Portuguese households' diet quality (adherence to ... - Nature

European Journal of Clinical Nutrition (2008) 62, 1263–1272

& 2008 Macmillan Publishers Limited All rights reserved 0954-3007/08 $32.00 www.nature.com/ejcn

ORIGINAL ARTICLE

Portuguese households’ diet quality (adherence to Mediterranean food pattern and compliance with WHO population dietary goals): trends, regional disparities and socioeconomic determinants SSP Rodrigues1, M Caraher2, A Trichopoulou3 and MDV de Almeida1 1 3

Faculty of Nutrition and Food Sciences, Porto University, Porto, Portugal; 2Centre for Food Policy, City University, London, UK and Department of Hygiene and Epidemiology, School of Medicine, University of Athens, Greece

Objective: To evaluate households’ diet-quality trends and regional disparities, as well as to identify the influence of its socioeconomic determinants. Methods/Subjects: Two dietary indexes were applied: a revised Healthy Diet Indicator version (HDIr) and the Mediterranean Adequacy Index (MAI). The socioeconomic differences between households with low-diet quality (if their scoring in HDIr or MAI was below the national median) and the remaining ones were analysed by means of logistic regression. Representative samples of households among each region were analysed. Results: Adherence to a Mediterranean diet was poor in all Portuguese regions; the proportion of households adhering to it has diminished in the North, Centre, Lisbon and Tejo valley. The proportion of households with high HDIr decreased in the North, Centre, Lisbon and Tejo valley and Algarve. In multivariate analysis, households whose head had secondary or higher education, living in semi-urban or urban areas, from the Azores region, within the higher income or outside home food expenditures groups were more likely to be in low-quality diet. Conclusions: The decreasing adherence to a Mediterranean food pattern and the diminishing compliance with World Health Organization dietary goals show that Portuguese households have reduced their diet quality. The observed influence of urbanization and region on diet quality highlights the importance of considering regional nutrition strategies.

European Journal of Clinical Nutrition (2008) 62, 1263–1272; doi:10.1038/sj.ejcn.1602852; published online 1 August 2007 Keywords: diet quality; household budget survey; Mediterranean diet; WHO goals; socioeconomic determinants; Portugal

Introduction Owing to the multidimensional nature of dietary habits, prevention and treatment of diet-related diseases have evolved to incorporate a more holistic approach than the traditional focus on single dietary components. To evaluate diet quality, researchers have developed a number of composite indexes in which dietary behaviour can be analysed as a single factor evaluating at the same time many relevant items (Trichopoulou et al., 1995; Huijbregts et al., 1997; Alberti-

Correspondence: Dr SSP Rodrigues, Faculty of Nutrition and Food Sciences, Porto University, Rua Dr Roberto Frias, Porto, Portugal. E-mail: [email protected] Received 23 November 2006; revised 18 June 2007; accepted 20 June 2006; published online 1 August 2007

Fidanza et al., 1999; Lagiou et al., 1999). Traditionally developed for the study of the relationships between diet and disease, such indexes may also help on the interpretation of dietary trends and disparities (Kant, 1996, 2004). Different perspectives may contribute to the definition and multi-method evaluations of diet quality. The attributes selected for its measurement are based on current scientific knowledge, the population to be assessed, the resources available for dietary analysis and the researcher viewpoint (Kant, 1996, 2004). In the present study, evaluating diet quality by means of conformity with scientific recommendations incorporating both a food-based and a nutrient approach was thought to be relevant. Despite of the recognition that behind the Mediterranean diet concept there is a large heterogeneity; adherence to this food pattern was considered because its major characteristics

