Morocco is undergoing a rapid nutrition transition, characterised by increasing obesity and a high prevalence of diet-related chronic diseases, such as type 2 diabetes, CVD and cancer( Reference Benjelloun 1 – Reference El Rhazi, Nejjari and Zidouh 4 ). According to one study, which investigated the burden of diseases attributable to low intake of fruit and vegetables and its association with different health outcomes, it was estimated that worldwide over 1·7 million deaths (2·8 %) were attributable to low fruit and vegetable intake, placing it among the top ten risks factors for mortality in middle- and high-income countries (the seventh and the eighth cause, respectively)( 5 ). More recently, the Global Burden of Disease study 2010 reported that in North African and Middle East countries, low consumption of fruit was the sixth overall leading cause of morbidity and the fourth behavioural factor contributing to morbidity( Reference Lim, Vos and Flaxman 6 ). The low consumption of vegetables was the sixteenth overall leading cause of morbidity. Therefore, it is crucial to focus on fruit and vegetable intake, as they have a potential preventive effect on weight gain and chronic disease development( Reference Mendis and Alwan 7 ).
Based on evidence that fruit and vegetables have a protective effect on diet-related chronic diseases, the WHO has recommended eating at least 400 g of fruit and vegetables daily( 8 ). But the health benefits of fruit and vegetables are not only a question of quantity, but also a question of diversity( Reference Thompson, Heimendinger and Diker 9 , Reference Bhupathiraju and Tucker 10 ). Indeed, at present, no studies have clearly identified the mechanisms for the benefits of eating fruit and vegetables, nor which fruit or vegetables are most effective( Reference Padayatty and Levine 11 ). Therefore, it is recommended to eat a wide variety of fruit and vegetables of different colours including red, green, yellow, white, purple and orange( 12 ).
To be able to promote consumption of fruit and vegetables, it is important to investigate what kinds of factors may influence their intake, in order to identify potential levers to increase their intake in populations.
Food choices are not only influenced by hunger or other physiological factors. On the contrary, they are influenced by a wide range of determinants acting at different levels. Systematic reviews have identified several factors influencing fruit and vegetable intake, such as: biological determinants (e.g. gender and age); economic determinants (e.g. income and cost); physical determinants (e.g. time, cooking skills, accessibility and living area); and social determinants (e.g. marital status, having children, education, family, peers and culture)( Reference Pollard, Kirk and Cade 13 – Reference Guillaumie, Godin and Vezina-Im 16 ).
Most of the studies investigating the relationship between fruit and vegetable consumption and age concluded that the amount of fruit and vegetables consumed increases with age( Reference Johansson and Andersen 17 – Reference Azagba and Sharaf 22 ). The influence of income/education on fruit and vegetable consumption has also been widely described in the literature and studies tend to conclude that people with higher income, socio-economic status or education are more likely to consume more fruit and vegetables( Reference Ricciuto, Tarasuk and Yatchew 23 – Reference Paalanen, Prattala and Palosuo 29 ). The household composition has also been demonstrated to influence the level of fruit and vegetable consumption, i.e. married couples are more likely to consume more fruit and vegetables( Reference Pollard, Kirk and Cade 13 , Reference Kamphuis, Giskes and de Bruijn 14 , Reference Friel, Newell and Kelleher 30 , Reference Kamphuis, van Lenthe and Giskes 31 ), but this relationship is less evident once they have children( Reference Kamphuis, Giskes and de Bruijn 14 , Reference Wandel 32 , Reference Devine, Wolfe and Frongillo 33 ).
Fruit and vegetable consumption may also be related to certain behaviours such as out-of-home eating or skipping meals. Indeed, several studies conducted in adults living in high-income countries reported inverse associations between out-of-home eating frequency and fruit and vegetable intake( Reference Bowman and Vinyard 34 – Reference Vandevijvere, Lachat and Kolsteren 37 ).
The aims of the present study were to characterise fruit and vegetable consumption in terms of quantity, quality and variety (all estimates of diversity), and to investigate determinants of fruit and vegetable consumption such as socio-economic and behavioural determinants, in urban Moroccan women living in a context of high prevalence of overweight and obesity.
