COVID-19 home isolation and food consumption patterns: Investigating the correlates of poor dietary diversity in Lebanon: a cross-sectional study

Statistical analysis

Respondents’ characteristics classified as categorical variables were presented as frequencies (percentages) while means ± standard deviation (SD) were used for continuous variables. Different statistical tests were used: The chi-square test was used to investigate differences for categorical variables between groups (gender) while an independent t-test was applied for continuous variables. A p-value lower than 0.05 was considered significant. Statistical analysis was conducted on IBM SPSS Statistics (IBM Corp, SPSS statistics for Mac, Version 24. Armonk, New York) (RRID:SCR_019096).

Ethical consideration

A consent form was attached at the beginning of the online survey and was required to be completed before continuing. The consent form protected participants by letting them know their rights and responsibilities while keeping their information confidential. Moreover, this study received the approval of the ethics committees at the University of Antwerpen (SHW_46) (the country leader of this project) and at all the concerned countries, including Lebanon (#CUER 22-2020) given that it was observational with respect of confidentiality and no traceability of respondents.

Socio-demographic and economic characteristics

The survey was completed by a convenient sample of 2449 Lebanese participants. A total number of 167 participants were excluded from the study due to either: being aged below 18 years or not completing the survey. Thus, 2282 Lebanese participants were included in the present study.⁴³ Among them, 80.9% were females. Most participants were either adults (25 to 46 years old, 52.1%) or youth (19 to 20 years old, 37.7%). A very low percentage of adolescents (18 years) (8.4%) and elderly people (1.8%) had registered (Table 1). The mean age of the respondents was 29.36 years with a SD of 12.221. Males who responded to this survey were significantly older than females who did (p-value=0) (Table 1). In regard to the educational level, around 40.4% of the respondents had a bachelor’s degree when analyzing both genders together. In addition, a significant higher percentage (41.1%) of females had a bachelor’s degree compared with males (37.4%) (p-value=0) (Table 1). The household composition was also analyzed before and during the lockdown in which most households were composed of three to five adults (more than 45%). This trend has been also observed in males and females each separately, except for female living in the same household before the lockdown in which the majority were composed of less than three female adults per household (47.1%) (Table 1). When looking at economic characteristics, a similar percentage of respondents were students (40.5%) and active workers (40.9%) while the minority were unemployed (18.6%) before the COVID-19 lockdown. The percentage of unemployed individuals had increased to 32.6% during the lockdown. This increase has been also observed for males (8.7% before lockdown to 28.5% during lockdown) and females (20.9% before lockdown to 33.5% during lockdown). Moreover, the COVID-19 lockdown has also induced a loss of income among 62.9% of respondents. This loss of income was significantly higher (p-value=0.007) for men (68.6%) compared to women (61.6%). In addition, most respondents, when taken either all together (73.8%) or categorized as males (74.3%) and females (73.7%), struggle to make their wages last a month. And most even struggle to earn enough money for shopping (65.6%, 63.1% and 66.1% for all respondents, men, and women respectively) (Table 1).

Consumption of food groups and food consumption score

Tables 2 and 3 show the food groups consumption and the FCS in the overall population, in Lebanon.

Fruits group

Regarding the consumption of fruits (fresh or frozen), 57% of the Lebanese population consumed fruits equals or less than 4 times per week during the lockdown, compared to 51.6% before the lockdown (p-value=0). It was also observed that the percentage of people consuming fruits five times and more per week, was relatively higher before the lockdown (48.4%) compared to the consumption during the lockdown (43%; p-value=0) (Table 2).

Vegetables group

As seen in Table 2, it appears that 64% of the population had higher consumption of vegetables (more than five times per week) before the lockdown compared to 57.1% during the lockdown (p-value=0). This statistic lead to the increase in the number of people in the four times or less category observed during the lockdown (42.9% vs 36% before the lockdown; p-value=0).

Legumes and pulses group

81% of people living in Lebanon, from the current study, were consuming legumes and pulses equals or less than 4 times per week before the lockdown; this category saw a decrease during the lockdown (77.8%). Consequently, the percentage of people frequently consuming legumes and pulses (five times and more per week) increased during the pandemic by 3.2% (22.2% during vs 19% before; p-value=0.001) (Table 2).