Portuguese households’ diet quality SSP Rodrigues et al

1264 have been identified and accepted as a dietary pattern to be preserved and encouraged for its health benefits (Willett et al., 1995; Trichopoulou, 2001; Serra-Majem et al., 2006; Willett, 2006). Furthermore, although its Atlantic geographical situation, Portugal is Mediterranean by history and nature; this has been corroborated not only for its cultural and social practices (Ribeiro, 1998) but also for its food habits (Nestle, 1995; Willett et al., 1995; Serra-Majem et al., 1997; Assman et al., 2000). The increases in the incidence of non-communicable diseases (NCDs) are strong motivations for the public health promotion of healthy dietary habits (Robertson et al., 2004; WHA, 2004). International and European bodies have developed population dietary goals consistent with its prevention (CEC, 1993; EuroDiet, 2001; WHO/FAO, 2003). It was thus considered to evaluate existing dietary patterns against the standards in the most recent one (WHO/FAO, 2003). The importance of socioeconomic disparities on the diet has long been reported. Identifying attributes of the population groups with unfavourable dietary behaviour may prove to be of maximum relevance for the delivery of effective prevention messages and strategies to improve healthy lifestyles (Mennel et al., 1992; Fieldhouse, 1995; Roos and Prattala, 1999; CEC, 2005). Food habits learned and habituated throughout the lifespan occur within the socialization process, where several influences act concurrently (Fieldhouse, 1995). Among them, family plays an important role as not only the first but also the most continuous influence, which makes decisions around food at the household level to be considered when efforts to carry dietary interventions are planned. Although several studies have attempted to identify the influence of socioeconomic factors on individual’s dietary intake (Moreira and Padraõ, 2004; Marques-Vidal et al., 2006), few have focused on socioeconomic determinants of household dietary patterns. The majority have focused on particular foods or food groups (Rodrigues and de Almeida, 2001; Trichopoulou et al., 2002; Ricciuto et al., 2006) but few on overall diet quality (Fuentes-Bol, 2002; Naska et al., 2006). This study aimed to evaluate trends and regional disparities in Portuguese households’ diet quality, namely, adherence to the Mediterranean food pattern and to World Health Organization (WHO) population dietary goals. It was also intended to identify the influence of socioeconomic determinants on diet quality.

1989/1990, 10 554 from 1994/1995 and 10 020 from 2000/01 (INE, 1990, 1997, 2002). As defined in the official territorial division (NUT II, Decreto-Lei n 46/89, 1989), the seven main Portuguese regions were considered (in brackets after each region name, the proportion of households to the overall population, which is the same for the three survey years, is presented): mainland Portugal—North (33%), Centre (18%), Lisbon and Tejo valley (35%), Alentejo (6%), Algarve (4%)— and the autonomous regions of Azores (2%) and Madeira (2%) (Figure 1). Data were managed and analysed according to Data Food Networking (DAFNE) methodology (Lagiou and Trichopoulou, 2001): the availability of foods and beverages was estimated as the sum of purchases, contributions from the household’s own production and food items received as gifts, without accounting for food consumed outside the house, food wasted, spoiled or fed to pets. The per capita availability was computed assuming intra-household food distribution to be equal among all members. Conversion into energy and nutrients was performed using the program Microdiet Plus for Windows Version 1.1 2000, based on the McCance and Widdowson’s food composition table fifth edition (Holland et al., 1991; Rodrigues et al., 2007). As by exploring the wider social context of decisions, socioeconomic variables available at the HBS datasets were also used as indicators of changes in the population at large such as increasing wealth and urbanization as well as proxies for the family social welfare and lifestyle. The urbanization degree of the residential area (rural, semi-urban, urban), completed education level (illiterate/elementary, secondary, higher) and occupational status (manual, non-manual, retired, unemployed, other: housewives, students, invalid) were classified under the previously established DAFNE categories (Lagiou and Trichopoulou, 2001). Household composition type was also included (1 adult, 2 adults, 1

Methods Data source Data used in the present study were taken from surveys of national representative samples of private households (defined as a person or group of persons living in the same dwelling and sharing food expenditures) accessed through the data collected by the National Statistic Institute (INE) for the Portuguese Household Budget Survey (HBS): 12 403 from European Journal of Clinical Nutrition

Figure 1 Oficial Portuguese territorial division NUT II (as adopted by Decreto-Lei n 46/89, 1989).


1265 adult with children, adults with children, adults with elderly, adults with children and elderly, 1 elderly, 2 elderly, other; considering children p18, adults 19–64 and elderly X65 years old) as the presence of children or the sole increase in the number of household members shown to result in a decrease in the quantity of food available to each member (Deaton and Paxson, 1998). The classification used attempted not only to reflect the effect of age and number of household members on food availability but also to distinguish food choices of lone parents and single households. Other variables also included for analysis were: household head age (p34, 35–49, 50–64, X65 years old), per capita household income (in quintiles), outside home food importance evaluated as the ratio of outside home food expenditures to total food expenditures (in quintiles), own food production importance evaluated as the ratio of own production money value to total food expenditures (in four categories: no production at all, terciles for those producing any quantity) and season of survey participation.