Experimental methods
The study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects were approved by the Ethical and Deontological Consultative Committee of the Institute of Research for Development (July 2009) and by the Moroccan Ministry of Health (letter n°623, March 2009). After being thoroughly informed of the purpose and procedures of the survey, written or verbal informed consent was obtained from all participants. Verbal consent was witnessed and formally recorded. Women did not receive any incentive for their participation.
The questionnaires were administered in Arabic by trained interviewers.
Participants and settings
The target population was non-pregnant women of child-bearing age (20–49 years) living in the urban area of Rabat-Salé. Within this area, a three-stage random cluster sample was used. Forty-five clusters of about fifty households were randomly selected among census enumeration areas by the Ministry of Statistics and Planning. In each cluster, addresses were numbered. Then in each cluster a starting point, based on the address list, was randomly selected. From this starting point, investigators proceeded to adjacent households until twenty eligible households, i.e. with at least one non-pregnant woman aged 20–49 years, were selected. If several women were eligible in a household, one woman was randomly selected to participate.
Measures
Fruit and vegetable consumption
A single 24 h recall was administered to each participant by trained interviewers using the multi-pass method( Reference Dennis, Ernst and Hjortland 38 ) (see online supplementary material). In this procedure, participants initially reported the time of eating, as well as all foods and beverages consumed during the last 24 h. At that stage, information about quantity and other details was not reported. After reviewing the list of food items for completeness, a second interview was undertaken with the participant to record details of recipes cooked and quantities of all food items consumed on the recall day (using diverse methods: photographs of food portion size( Reference Hercberg, Deheeger and Preziosi 39 , 40 ), household measures, known weight or prices). Lastly, specific information for fruit and vegetables was extracted from the recall. A Fruit and Vegetable Diversity Score (FVDS) was calculated by summing the number of different fruit and vegetables consumed over the last 24 h without any minimum quantity.
Socio-economic factors
Data on age, marital status, parity, educational level, employment and living area were collected.
An economic index was computed from six variables concerning housing (number of persons per room, presence of toilets, source of drinking water, kitchen and bathroom at home) and eleven variables concerning equipment at home (fridge, washing machine, dishwasher, satellite dish, Internet access, television, heating, air conditioning, telephone, car, computer) using correspondence analysis( Reference Traissac and Martin-Prevel 41 ). Then households were classified into tertiles, corresponding to low, medium and high economic level.
Behavioural factors
The behavioural factors investigated were: (i) the number of out-of-home eating occasions during the last week (including eating at other people’s houses); (ii) usual consumption of the three possible main daily meals (breakfast, lunch and dinner); (iii) eating from a shared bowl; and (iv) the number of processed foods consumed during the previous day. The processed foods taken into account were biscuits, cooked meats, cream cheeses, yoghurts and soft drinks.
Anthropometry
Weight was measured using BodyUp digital scales accurate to 100 g (Tefal™, France), which were verified daily. Height was measured using a portable stadiometer (Seca® 214) to the nearest millimetre (Seca®, Germany). All anthropometric measurements were performed by the interviewers. BMI was assessed from measured weight and height, and women were classified into four groups based on the WHO classification( 42 ).
Data entry and data management
A data entry file was set up with EpiData entry, version 3·1( Reference Lauritsen and Bruus 43 ). Data from questionnaires were entered twice, into two separate files, by the same operator and then compared for errors. Errors were corrected and the comparison between the two files was performed until differences no longer existed between them.
Data analysis
All statistical analyses were performed with the Stata statistical software package version SE 11·2( 44 ). Summary statistics were used to describe the characteristics of the sample. Associations between socio-economic characteristics, behavioural factors and fruit and vegetable consumption were tested using linear regression analysis.
All analyses took into account both sampling design and sampling weight and were adjusted for all of the socio-economic variables and energy. P<0·05 was considered as significant.
