These policies are, however, limited to only a few of the multiple factors involved in eating behaviour (Dhurandhar et al., 2015a). Solutions are needed that address more broadly the immediate context of food/drink consumption. Identifying contextual features of common patterns of eating/drinking in a culture is a first step towards further investigation of environmental influences on such prevalent behaviour in order to inform nutritional policy interventions.

People remember well everyday eating episodes from the preceding week span. Such recent episodic memories are very accessible near-experience perceptions of past personal events (Conway, 2009; Zacks, 2020). There is good agreement between records in real time and recalls of recent daily activities (Kristo et al., 2009; Radvansky et al., 2022), including eating episodes (Armstrong et al., 2000). Likewise, wider research shows that although errors occur about what is perceived, these are small (e.g. Jussim, 2012).

Misconceptions are common about retrospective and self-reported data. For example, errors in autobiographical memory involve events dated many months/years ago (Brewer, 1994). In the case of self-report, biases relate to leading questioning (Schwarz, 2007) as well as opportunity to retrieve information from semantic memory (Schacter & Madore, 2016). These errors are minimising targeting memories of recent autobiographical episodes (Conway, 2009) as well as using non-suggestive open-ended questions (Fisher & Geiselman, 1992). Moreover, inducing the recall of features of a past autobiographical episode further enhances retrieval of real details of that event (Schacter & Madore, 2016). It is the estimation of portion sizes, energy and nutrient intake from recalled or self-reported intake that can be grossly inaccurate (Dhurandhar et al., 2015 b; Krall & Dwyer, 1987; Krall et al., 1988). However, this may partly result from people recalling or reporting the most salient foods/drinks as well as contextual features that configure a past eating episode (Alba & Hasher, 1983; Brown-Kramer et al., 2009; Zacks, 2020).

A memory of a recent eating episode is elicited with questions that prompt perceived objective (such as place, day and time, people present, food consumed, etc.) and subjective (such as mood, appetite, healthiness, etc.) features of the event (Fisher & Geiselman, 1992). Words selected by people to describe these features of an everyday eating episode within a culture are ecologically valid (Booth & Booth, 2011). An indicator of this validity is the percentage of people who mention the same or similar terms for a feature of a particular episode. A report of a past personal event can be verified by video recording that instance from start to end in order to test the match between recalled and actual physical features of the behaviour scenario (Mair et al., 2017). There is also the perspective that the subjective experience (and its recounting), including awareness of salient features of an event, is an objective performance of analogies with a shared reality with others (Booth, 2003).

The ability to re-construct in memory a salient scheme of a past autobiographical (eating) episode provides the basis for behaviour change approaches. For instance, the habit theories propose that memories of repeated behaviour in consistent context influence or guide behaviour when features of that context are encountered (Booth, 2022; Gardner, 2015; Papies et al., 2022).

Self-reported memories of recent eating episodes have shown differences between (un)healthy meals that individuals consume in normal life. While reports of healthy eating episodes were characterised by more mentions of water, fruits and vegetables consumed at home with family members, those of unhealthy eating episodes were characterised by more mentions of high-calorie foods/drinks consumed out of home with friends (Laguna-Camacho & Booth, 2015; Laguna-Camacho et al., 2018). These findings agree with those from momentary ecological assessments showing that people consume high-calorie food/drink in company of other people out of home (Elliston et al., 2017; McNaughton et al., 2020). However, limitations of the momentary ecological assessment are behavioural reactivity from recording an eating episode in real time (Booth & Laguna-Camacho, 2015) and high cost of mobile technology. Also, previous studies have only assessed context and intake of general rather than particular sorts of (un)healthy eating episodes.

Correlational analyses can be applied to further explore connections between self-reported contextual features and intake of eating episodes. For instance, former studies from food diaries showed positive correlation between number of people present and estimated calorie intake at eating episodes (de Castro, 1994). There is additional evidence for negative correlation between perceived healthiness and estimated calorie intake of recalled eating episodes (Laguna-Camacho et al., 2018), indicating that people are sensitive about how healthily they eat/drink.

