Origin of Mediterranean Diets

Comparison between Sardinia and Malta: the Mediterranean diet revisited

Running title: Mediterranean diet revisited[i]

Sophie TESSIER1* and Mariette GERBER1

1Groupe d’Epidémiologie Métabolique

Centre de Recherche en Cancérologie

INSERM-CRLC

34 298 Montpellier Cedex 5

FRANCE

* Corresponding author: Tel: +33 (0)4 67 61 37 58 Fax: +33 (0)4 67 61 37 87

email:

Abstract

The Mediterranean diet is often characterized in terms of food items only. In this paper, the comparison of meals, in-between meals snack consumption and total daily food intake is presented between Sardinia and Malta in terms of structure, social environment and hours, together with their changes. Data were collected in 2001 in Sardinia and 2002 in Malta. A structured qualitative questionnaire, articulated around four main themes: food supply, transformation, preparation and consumption habits, was administered by face-to-face interviews with the help of a local person. Inquiries were carried out over two generations with 30 «mother-daughter» couples in each island.

We highlighted some Mediterranean characteristics in Sardinia by showing striking contrasts between Sardinian and Maltese food habits such as meal preparation times, both breakfast and main meal structures, total daily food intake profiles, mealtimes and commensality in the mothers’ generation. Some of these characteristics were also maintained in the daughters’ generation.

This investigation has presented evidence that beyond foods, meal pattern, structure and commensal rituals appeared as fundamental aspects of Sardo-Mediterranean food habits, which are important to consider when trying to maintain or implement the Mediterranean dietary model.

Keywords:

2

Mediterranean diet

Meal preparation times

Meals structures

Total daily food intake profiles

Meals patterns

Commensality

2

Introduction

Even though regional and temporal disparities do exist, some nutritional compounds common to all Mediterranean countries emerge to create the Mediterranean dietary model which is based on the comparison between Northern European and Southern European countries (Padilla, 1996; Corpet et Gerber, 1997). Generally speaking and in a unified way, the traditional Mediterranean diet is renowned for its frugality, and is characterized by a high consumption of cereals, pulses and varied fresh and dried fruit and vegetables. Meat consumption is low, especially those from bovines; the meat consumed is rather from sheep and goat given the important presence of this livestock. On the other hand, fish is often consumed, and more particularly along the coastline. Inland, smoked or salted fish is mostly used because of the limited availability of fresh sea products. There is virtually no milk, nor butter but some cheeses and yoghurts, which are rather from sheep and goat. Olive oil is the only added lipid, and wine is regularly consumed with main meals in moderated amount for men, whereas women hardly drink any. It is also important to note the role of allium and wild herbs, providing this cuisine with flavours and precious nutrients; a cuisine that is usually eaten at the table with one’s family (Gerber et Corpet, 1997; Vague, 1999; Gerber 2004).

The eating act involves a set of rituals defining what is qualified by Poulain (2002-1) as the “space of food consumption habits”. The number of food intakes, their structure, the time and the social context in which they occur are key elements that contribute to define the profile of the total daily food intake. Aymard et al. (1993) report that daily chronology of meals, time at which they are taken and the time devoted to their preparation are indeed fundamental aspects to take into consideration as they are directly linked to the importance which is attached to them. Besides, these aspects also contribute to determine a meal’s composition.

Since the Mediterranean diet is generally recognised as beneficial for health, we implemented a study to identify and understand actual characteristics of the Mediterranean lifestyle in term of food habits. In addition, because it tends to disappear in several Mediterranean regions (Scali et al., 2000; Ferro-Luzzi et al., 2002), we studied factors of change. This study was conducted in two societies of seemingly comparable geographical contexts (Mediterranean islands) but historically different, Sardinia and Malta (Tessier, 2004). Over the second half of the last century, the evolution of Sardinian and Maltese societies has been different and has led to diverse eating habits. Sardinia is known to have preserved a traditional Mediterranean diet (Carbini, 1998) whereas Malta has evolved towards a Northern European type food consumption (Serra-Majem et al., 1997).