Evaluation of diet quality The food and nutrient availability obtained as described above for the Portuguese households was then summarized into two dietary scores: the Healthy Diet Indicator revised (HDIr), based on the extent to which households met with WHO population dietary goals for the prevention of chronic diseases (a revised version from the one proposed by Huijbregts et al., 1997) and the Mediterranean Adequacy Index (MAI) evaluating the extent of adherence to the Mediterranean food pattern (as proposed by Alberti-Fidanza et al., 1999) (Table 1). Originally developed to be used at an individual level, in this study dietary indexes were applied to the average per capita daily availability values of each household considered in the HBS samples. Subsequently, to join the food-based and nutrient approaches, overall diet quality was broadly defined as households with either low adherence to Mediterranean food pattern (MAIp2) or low compliance with WHO population goals (HDIrp4).

Statistical analysis For 1990, 1995 and 2000, MAI and HDIr were computed for the overall population and for each Portuguese region as median values, 10th and 90th percentiles. As the trends for HDIr are virtually non-existent and to better describe the Portuguese households’ dietary patterns, three different range groups (low, intermediate and high) were established (Table 2) and the proportion of families within each of them was also computed. To analyse tendencies over the three survey years under study, linear trend was computed using w2-test with significance set at the 0.05 level. The magnitude of the socioeconomic differences between households with overall low-diet quality and the remaining

Table 1

Criteria used for obtaining household dietary indexes

Healthy Diet Indicator revised a,b Mediterranean Adequacy Index c 10–15% energy from protein 50–70% energy from complex carbohydrates o10% energy from saturated fatty acids 6–10% energy from polyunsaturated fatty acids (PUFA) o10% energy from simple sugars o300 mg cholesterol 425 g dietary fibre X400 g of fruit and vegetables X30 g of pulses and nuts o2000 mg sodium o4% energy from alcohol

Ratio of the energy contribution from the ‘Mediterranean food groups’ (A) to the sum of the energy contribution from the ‘non-Mediterranean food groups’ (B), defined as: A (cereal and cereal products (excluding sweet bakery products), pulses, potatoes, vegetables, fruit, nuts, fish, vegetable oils and wine); B (milk and dairy products, meat and meat products, eggs, animal fats, soft drinks, sugar and sugar products (including sweet bakery products)).

a Revised version from the one proposed by Huijbregts et al. (1997), where cutoff points for PUFA and fibre were changed in accordance with the new 2003 WHO/FAO recommendations, and values for sodium (WHO/FAO, 2003) and alcohol energy contribution (WHO, 1990) were also taken into account. b A 1-point value was attributed if the average availability of a household was within the recommended intervals and a 0 in the other cases. c As proposed by Alberti-Fidanza et al. (1999), food grouping based on the DAFNE food classification system (European Commission, 2005).

Table 2 Criteria for forming three range groups within dietary indexes scores Mediterranean Adequacy Index a Low (p2) Equal or below the national median rounded off to the nearest whole number High (X4) Equal or above the lower band of the reference Mediterranean diet settled by Fidanza et al. (2004) (based on the pattern observed in the southern Italian rural village of Nicotera in the late 1950s, due to its similarity to the one found in the Greek island of Corfu) Intermediate The remaining possible values (42 and o4) Healthy Diet Indicator revised b Low (p4) Equal or below the national median, which includes the first five possible values of the scale High (X7) Equal or above the last five possible values of the scale, to form a group admitting similar number of possible values as those in the low category Intermediate The remaining values of the scale (5–6) a May vary from 0 to positive infinite; the higher the score, the closest to the Mediterranean diet concept was considered to be the family. b Allowing only whole numbers, the range score that a family could obtain varied from 0 (lowest compliance with recommendations) to 11 points (highest compliance with recommendations).

ones was then analysed for the most recent survey year (2000). By means of logistic regression, odds ratios (OR) with 95% confidence intervals (CI) were calculated both at a univariate and a multivariate level, where the stepwise procedure was applied. Significance was set at the 0.05 level. European Journal of Clinical Nutrition