Results
Of the women who were randomly selected to participate in the study, fifty-six refused (5·9 % refusal rate). Eight hundred and ninety-five women were interviewed within forty-five clusters. One woman was excluded from the overall analysis because no food consumption data had been recorded for her and thirty-nine women considered as outliers (energy intake outside the allowable range of 2092–14 644 kJ/d (500–3500 kcal/d)) were excluded from the analyses that investigated the relationship between socio-economic and behavioural determinants and fruit and vegetable consumption.
Population characteristics
Over two-thirds of the respondents were married (66·1 %; 95 % CI 61·6, 70·7 %) and had at least one child (70·0 %; 95 % CI 59·8, 80·1 %; Table 1). About two-fifths of the women had never attended school (41·3 %; 95 % CI 34·6, 48·1 %) and the majority of women were unemployed (80·1 %; 95 % CI 76·0, 84·3 %). Slightly less than two-thirds of the sample (65·0 %; 95 % CI 50·2, 79·0 %) lived in the medina (traditional living area in Morocco; Table 1). Two-thirds of the respondents were either overweight or obese (overweight: 33·7 %; 95 % CI 30·3, 37·1 %, obese: 32·4 %; 95 % CI 28·5, 36·4 %).
Characteristics of the study participants: Moroccan women of child-bearing age (20–49 years) from the urban area of Rabat-Salé (n 894)
* Traditional Moroccan living area.
† Precarious living area and shanty town.
‡ Data based on 24 h recalls (n 855); processed foods are biscuits, meat products, processed cheese, yoghurts and soft drinks.
One-third of women ate out of home at least once during the previous week (33·7 %; 95 % CI 26·5, 40·8 %), another third usually skipped one of the three main daily meals (30·2 %; 95 % CI 24·0, 36·5 %) and most of them usually ate from a common dish (86·6 %; 95 % CI 82·4, 90·7 %), i.e. from a shared bowl. Almost two-thirds of the women did not eat any processed foods during the last 24 h (59·1 %; 95 % CI 54·0, 64·9 %; Table 1).
Fruit and vegetable intake
During the previous 24 h, nearly two-thirds of the respondents ate fresh fruit (60·3 %; 95 % CI 54·6, 66·0 %), but much fewer (13·0 %; 95 % CI 10·2, 15·8 %) ate dried fruit; nearly all women ate vegetables (94·2 %; 95 % CI 92·8, 95·6 %) and slightly more than a quarter ate beans or pulses (28·0 %; 95 % CI 23·7, 32·3 %). Only a small proportion of women (2·6 %; 95 % CI 1·5, 3·7 %) did not eat fruit or vegetables during the previous day.
The mean fruit and vegetable intake was 331 (95 % CI 306, 357) g/d. The mean fruit intake was 155 (95 % CI 136, 175) g/d and the mean vegetable intake was 176 (95 % CI 164, 189) g/d. Slightly less than one-third of the women (32·1 %; 95 % CI 28·0, 36·2 %) met the WHO recommendations, i.e. ate ≥400 g of fruit and vegetables daily. Half of the sample (50·6 %; 95 % CI 45·3, 55·9 %) were low consumers, i.e. consumed <280 g/d.
The weight of a mean fruit portion size was 155 (95 % CI 144, 166) g and the weight of a mean dried fruit portion size was 31 (95 % CI 24, 39) g. The weight of a mean vegetable portion size was 39 (95 % CI, 37, 41) g and the mean weight of a portion size of beans or pulses was 126 (95 % CI 113, 139) g.
Regarding diversity in the previous day, women ate slightly less than five different fruits and vegetables (4·77; 95 % CI 4·52, 5·04). They ate roughly one fruit (1·08; 95 % CI 0·97, 1·18) and more than three different kinds of vegetables (3·70; 95 % CI 3·49, 3·91).
Socio-economic and behavioural determinants of fruit and vegetable intake
The mean daily intake of fruit and vegetables combined was not associated with any of the socio-economic determinants investigated, except for economic status (Table 2). Indeed, women with a higher economic status ate significantly more fruit and vegetables (296 g/d for women belonging to the low economic group v. 390 g/d for women belonging to the high economic group, P=0·049). The same finding was reported for fruit. Thus, women with a higher economic status ate a significantly larger amount of fruit (117 g/d for women belonging to the low economic group v. 203 g/d for women belonging to the high economic group, P=0·031). The mean daily vegetable intake was not associated with any of the socio-economic determinants investigated.