Nuts group

As per Table 2, the percentage of people frequently consuming nuts and derivatives was higher before the lockdown (15.9% before vs 14.7% during), while, a slight decrease of 1.2% was observed during the lockdown; 85.3% of the population was consuming nuts and derivatives 4 times or less during the lockdown compared to before (84.1%). Yet, these results were not statistically significant (p-value=0.147).

Processed meat/poultry/fish/vegetarian alternatives

The frequency of consumption of processed meat/poultry/fish/vegetarian alternatives was higher (more than five times per week) before the pandemic (18.6% before vs 12.8% during; p-value=0). In addition, a decrease of 5.8% in the consumption of this food group (four times or less per week) was observed during the lockdown (87.2% during vs 81.4% before; p-value=0) (Table 2).

Unprocessed fish, unprocessed poultry, and unprocessed meats

It was observed that the frequent consumption (five times and more per week) of unprocessed fish was higher during the pandemic (6.2% during vs 5.7% before) and less frequent consumption (equals to 4 times or less per week) was higher before the lockdown (94.3% before vs 93.8% during). Yet, the frequency of consumption of unprocessed poultry and meat remained unchanged before and during the lockdown (Table 2). However, these results were not statistically significant (fish-p-value=0.366, poultry-p-value=1 and meat-p-value=0.948).

White and whole wheat, bread, pasta, and grains

The frequency of consumption of wholewheat bread, pasta and grains was higher during the lockdown (25.8% during vs 23% before), yet this consumption was lower before the lockdown; 77% of the Lebanese population was consuming this food group equals to four times or less before the lockdown compared to 74.2% during the lockdown (p-value=0.003). As for the frequency of consumption of white bread, pasta and grains, there was no significant differences before and during the pandemic (p-value=0.79) (Table 2).

Milk and dairy products group

The consumption of dairy products was lower during the pandemic; 50.6% of the Lebanese population was consuming dairy products equals to four times and less during the lockdown vs 45.5% before. Moreover, we noticed a remarkable decrease during the lockdown, of 5.1% in the consumption of other dairy products (54.5% before vs 49.4% during; p-value=0.003). As for the milk consumption, no significant differences were stated before and during the pandemic (p-value=0.433) (Table 2).

Sugar group (products and beverages)

The frequency of consumption of sugary beverages and products in the studied population was high before the lockdown; 27.1% and 43.8% of the population consumed sweet snacks (p-value=0.001) and sugary beverages (p-value=0) respectively, five time and more per week (Table 2). However, a decrease of 3% in the consumption of this group was noted during the lockdown (snacks-p-value=0.001; beverages-p-value=0) (Table 2).

Fats and oils group

A decrease by 3% in the consumption of fats and oils was observed during the lockdown (82.3% during vs 79.5% before; p-value=0.001). It was also observed that added fats and oils were more frequently consumed in the Lebanese population before the lockdown (20.5% before vs 17.7% during; p-value=0.001) (Table 2).

Calculation of the food consumption score

The FCS was calculated based on the equation explained previously in the methods. Compared to the period preceding the pandemic, the mean levels and standard deviation (SD) of the FCS in Lebanon was equal to 93.9±41.9 and it decreased to 88.1±44.4 during the pandemic (p-value=0). Additionally, the percentage of people having low FCS has increased during the lockdown from 9.2% to 13.8% (p-value=0). These results are presented in Table 3.

Food consumption patterns and food-related behaviors

Cooking practices and barriers

Table 4 shows that most cooking attitudes and practices (10 out of 13) showed a significant increase in the intra-COVID-19 pandemic. The highest increase was recorded for the response “not having access to foods/ingredients needed or wanted” (36.2% positive answers before lockdown compared to 51.1% during lockdown) (p-value=0). The only practices that showed a significant decrease during lockdown was “throwing away food leftovers” (30.4% before lockdown compared to 26.6% during lockdown) (p-value=0). Yet, a slight decrease was stated during the lockdown, in the attitude of “feeling confident about managing money to buy healthy food” and “feeling confident about cooking a variety of healthy meals” of 0.8% and 0.3% (p-value=0.474 and p-value=0.706) respectively. Moreover, it was observed that the attitudes of “don’t having the funds for the foods/ingredients needed or wanted” and “don't having the access to the facilities needed to cook or bake such as stove, oven, kitchen equipment” increased by 11% and 9.6% during the lockdown (p-value=0 and p-value=0), respectively, despite the null change in the practice of “cooking meals using healthy ingredients” (p-value=0.611). Similarly, a significant increase of around 8.7% and 6.7% was observed in the practice of “cooking with leftover food” and “using other parts of food label to make food choices” (p-value=0 and p-value=0), respectively. More than 66% of people were “changing recipes to make them healthier”, were “thinking about healthy choices when deciding what to eat “and were “planning meals to have a varied diet”. Additionally, 52.9% of people were “using the nutritional information panel (nutritional breakdown of the products) to make food choices “(Table 4).