This comparative study, in which anthropological methods were used to evaluate food habits in a holistic way, enabled us to identify key elements of the Mediterranean dietary model which are little or never encountered in the literature and that nutritional epidemiology methods, usually applied to the study of the Mediterranean diet, can not reveal. The comparison of meals, in-between meals snack consumption and total daily food intakes are presented in this paper between Sardinia and Malta in terms of structure, social environment and times, as well as their changes.

Methods

Subjects

In order to study the changes in eating habits, inquiries were carried out over two generations (mothers-daughters), since women still have the monopoly of household tasks and play a key role in the transmission of food habits.

In both Sardinia and Malta, it was planned to interview 30 «mother-daughter» couples, that is to say 60 people in total. Recruitment of women was done either in collaboration with general practitioners, or through parish play groups where mothers came with their young children, or through a network. The latter method presented the advantage of adapting itself to the complexity of various local social groups. It thus enabled us to include women who would not have been included in this study otherwise, and therefore created new interview opportunities. The direct approach of this recruitment method also allowed us to reduce the number of refusals, and therefore participation bias. Fieldwork was carried out over three months in Sardinia (April, July and September 2001) and one month in Malta (May 2002). Fieldwork was realised in a shorter time in Malta due to, on the one hand, previously acquired experience in Sardinia and, on the other hand, the smaller size of the archipelagos, compared to Sardinia, making contacts and travelling easier.

Questionnaire

In both Sardinia and Malta, a structured qualitative questionnaire which had previously been elaborated for a European pilot study COMER N°97/CAN/45916 (Gerber et Padilla, 1998) was used. It was particularly adapted to comparable anthropo-nutritional data collection in different countries. The questionnaire was articulated around four main themes: food supply, transformation, preparation and consumption and encompassed mainly open-ended questions, which allowed in-depth investigation of food habits. It was administered by face-to-face interviews with the help of a local person who translated Italian (or Sardinian) exchanges into French or English, and Maltese exchanges into English. Local translators had previously been thoroughly briefed on the study.

Data processing and analysis

For each meal, the analysis of the structure was based on the typology used in the DGAL/CIDIL/OCHA study (Poulain; 2002-2).

The social science adapted software, Le SPHINX Lexica 4.0, allowed to semi-quantify qualitative data by performing an inventory of the range of variable modalities answered for each question. These steps were carried out ensuring that the qualitative aspect of data was not lost in the computerization process. Chi2 tests were performed for spatial comparison of the qualitative variables between the two islands using SPSS 10.1. Generational differences were analysed performing Mc Nemar and Stuart-Maxwell tests, which allow comparison of matched data, using STATA 7.0. These tests allowed us to study changes in food habits at two different levels. Either at the inter-individual level, i.e. from one generation to the next (from mother to daughter) or/and at the intra-individual level, i.e. comparisons in time were made, for each woman, between food consumption habits during childhood (i.e. in the parental home) and current one (i.e. in the present family home).

Results

Socio-economic profile of Sardinian and Maltese samples

In total, 63 women were interviewed in Sardinia and 60 in Malta. There was no significant difference in terms of socio-economic characteristics between women of the two islands. The sample was composed of mothers (mean age 66) and daughters (mean age 39); most of them lived with spouse and children (74%) rather than alone. A majority of women (64%) did not work outside of home.

Meal preparation times

Table 1 here

When looking at meal preparation times, a distinction was made between lunch and dinner as well as between weekdays and weekend.

During the week, Sardinian women were more likely to spend more time on lunch preparation than Maltese women (60% vs. 35% respectively spent between half an hour to an hour; p<0.001). No difference was observed between weekend habits and week ones for any of the Sardinian women whereas there were differences for 32% of Maltese women (p<0.001). Maltese women were more likely to report that a proper meal was being cooked on weekends rather than during the week.

If lunch was judged as being the most important meal of the day by 75% of Sardinian women, it was, on the contrary, dinner for a majority of Maltese women (59%) (p<0.001).