Relationship between socio-economic factors and daily fruit and vegetable intake controlling for age, marital status, number of children, education, employment, economic status, living area and energy intake among Moroccan women of child-bearing age (20–49 years) from the urban area of Rabat-Salé (n 855)
The only socio-economic factor associated with FVDS was age (data not shown). Indeed, older women ate significantly more different types of fruit and vegetables compared with younger women (4·9 v. 4·7, P=0·032). Fruit Diversity Score (FDS) increased with age and economic status, i.e. older and wealthier women ate significantly more types of fruit (1·1 v. 1·0, P=0·038 and 1·35 v. 0·90, P=0·022; respectively). Vegetable Diversity Score (VDS) was not associated with any of the socio-economic factors investigated.
Mean daily fruit and vegetable intake was not associated with out-of-home eating frequency or with the fact that women ate from a common dish (Table 3). Women who did not consume any processed food during the previous day tended to eat slightly more fruit and vegetables compared with those who ate two or more processed foods (334 g/d v. 323 g/d). However, this association was not significant. Women who usually skipped at least one of the three main daily meals ate significantly lower amounts of fruit and vegetables (305 g/d for women who skipped at least one meal v. 343 g/d for women did not skip any meal). Mean daily fruit intake was not associated with any of the behavioural determinants investigated. Mean daily vegetable intake was not associated with whether women usually ate from a shared bowl; however, women who ate out of home more frequently, as well as women who usually skipped at least one of the three main daily meals and women who ate more processed food ate significantly less vegetables (148 g/d v. 180 g/d, 152 g/d v. 187 g/d and 152 g/d v. 182 g/d, respectively).
Relationship between behavioural factors and daily fruit and vegetable intake controlling for age, marital status, number of children, education, employment, economic status, living area and energy intake among Moroccan women of child-bearing age (20–49 years) from the urban area of Rabat-Salé (n 855)
* n 854 (one missing value).
FVDS, as well as FDS and VDS, were not associated with any of the behaviours investigated.
Discussion
In terms of quantity and quality, fruit consumption was vastly different from vegetable consumption. Indeed, while almost all women ate vegetables on a daily basis, less than two-thirds of them ate fruit every day. Overall women ate only one portion of fruit per day with a mean portion size weighing 155 g, i.e. about twice the weight of the reference portion size (80 g). In comparison, vegetable diversity was higher but the weight of the mean vegetable portion size was 39 g, i.e. about half the weight of the reference portion size. Therefore, contrary to other studies where one eating occasion was assimilated to be one portion( Reference Yarnell, Fehily and Milbank 45 – Reference Thompson, Demark-Wahnefried and Taylor 47 ), in the present context, it is important to measure the exact quantity of fruit and vegetables consumed rather than just recording or recalling the frequency of consumption, as one occasion was completely different from one portion.
The mean fruit and vegetable intake was relatively high (331 g/d); however, only one-third met the ≥400 g/d recommended by the WHO and half of the sample were considered as low consumers, indicating that in the present context the consumption of fruit and vegetables is still a public health issue. In comparison, in Brazil, a country also experiencing rapid nutrition transition and with higher economic development compared with Morocco, one in five adults meet the WHO daily recommendations (20·5 % of women)( Reference Ministério da 48 ). In a high-income country, such as the USA, less than one-third of adults eat the recommended amount of fruit and vegetables (26·3 % eat ≥3 servings of vegetables and 32·5 % eat ≥2 servings of fruit per day)( 20 ). It is worth noting that all these results are based on different dietary assessment methods and therefore are not comparable, strictly speaking.