European Journal of Clinical Nutrition

36.3 10.2 53.5 67.9 21.7 48.7 29.6 82.2 5.3 94.3 0.4 21.6 22.8 74.6 43.4 26.7 12.7 40.7 9.8 49.6 71.6 26.2 35.4 38.4 81.3 4.2 95.6 0.2 26.1 31.7 69.6 37.8 26.3 15.8 37.2 13.8 49.0 74.0 17.9 51.1 31.1 75.3 7.0 92.3 0.7 22.0 38.9 63.7 41.2 31.6 17.3 53.2 11.5 35.4 79.3 26.2 54.8 19.0 92.6 22.4 74.3 3.3 21.4 13.0 93.0 58.2 18.2 2.5 53.5 11.1 35.4 80.8 21.8 49.3 28.9 91.0 16.6 78.7 4.7 20.6 13.1 93.2 59.2 17.9 4.0 42.8 28.7 28.5 80.1 12.3 50.2 37.5 91.4 28.5 64.5 6.9 30.9 10.4 97.5 70.6 9.0 2.7 46.1 10.9 43.0 58.1 26.1 49.8 24.1 80.3 7.3 92.4 0.3 13.0 23.8 82.5 34.3 12.0 14.2 51.6 8.4 40.1 57.2 31.1 41.9 27.0 74.9 3.4 96.5 0.0 10.9 30.1 78.9 29.9 11.1 20.7 48.3 19.1 32.6 62.4 15.0 54.0 30.9 81.5 11.9 87.3 0.8 12.5 22.2 90.0 41.8 8.7 14.3 33.9 9.0 57.1 64.3 23.3 46.3 30.4 85.4 5.8 93.7 0.5 20.6 19.0 82.9 42.6 23.8 10.8 43.8 8.3 47.9 63.4 32.6 28.4 39.1 81.7 2.5 97.3 0.1 29.6 22.3 80.3 40.9 22.7 13.1 41.4 11.0 47.6 71.2 23.9 41.6 34.5 71.3 2.1 97.7 0.2 14.2 45.0 70.0 31.7 45.6 18.4 40.7 12.3 47.0 67.6 25.4 46.4 28.2 84.4 7.1 92.7 0.2 24.0 22.3 87.1 34.6 18.1 17.2 47.3 10.5 42.2 69.5 30.9 39.7 29.4 83.7 7.2 92.3 0.4 23.6 27.4 83.2 31.7 19.7 16.5 49.4 15.1 35.6 64.2 21.8 40.8 37.3 72.3 2.5 97.1 0.4 15.9 39.5 79.9 28.1 30.9 20.0 30.5 7.4 62.1 53.4 23.5 52.4 24.1 75.0 2.5 97.4 0.1 17.7 25.8 78.5 40.6 34.8 11.4 35.4 7.2 57.4 57.6 26.7 37.4 36.0 75.1 2.9 97.0 0.0 20.5 28.3 76.0 39.6 29.9 12.3 31.5 8.2 60.3 61.9 16.4 53.9 29.7 69.4 3.4 96.4 0.2 16.5 32.2 73.0 41.7 38.6 16.4 37.3 12.3 50.4 71.5 22.2 47.3 30.5 85.0 6.9 92.3 0.7 27.9 22.5 70.6 40.9 23.6 18.7 39.4 10.4 50.2 79.0 28.1 38.1 33.8 86.6 4.4 95.3 0.2 32.2 41.1 62.7 33.7 27.6 24.3 41.9 17.0 41.1 79.3 20.6 51.9 27.5 80.2 7.2 92.3 0.5 27.2 50.7 55.4 36.8 30.5 23.3 39.0 11.6 49.4 79.0 18.5 45.9 35.6 85.1 5.4 94.2 0.4 22.7 20.8 67.5 47.5 23.3 10.9 43.2 11.9 44.8 81.8 23.5 31.0 45.6 82.9 4.1 95.8 0.1 30.0 33.0 60.7 37.3 24.9 15.3 36.7 16.2 47.1 84.3 17.4 50.3 32.3 77.0 9.5 89.6 0.8 26.4 41.5 50.9 43.4 27.4 15.6

1995 1990 1995 2000 1990 1995 2000 1990

Madeira Azores Algarve Alentejo Lisbon and Tejo valley Centre North

Abbreviations: PUFA, polyunsaturated fatty acids; SFA, saturated fatty acids. *Values with significant over time trend (Po0.05) are underlined.