The aim of the present work was to assess the characteristic context and intake of (un)healthy eating episodes recalled by adults from the (urban) Mexican midlands. The expectation was that similar to former studies (Laguna-Camacho & Booth, 2015; Laguna-Camacho et al., 2018) there would be in general more consumption of healthy food/drink with family at home and more consumption of unhealthy food/drink with friends out of home, however, there would be a more detailed characterisation of these features sorting (un)healthy eating episodes according to conventional meal periods of the day.

The present work assessed also if there was a relationship between the basic contextual features (place, time and people present) with the estimated calorie intake across recalled (un)healthy eating episodes, as wider research indicates (e.g. Lachat et al., 2012). The expectation was that higher estimated energy intake would correlate with carrying out episodes out of the home, later in the day and with people present. Also, the perceived healthiness of eating episodes was expected to correlate negatively with their estimated energy intake, as previously observed (Laguna-Camacho et al., 2018).

The current investigation explored additionally if “meals” differ from “snacks” simply by drawing a line a priori through energy estimates (e.g. at 250 or 300 kcal) because “meal” means substantial intake and so meal periods should usually have larger intakes than between-meal periods.

2 Methods

3 Results

3.5 Relationships between Contextual Features, Kilocalorie Intake and Healthiness

As expected, eating out of home, later in the day and with more people correlated positively with estimated kilocalorie content across (un)healthy eating episodes (Table 6). Also, the perceived healthiness of (un)healthy eating episodes correlated negatively with their estimated kilocalorie content (Figure 1).

Table 6

Correlations between remembered context, estimated energy content and perceived healthiness across recalled healthy and unhealthy eating episodes (N = 816)

Model variables	1	2	3	4	5
1. Place	—	0.04	0.07*	0.17**	−0.33**
2. Time of day		—	0.15**	0.19**	−0.26**
3. People present			—	0.18**	−0.11**
4. Kilocalories				—	−0.31**
5. Healthiness					—

*p < 0.05, **p < 0.01.

Figure 1

Association between perceived healthiness and estimated energy content of recalled healthy and unhealthy eating episodes (see corresponding statistics in Table 6).

3.6 Energy Intake at Meals and between Meals

For healthy eating episodes, mean estimated energy consumption was higher at meal than between-meal periods, however, for unhealthy eating episodes, no difference in mean estimated energy consumption between meal and between-meal periods was observed (Figure 2).

Figure 2

Estimated energy consumption of recalled eating episodes at meal and between-meal periods. (Left) Episodes at meal periods n = 390, episodes between-meal periods n = 57. F = 5.44, p < 0.02. (Right) Episodes at meal periods n = 240, episodes between-meal periods n = 129. F = 1.1, p = 0.30.

4 Discussion

The present investigation assessed the context and food/drink intake of sorts of (un)healthy eating episodes recalled by Mexican adults. There were more occurrences at home for healthy than unhealthy breakfasts, episodes between breakfast–lunch and lunches, and more occurrences out of home for all sorts of unhealthy than healthy eating episode except episodes between breakfast–lunch. Healthy eating episodes featured, however, less at home between breakfast–lunch than at other times of the day, which would be the interval when most participants attended school/work. For lunches and dinners, there were more occurrences of healthy than unhealthy eating with family and more occurrences of unhealthy than healthy eating with friends. Hence although there is a general tendency for healthy eating episodes to be with family at home and unhealthy eating episodes to be with friends out of home (Laguna-Camacho & Booth, 2015; Laguna-Camacho et al., 2018), the findings of the current study confirmed, as hypothesised, a particular context by sort of (un)healthy eating episode.