Criteria for recipe selection

With regards to the criteria for recipes selection, the lockdown was found to significantly increase the percentage of agreeing responses for all criterions. For instance, during the lockdown, more than 70% of people selected their recipes had few ingredients (73.7%) that are easily available at home (81%) or at store (79.8%), inexpensive (71.8%) or at good price (81.6%), and healthy (77.5%) (Table 5). Before the pandemic, more than half to three quarters of the people living in Lebanon, agreed that they were selecting recipes that were achievable with few ingredients, were available at home or can be easily found at the store, inexpensive to prepare, healthy and cheap. During the lockdown, an increase ranging between 2.3% and 13.1% was observed in these patterns. The selection of recipes that were based on ingredients easily available at the store, was the only decrease and saw a decrease of 1.2% (Table 5).

Shopping

Shopping practices have also been affected during the lockdown in Lebanon. Indeed, more than 76% of respondents agreed that they search more for cheapest prices before and during the lockdown (Table 5).

During the lockdown, respondents showed a significant reduction of 7.3% less going physically to select and buy food, and that they preferred to order their food products online (variation of 3.6%) they also preferred having it delivered to their home rather than being delivered at a seller’s point (4.1%) (p-value=0.008). Regarding places of groceries shopping, there was a significant decrease during the lockdown for all the places except for shopping straight from the farmer/producer. In addition, it is remarkable that respondents had shown a disinterest in buying food at organic/fair trade shops or specialty stores during lockdown (p-value=0) (Table 6).

Food stock

Stocking up foods was also affected during the pandemic. Thereby, we observed a huge increase in stocking up pasta, rice or other grains, for flour, and for legumes/pulses (such as beans, lentils, chickpeas: dried or tinned). This increase was less significant for water, potatoes, yeast, milk, bread, vegetables, and meat. On the other hand, ready-made meals, salty snacks, nuts or nut spread, fish (fresh, frozen and canned), sweet snacks, and plant-based drinks were stored in smaller amounts during the lockdown. Furthermore, lockdown did not show any impact on stocking up non-alcoholic drinks, eggs, other dairy products, fruits in any of its forms as well as vegetarian alternatives (Figure 1, Table 7).

Figure 1. Food stock patterns during the pandemic in Lebanon.

Determinants of food insecurity in the overall population

Many factors affected the FCS of the people living in Lebanon. Before the pandemic, the percentage of people having low FCS was 9.2% and increased to reach 13.8% during the pandemic (Table 3). The determinants of poor dietary diversity indicated by “low FCS” in the overall population, before and during the pandemic, is conditioned by many variables including the gender, age categories, some cooking practices, cooking with leftovers, education, and employment. Prior to the pandemic, and compared to adolescents (18 years), elderly people were 86% less likely to be affected by the decrease in the FCS [OR:0.14; 95% CI (0.01-1.07)]. The cooking practice that affected the FCS of the population studied the most, was meal planning. 55% of people who planned their meals to include a variety of food [OR:0.45; 95% CI (0.31-0.65)] were less affected by the decrease of the FCS compared to those not practicing this pattern. As for the education, it appears that people with a high school diploma [0.3; 95%CI (0.11-0.82)] were 70% less likely to be affected by the decline in the FCS compared to those who attained lower levels of education. During the pandemic, the binary logistic regression analysis showed that the decrease in the FCS doesn’t seriously affect the elderly [OR: 0.17; 95%CI (0.04-0.72)], when compared with adolescents. Similarly, the decrease in the FCS didn’t affect people with a high school diploma [OR:0.26; 95% CI (0.1-0.66)] and a bachelor degree [OR:0.37; 95% CI (0.15-0.92)], when compared to people with a lower educational level. It was noticed that there was no impact of gender, cooking practices (except for planning meals to include a variety of food before the pandemic), cooking or throwing away leftovers, and employment on the FCS, neither before nor during the lockdown (Table 8).