This study results showed that adherence to the Mediterranean food pattern was low in all Portuguese regions. Varying from 1.7 to 2.9, MAI values were far from 4.0, the lower band

Protein 10–15% o10% 415% SFA o10% PUFA 6–10% o6% 410% Cholesterol o300 mg Complex CHO 50–70% o50% 470% Sugar o10% Fibre 425 g Alcohol o4% Sodium o2000 mg Fruit and Vegetable X400g Pulses and Nuts X30 g

Discussion

% families complying with *

Across all regions lower compliance percentages were found for complex carbohydrates and for pulses and nuts. Conformity with fibre recommendations significantly decreased in all regions except in Madeira and Azores. Compliance with alcohol contribution increased in all the regions but decreased in Azores and Madeira. While compliance with sodium upper limit improved in the Centre, Alentejo and Algarve, it reduced in Madeira and the Azores (Table 3). In general, around two-thirds of the families scored HDIr below or equal to 4. Madeira presented the lowest proportion of households within the lower HDIr group. During the decade, significant increases in the proportion of households within the lower HDIr group, and consequent decreases in the proportion of those with high HDIr values, were observed in the North, Centre, Lisbon and Tejo valley and Alentejo (Table 4). During the decade, Madeira presented the highest MAI scores and Lisbon and Tejo valley and the Azores the lowest ones. In all regions, more than 60% of the families presented scores below 4 (Table 5). Table 6 presents the proportion of households with low HDIr, low MAI and low overall diet quality by socioeconomic categories. In this descriptive analysis it is visible that values of both dietary indexes increase along with increasing education level, urbanization degree, household income and outside home food consumption expenditures. The influence of geographic and socioeconomic variables in determining households’ overall low diet quality is presented in Table 7. In the univariate analysis, it is interesting to notice that the odds of belonging to the low diet-quality group increased when education, urbanization, income or reported expenditure on food outside the home increased. On the other hand, the odds decreased with age and own food production importance. In the mutually adjusted model, households whose head had secondary or higher education, from semi-urban or urban areas, from the Azores region, with income within the fourth or fifth quintile and outside home food expenditures higher than the first quintile were more likely to be at the low dietquality group. On the contrary, households whose head was 50–64 years old, with only elderly members, from Madeira, interviewed in summer, or with own food production were more likely to have a better diet-quality level.

Table 3 Proportion of families complying with WHO population dietary goals (based on per capita daily household availability) by Portuguese region from 1990 to 2000

Results

Portugal

Statistical analysis was performed using SPSS 14.0 for Windows.

2000 1990 1995 2000 1990 1995 2000 1990 1995 2000 1990 1995 2000 1990 1995 2000

1266


1267 Table 4 Portuguese households’ compliance with the WHO population dietary goals (based on per capita daily household availability) by Portuguese region from 1990 to 2000 Household diet quality

North

Centre

Lisbon and Tejo valley

1990 1995 2000 1990 1995 2000 1990 HDIr (median) Percentile 10 Percentile 90 % families with* Low p4 Intermediate 5–6 High X7

4.0 2.0 6.0

4.0 3.0 6.0

4.0 3.0 6.0

5.0 3.0 6.0

5.0 3.0 6.0

1995

Alentejo

Algarve

Azores

Madeira

Portugal

2000 1990 1995 2000 1990 1995 2000 1990 1995 2000 1990 1995 2000 1990 1995 2000

4.0 2.0 6.0

4.0 2.0 6.0

4.0 2.0 6.0

4.0 2.0 6.0

55.6 54.0 61.3 48.7 47.6 56.6 36.7 39.3 33.2 42.8 42.4 36.5 7.7 6.7 5.5 8.5 10.0 6.9

63.3 31.1 5.6

63.6 31.8 4.7

65.6 31.5 2.9

4.0 2.0 6.0

4.0 3.0 6.0

4.0 3.0 6.0

4.0 2.0 6.0

4.0 3.0 6.0

4.0 2.0 6.0

4.0 2.0 6.0

4.0 2.0 6.0

4.0 2.0 6.0

5.0 3.0 6.0

5.0 3.0 6.0

5.0 3.0 6.0

4.0 2.0 6.0

4.0 3.0 6.0

4.0 2.0 6.0

56.8 52.2 57.2 54.5 55.9 63.8 62.4 64.6 66.1 43.0 44.6 42.0 56.8 56.1 61.4 36.3 41.7 37.1 37.5 37.4 30.5 33.6 31.8 30.4 48.2 49.5 48.3 36.1 37.5 33.6 6.9 6.1 5.7 8.0 6.7 5.7 4.0 3.6 3.5 8.8 5.9 9.8 7.1 6.5 4.9

Abbreviation: HDIr, Healthy Diet Indicator revised. * Values with significant over time trend (Po0.05) are underlined.