Significant generational differences were observed in both islands (Table 1). Regarding dinner preparation during the week, both Sardinian and Maltese mothers were more likely to spend less than half an hour on its preparation (65% and 47% respectively). A significantly higher proportion of Maltese mothers (50%) than daughters (20%) spent between half an hour to an hour on lunch preparation (p=0.004) and less than half an hour on dinner (47% vs. 7% respectively) (p=0.001). Moreover, 43% of Maltese mothers spent as much time on meal preparation during the week as on weekends whereas it was only the case for 17% of Maltese daughters (p=0.002). Sardinian daughters only differentiated themselves from their mothers as far as dinner preparation was concerned and were more likely to spend between half an hour to an hour (45% vs. 23% respectively; p=0.027). Though the Sardinian second generation was behaving significantly different from the first Sardinian generation as far as dinner preparation time was concerned, it remained significantly different from its Maltese homologue who was 77% in that time range (p=0.034).

Meal structures

Breakfast

Three main breakfast structures were identified. The continental type which was composed of a hot drink (coffee, tea or milk), toasts, biscuits or pastry. A citrus fruit could also be present, either in the form of juice or fruit. The Anglo-Saxon type which was based on the breakfast of British and Northern European countries. To the continental type of breakfast was added breakfast cereals and/or salted foods such as delicatessen, cheese or cooked breakfast (i.e. bacon, eggs and baked beans). And lastly, the light type which corresponded to a simplified breakfast. A drink (hot or cold), which could be composed of milk or accompanied by a dairy product (e.g. yoghurt).

A significant difference was observed in the breakfast structure between both islands. The continental type was the predominant structure type in Sardinia whereas in Malta, the Anglo-Saxon type prevailed (p<0.001). A light type of breakfast was characteristic for the same amount of Sardinian women as for Maltese women. In both islands, a minority of women reported having no breakfast (5% and 3% respectively for Sardinia and Malta). No significant difference was observed between generations in both islands.

Lunch and dinner

Taking a three-course meal as a reference, two main structure types were distinguished: complete meals and simplified meals. From these two main structures, different combinations were described according to the various set of situations encountered in both samples.

Despite the current preponderance of simplified meals (lunch and dinner) in Sardinia, we could observed more complete meals there than in Malta. In Malta, simplified meals were part and parcel of food habits for 88% of interviewed women (vs. 63% in Sardinia) at lunch time (p<0.001) and 97% (vs. 84% in Sardinia) at dinner time (p=0.011). No significant change was observed for women of any nationality in the structure of main meals. Though small changes occurred, equal proportions of women declared having shifted from a simplified to a complete meal and vice-versa. It thus statistically annihilated the effect of change on both samples. No significant difference was observed between generations in both islands.

A different behaviour was observed between both islands in the way main meals were simplified. In Sardinia, the most frequent way to simplify a meal meant having a two-, instead of a three-, course meal (for 38% of interviewed women) whereas in Malta, this meant having a sandwich, on the Anglo-Saxon dietary model for lunch (for 38% of interviewed women).

Snack consumption in-between mealtimes

For both morning and afternoon, past and present, a higher proportion of Maltese than Sardinian women reported eating snacks in-between meals. Except for the morning snack in the past, a significant difference was observed between both nationalities (65% vs. 46% respectively for the afternoon snack in the past, p=0.034; 73% vs. 35% respectively for the morning snack nowadays, p<0.001; 90% vs. 44% respectively for the afternoon snack nowadays, p<0.001).

From the childhood to the current period of time, changes in snack consumption were opposed between Sardinia and Malta. On the one hand, during childhood, 41% of Sardinian women ate a morning snack; they were only 35% to do so currently. On the other hand, Maltese women who were already 55% to consume a morning snack in the past, shifted to 73% at present. No change was observed regarding the afternoon snack consumption in the Sardinian sample. Maltese women sample shifted from 65% to 90% (p=0.007).

No significant difference was observed between generations within both islands. However, significant differences were observed between nationalities for each generation. Maltese mothers were currently more likely than Sardinian mothers to have both morning and afternoon snacks (77% vs. 35% respectively for the morning snack, p=0.001; 87% vs. 39% respectively for the afternoon snack, p<0.001). Likewise, this difference was also found at the daughters’ generation (70% vs. 34% respectively for the morning snack, p=0.005; 90% vs. 50% respectively for the afternoon snack, p=0.001).

It was reported by some Maltese women that delicatessen could be consumed as an evening snack after dinner due to the early timing of the evening meal as we will see further on.