Comparison with other countries

When comparing our results with findings from other countries’ and studies, we observed that the home isolation due to the pandemic induced changes in food-related behaviors in Russia, of which a decrease of meat consumption and sweet products was observed in 1047 adults along with an adoption of healthier consumption patterns.²¹ In addition, an American survey (n=484 adult participants), suggested people facing food insecurity reduced their intake of fruits and vegetables since the start of the pandemic.²² The findings of the American study, as well as our findings, supported two Italian surveys. One of which stated 18% of the respondents reported a decrease in the consuming of fresh fruits while in another study, 8.7% of the respondents reported a decrease in the consumption of fresh fruits and vegetables.^23,24 However, a third study conducted in Italy stated that 2768 adults showed an improvement in their diet quality in which they increased their consumption of fruit (24.4%), vegetables (28.5%), legumes (22.1%), nuts (12%), and fish or shellfish (14%). On the other hand, an excessive consumption of sweets or pastries (36.9%) increased in this study.²⁵ A survey by Enriquez-Martinez et al., conducted in Argentina, Brazil, Mexico, Peru, and Spain on 6,325 adults revealed that most Spanish participants (61.6%), didn’t show any improving or worsening in their food pattern.²⁶ People living in Argentina and Brazil reported some amelioration in their diets. Lower change in food consumption patterns was observed among Peruvians and Mexicans (OR: 0.51; 95% CI: 0.4–0.6 and OR: 0.69; 95% CI: 0.4–0.8, respectively), when compared to Argentinians.²⁶ A Polish survey enrolled in Poland (n=2381 adults) shows that 30% of respondents improved their food patterns through a rise in the consumption of vegetable, milk and other dairy products. In addition, the consumption of fast food, and commercial pastry along with confectionary and salty snacks decreased in more than 50% of respondents. Moreover, one-fifth of the respondents reduced the intake of sugar-sweetened beverages and alcohol.²⁷ In France, in comparison to the period preceding the pandemic, the nutritional quality of the respondents (n=938) diets was affected strongly; a sharp increase in processed meat, sweet-tasting beverages, and alcoholic beverage consumption had negatively decreased the overall diet quality.²⁸ In addition, a French web-based survey that encompassed 37,252 adults showed a decline in the intake of fresh food of where 17% of participants reported a decrease in consumption of fresh fruits, 18% reported a decrease in consumption of fresh vegetables, 22% in fresh red meats, and 31% in fresh fish. On the other hand, 22% of people reported an increase in the consumption of sweets and chocolate, 20% in cookies and cakes, and 18% in cheese along with a decline in the consumption of sandwiches, pizzas, or savory pies (17%).²⁹ The Lithuanian COVIDiet Study (n=2447 participants) that revealed frequent and high consumption of sweets, biscuits, cakes, frozen and canned foods and low intake of fruits and vegetables, also reported frequent cooking and eating out of control. However, this was associated with some positive changes such as high consumption of fresh fruits, vegetables, fish, legumes, and white meat and low consumption of processed meat, carbonated and sugary drinks.³⁰ The findings from a survey conducted by three European countries (n=1071 adults): from Poland (n=407), Austria (n=353) and the United Kingdom (n=311) showed a frequent purchase of frozen foods with long shelf life and high consumption of dairy products, grains, fats, vegetables, and sweets (p-value<0.05).³¹ In Belgium, (n=8640 adults), 10.4% of participants reported food shortages and 10.3 % were incapable of eating a nutritious diet during lockdown. This status of food insecurity was associated with a change in most dietary behaviors.³² The people living in Denmark, Germany, and Slovenia witnessed, during the pandemic, a decline in the intake of fresh food.³³ Three studies in Brazil were conducted between 2020 and 2021 in which more than 50,000 adults were interviewed through web-based questionnaires. The first study highlighted the increase in consumption of processed and energy dense foods (potato fries, chocolate, and ice cream) among Brazilian adults.³⁴ In the second study, an increase in the consumption of vegetables, fruits, and legumes along with a steadiness in the consumption of processed foods was observed.³⁵ The third study described the reduction in consumption of fruits and vegetables along with an increase in the consumption of candies and fast-food.³⁶ In Chile (n=700 adults), it was shown that there was a low consumption of legumes and water and high consumption of junk food (e.g., food with low food quality, low contribution of micronutrients and with a high contribution of sugar, saturated fat, and sodium) and fried foods.³⁷ In China (n=2702 adults), many participants didn’t witness any modification in their habitual diet, while 38.2% increased their snack intake, during their home isolation. This could be explained by the fact that basic food supplies were offered by the government in China during the lockdown.³⁸ Similarly, in the Netherlands, 83% of participants did not modify their eating habits during their confinement.³⁹ A scoping review on many populations (US, Asia including Palestine, India, and China, Europe including Italy, France, Spain, Poland, and the UK, Australia, and Zimbabwe) showed an increase in the consumption of meals and snacks, an amelioration in dietary patterns associated with an increase in fresh produce and home cooking and reductions in comfort food and alcohol consumption. However, in the same review, nine studies showed a reduction in fresh produce, with a further six describing an increase in the consumption of comfort foods such as sweets and processed foods. An increase in alcohol consumption was reported in two studies and weight gain was reported in fifteen studies a reduction in physical exercise.⁴⁰