Table 5 Portuguese households’ adherence to Mediterranean food pattern (based on per capita daily household availability) by Portuguese region from 1990 to 2000 Household diet quality

North

Centre

Lisbon and Tejo valley

1990 1995 2000 1990 1995 2000 1990 MAI (median) Percentile 10 Percentile 90 % families with * Low p2 Intermediate 2–4 High X4

2.6 0.9 7.4

2.6 1.2 5.9

2.3 0.9 5.4

2.5 1.0 7.5

2.7 1.2 6.6

1995

Alentejo

Algarve

Azores

Madeira

Portugal

2000 1990 1995 2000 1990 1995 2000 1990 1995 2000 1990 1995 2000 1990 1995 2000

2.3 0.9 6.0

1.9 0.8 4.7

2.0 0.9 4.4

1.8 0.8 4.1

36.5 32.8 42.0 36.4 30.6 41.6 35.3 43.6 39.1 37.5 43.0 38.8 28.2 23.6 18.8 26.1 26.4 19.7

54.2 32.3 13.5

48.9 37.6 13.4

57.3 32.0 10.7

2.1 0.9 4.7

2.6 1.1 5.4

2.4 1.0 5.3

2.1 0.9 4.7

2.6 1.2 5.6

2.2 1.0 5.3

1.7 0.7 4.3

1.7 0.9 3.2

1.7 0.8 3.6

2.9 1.0 9.9

2.6 1.2 5.6

2.8 1.3 7.0

2.2 0.9 6.1

2.4 1.1 5.5

2.1 0.9 5.0

46.6 32.5 38.6 45.1 33.2 43.7 62.4 67.2 62.0 32.4 31.9 28.5 43.8 38.6 47.1 38.2 46.1 41.2 38.0 43.8 37.9 26.1 27.3 30.9 30.2 46.4 43.2 34.6 41.3 36.6 15.2 21.4 20.1 16.9 23.1 18.4 11.5 5.5 7.1 37.4 21.6 28.3 21.6 20.0 16.2

Abbreviation: MAI, Mediterranean Adequacy Index. * Values with significant over time trend (Po0.05) are underlined.

of the reference Mediterranean diet settled by Fidanza et al. (2004). In addition, except for Alentejo, the proportion of families with MAI values equal or higher than 4.0 has either significantly decreased or remained stable from 1990 to 2000, supporting the idea that the country’s traditional Mediterranean food pattern is in transition, a situation which has also been suggested by others (Marques-Vidal et al., 2006). To evaluate adherence to Mediterranean food pattern, several indexes have been proposed, among which the Mediterranean Diet Score proposed by Trichopoulou et al. (1995, 2005) is the most extensively employed (Bach et al., 2006). In this study the MAI was preferred due to the use of a quotient measure instead of an add/subtract score, which makes cut-off points unnecessary. Using cut-off points based on the distribution of selected food groups in the population being evaluated would make it difficult to appraisal time trends and comparison between groups or even studies (Bach et al., 2006). Furthermore, the use of energy contributions instead of absolute values was also taken as an advantage, allowing for adjustment to the total energy available and thus for more accurate comparisons. Diet-quality scores may either be based on a priori assumptions, derived from recommended diets or guidelines

(Lagiou et al., 1999), or in a posteriori derived dietary patterns (Naska et al., 2006). This latter approach often involves the use of statistical reducing methods such as cluster or factor analysis (Knoops et al., 2006). Beside the limitations that have been pointed out to any of the methods (Kant, 2004), inasmuch as the components of the score are assumed to contribute equally, the a priori approach was chosen for the present study as it allows to evaluate compliance with the scientifically developed recommendations. In addition, such dietary indexes are easily reproducible and comparable between studies, not depending on the arbitrary decisions often required in various steps of the statistical procedures involved in the a posteriori approach (Kant, 2004). Conformity with WHO recommendations was not high; around 60% of the families of each region presented HDIr values lower or equal than 4, out of the maximum possible score of 11. Within the components of this index, the increase in the proportion of energy coming from saturated fatty acids, and the reduction in the availability of fibre should be noted. An opposing positive trend can be seen in the decreases in sodium and alcohol contribution. However, the slight decreases observed in both national supply (INE, 1999) and self-reported intake (Marques-Vidal and Dias, 2005) of alcoholic beverages indicate that our positive results European Journal of Clinical Nutrition