The present study identified as well particular food/drink intake that varied in mentions by sort of (un)healthy eating episode of the day. These variations had to do most likely with the patterns of intake, i.e. usual preparations, menus or combos of food/drinks in the studied culture. Meat and vegetables were common in all sorts of healthy eating episode, indicating a particular pattern of healthy intake. There were more mentions of pasta, Mexican homemade food, tortillas and water in healthy lunches and episodes between lunch–dinner, which would relate to having a typical meal after attending school/work. Fruit, cereal, milk, yogurt, egg preparations and sandwich were less common in healthy lunches than in other sorts of healthy eating episode. Mexican street food was mentioned more in all sorts of unhealthy eating episodes, which indicated a popular pattern of unhealthy intake. Fast food was mentioned less in unhealthy breakfasts than in unhealthy eating episodes at other periods of the day, possibly because most fast food outlets operate after breakfast time. Savoury snacks and sweets were commonly mentioned in unhealthy eating episodes, particularly in episodes between breakfast–lunch and between lunch–dinner. Soft drinks were common in all sorts of unhealthy eating episode, which is consistent with the reported high-sweetened beverage consumption in Mexican population (Sánchez-Pimienta et al., 2016). Such identified categories of contextual features and food/drink intake in (un)healthy eating episodes contrast with intake patterns in epidemiology based on energy estimations from food frequency questions out of context (Schulze et al., 2018).

Even so, there is still a more specific context and food/drink intake for the (un)healthy eating episodes than the assessed in the current study. The higher intake at meal than between-meal periods found for healthy but not unhealthy eating episodes indicated that intake patterns were highly variable through the day and perhaps beyond the conventional classification of intakes into meal and between-meal episodes held by nutritionists. Here the categorisation of contextual features was simplified and so a more detailed categorisation is needed of the places, sorts of occasion and people present that characterise (un)healthy eating episodes. Combinations of (un)healthy food/drink intake categories could also have been identified in the present work. Hence future research is warranted on characterisation of context of specific eating patterns or habits.

In addition, the present study found the expected positive correlations of eating out of home, later in the day and with people present with estimated calorie intake of recalled (un)healthy eating episodes and negative correlation of estimated calorie intake of recalled (un)healthy eating episodes with perceived episode healthiness. These findings extend former evidence (de Castro, 1994; Laguna-Camacho et al., 2018) and show the usefulness of data from recall of recent eating episodes for testing relationships between features that characterise those events.

Further evaluation is required on how notions about inaccuracies in retrospective reports of eating episodes relate to theoretical and methodological deficiencies (Booth, 2022; Brown et al., 2021; Munafò et al., 2017; Podsakoff et al., 2003; Scheel et al., 2020, 2021). Recent research applying open science practices supports the precision of episodic memory. For instance, a registered report found losses in accessibility but not in precision of recalled features of personal past events that are remembered (Berens et al., 2020). Moreover, memory could be better for eating than non-eating episodes (Seitz et al., 2021). Nonetheless, research is still needed on biases in self-reported memories of recent eating episodes. For instance, distortions in recall of eating episodes can surge from semantic memory (Schacter & Madore, 2016). Demand effects may contribute to consistent reports, e.g. when asked about detail(s) of a past personal event, to construct such answer people would fill gaps in episodic memory with semantic schemas or beliefs (e.g. Delivett et al., 2022) mainly if little attention was paid to such detail(s) during the episode.

Since the findings of the present work were mostly derived from undergraduate students in Mexican midlands, potential applications are limited to this group. The identified contexts of (un)healthy food/drink intake can help formulate guidance about what foods/drinks these student group can eat, where, when and with whom, as well as to design scenarios for eating more healthily in their campus and surroundings.

The present study supports the value of systematically assessing memories shared by people about their everyday eating episodes. This work moved towards a cultural perspective to describe the context of customary eating patterns. Dietary recommendations based on contextualised food/drink combinations are a novel intervention for improving eating habits in a culture (e.g. Booth et al., 2004; Laguna-Camacho & Booth, 2021; Laguna-Camacho & Serrano-Plata, 2021). Monitoring episodes of such specified eating patterns applies also to measure the effect on health outcomes of maintained change in their frequency as well as to identify contextual triggers for lapsing (Booth & Booth, 2011; Booth & Laguna-Camacho, 2022).