Comparison with other Eastern Mediterranean regional countries

At regional level, in Jordan, among a total of 3129 Jordanians, 23.1% of the respondents were facing severe food insecurity, 36.1% were moderately food insecure and 40.7% were food secure. Food insecure people were consuming fewer meats and carbohydrates compared to food secure people.¹¹ In Iran, the household’s dietary diversity and food security status were studied and a decline in the consumption of white roots and tubers, dark green leafy vegetables, other fruits, organ meat, legumes, nuts and seeds, sweets, spices, condiments, and beverages was shown.⁴¹ In addition, the impact of COVID-19 pandemic on food purchasing and dietary behaviors was studied in three Kuwaiti surveys with 1935 respondents in Kuwait. In the first study, an increase in the consumption of vegetables, fruits, and carbohydrates along with a decrease in the consumption of fish and sugary drinks was observed.⁴² In the second study, a change in the consumption of only fish and seafood was observed.¹⁷ In the third study, 50% of respondents reported no changes at all in the consumption of all food groups (44%). In addition, more than half the participants ate more fruits and vegetables, 41.5% ate more legumes and pulses, and 2.3% ate fewer fast food.¹⁸ In Lebanon, more than 35% of people did not eat fruits and vegetables daily. In addition, 28% consumed sweets or desserts and 24% drink sweet beverages at least once per day and 30.9% consumed salty snacks (nuts, crackers, chips).¹⁹ Moreover, according to Hoteit et al., 9 in every 16 households ate less than one or two meals daily and the majority of them were omitting their meals to spare food. In addition, more than 50% of the Lebanese population were having a low food consumption score.¹⁵ Another study conducted in three countries of which Lebanon was included, showed that 33% of people were shopping for their food once per week and 31% of the respondents stated that they were shopping two to three times a week during the COVID-19 pandemic. One in every four were purchasing food less than once a week. Less than 10% of people living in Lebanon, Tunisia, and Jordan were obtaining food and groceries from supermarkets or food delivery services and was significantly the least reported among the Tunisians.²⁰

Limitations

Some limitations should be considered when evaluating the results of this study. This is a cross-sectional study of which it included retrospective data on the time before lockdown. Thus, due to the respondents’ memory, the presented eating habits could be biased due to the type of questions asked concerning the food patterns before the pandemic. Another limitation concerns the determination of the portion size. Thus, it was not possible to use home measurements or the photographs of products to assist respondents in determining the food portion size.

Underlying data

OSF: CORONACOOKING COVID-19 Lebanon. https://doi.org/10.17605/OSF.IO/PRTDQ.⁴³

The project contains the following underlying data:

Extended data

OSF: Supplementary materials CoronaCookingSurvey April 2020 https://doi.org/10.17605/OSF.IO/8GPZX.⁴⁴

The project contains the following extended data:

Data are available under the terms of the Creative Commons Zero “No rights reserved” data waiver (CC0 1.0 Public domain dedication).