Contrary to what was usually reported in other studies( Reference Pollard, Kirk and Cade 13 , Reference Kamphuis, Giskes and de Bruijn 14 , Reference Estaquio, Druesne-Pecollo and Latino-Martel 19 , 20 , Reference Azagba and Sharaf 22 ), in the present context, the amount of fruit and vegetables eaten was not associated with age, marital status, parity, education or living area. For age, one of the possible reasons for not finding any association was that the population under investigation – women of child-bearing age – represented a relatively narrow age range. For marital status, the difference with other studies may be cultural, given that in Morocco, single people tend to stay living within families until they are married.
As reported in other studies( Reference Estaquio, Druesne-Pecollo and Latino-Martel 19 , Reference Ricciuto, Tarasuk and Yatchew 23 , Reference Elfhag, Tholin and Rasmussen 25 , Reference Hall, Moore and Harper 27 , Reference Lallukka, Pitkaniemi and Rahkonen 28 , Reference Ball and Mishra 49 ), women with a higher economic status ate more fruit and vegetables.
Similar to the amount of fruit and vegetables consumed, the determinants of fruit or vegetable consumption are completely different. Indeed, fruit consumption was positively associated only with economic status, whereas vegetable consumption was not influenced by any of the socio-economic factors investigated. Moreover, while fruit consumption was not associated with any behavioural determinants, vegetable consumption was independently and negatively related to out-of-home eating, skipping meals and eating processed foods. In other words, in the present context where fruit is more expensive than vegetables, fruit consumption was economically driven, whereas vegetables are part of the daily main traditional Moroccan dish called tajine that is made up of meat (more rarely of fish) and vegetables, and vegetable consumption was thus influenced more by behaviours characteristic of a modern dietary pattern.
There are some limitations associated with the present study. First, fruit and vegetable intake was based on data collected from a single 24 h recall. Therefore, the interpretation of results should be treated with caution, since a single 24 h recall gives no information on intra-individual variability in food intakes and so it is less likely to reflect true long-term individual intakes( Reference Willett 50 ). Furthermore, the 24 h recall relies on the participant’s memory, both for identifying food and beverages consumed and the evaluation of portion sizes. Additionally, in the present study, the amount of food consumed was assessed mainly using photographs of food portion size presented in an individual plate, whereas Moroccan women traditionally eat from a shared dish. As a consequence, this may have introduced a bias in the reported amount of vegetables consumed.
Conclusions
In a context where diet-related chronic diseases are highly prevalent and half of urban Moroccan women are considered low consumers of fruit and vegetables, it is crucial to focus on their intake as part of a healthy diet. Fruit consumption and vegetable consumption among urban Moroccan women are driven by distinct determinants. Indeed, while vegetable consumption was related only to behavioural determinants, fruit consumption was influenced only by economic status. Therefore, programmes that promote fruit and vegetables intake should account for these differences to enhance their effectiveness. For instance, programmes that would aim to increase vegetable consumption might promote the Moroccan traditional diet that is naturally high in vegetables; while programmes that would aim to increase fruit consumption might include economic measures such as subsidies or vouchers to make fruit more affordable for the population.
Acknowledgements
Acknowledgements: The authors would like to acknowledge all of the interviewers, as well as all of the interviewees. Financial support: This work was part of the Obe-Maghreb project ‘Understanding the nutritional transition in the Maghreb to contribute to the prevention of obesity and non-communicable diseases’ that was funded by the CORUS programme of the French Ministry of Foreign Affairs (contract 6028-2) and by the Institute of Research for Development (IRD). The funders had no role in the design, analysis or writing of this article. Conflict of interest: None. Authorship: E.L., A.B., A.G. and F.D. were involved in data collection. E.L. conducted statistical analyses and drafted the manuscript. A.G., F.M., F.D. and M.H. were involved in the critical revision of the manuscript. All authors were involved in the planning of the study and read and approved the final manuscript. Ethics of human subject participation: The study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects were approved by the Ethical and Deontological Consultative Committee of the Institute of Research for Development (July 2009) and by the Moroccan Ministry of Health (letter n°623, March 2009). Written or verbal informed consent was obtained from all participants and verbal consent was witnessed and formally recorded.
Supplementary material
To view supplementary material for this article, please visit http://dx.doi.org/10.1017/S1368980014001761