1268 Table 6 Portuguese households’ adherence to Mediterranean food pattern, compliance with WHO population dietary goals and overall diet quality by socio-demographic categories in 2000 Total Na (%)

Low HDIr (p4) %b

Low MAI (p2) %c

Low overall diet quality (HDIrp4 or MAI p2)%

Age of the household head (years) p34 35–49 50–64 X65

8 27 30 35

72 66 60 53

61 54 44 37

82 77 71 63

Education of the household head Illiterate/Elementary Secondary Higher

79 15 6

58 73 75

42 63 74

68 83 89

Occupation of the household head Manual Non-manual Retired Unemployed Other

32 22 36 3 7

63 71 53 62 51

47 60 37 47 43

73 82 64 71 64

Composition type 1 Adult 2 Adults 1 Adult þ children Adults þ children Adults þ elderly Adults þ children þ elderly 1 Elderly 2 Elderly Other

5 10 2 30 13 4 12 11 13

62 60 68 66 54 59 49 50 66

57 46 61 52 39 39 43 33 50

77 71 77 76 64 69 62 60 77

Locality urbanization degree Rural Semi-urban Urban

18 29 53

53 60 65

35 40 55

63 69 77

Region North Centre Lisbon and Tejo valley Alentejo Algarve Azores Madeira

33 18 35 6 4 2 2

61 57 66 57 64 66 42

42 42 57 39 44 62 29

71 67 77 66 72 79 52

Season of survey interview Winter Spring Summer Autumn

28 23 26 23

63 64 57 61

49 51 42 47

74 74 68 72

Income per capita First quintile Second quintile Third quintile Fourth quintile Fifth quintile

20 20 20 20 20

56 59 60 63 70

36 41 44 50 65

65 69 70 74 83

33 43 48 51 58

58 68 74 76 80

Household characteristics

Outside home food consumption expenditures (% of total food expenditures) First quintile 15 49 Second quintile 20 58 Third quintile 23 64 Fourth quintile 23 65 Fifth quintile 19 67



1269 Table 6 Continued Household characteristics

Total Na (%)

Low HDIr (p4) %b

Own food production importance money value (% of total food expenditures) No production 69 66 First tercile 10 58 Second tercile 10 50 Third tercile 11 48

Low MAI (p2) %c

Low overall diet quality (HDIrp4 or MAI p2)%

53 40 31 32

76 69 61 58

Abbreviation: HDIr, Healthy Diet Indicator revised; MAI, Mediterranean Adequacy Index. a Total N comprises 10 020 households. b Healthy Diet Indicator revised, evaluating compliance with WHO population goals (may vary from 0 to 11, the higher the score the better the compliance). c Mediterranean Adequacy Index, evaluating adherence to Mediterranean food pattern (may vary from 0 to positive infinite and the reference value of good adherence is equal or higher than 4).

may be counterbalanced by an increase in consumption outside the home. Due to its increasing importance (Gracia and Albisu, 2001; INE, 2002), the lack of information on eating out food consumption is an important limitation of HBS data. Although it may be argued that the type of food people choose to eat at home is not remarkably different from the food they choose when eating out, it can also be defended that people go out exactly to look for different experiences. Trying to overcome this drawback, the ratio outside home food expenditures to total food expenditures was used as a proxy for the importance of eating out within the household. However, as it is the case for the domestic food production expenditures ratio, no assumption of its quality or quantity can be made. Further research on the eating out habits of European subjects is indispensable for a comprehensive understanding of its real influence on dietary patterns. Together with the limitation inherent to the use of only 3 years data points, changes in the age structure of the Portuguese population among the time periods under study, increasingly older, may have introduced some bias in trend analysis. However, due to the relatively small time period, variation was not so large (HBS proportion of individuals o19 years old varied from 27 to 22% and those 464 years old increased from 14 to 16%) and is thus not expected to significantly impair comparisons. To our knowledge, HDI was for the first time used in HBSderived data in the present study. In addition, HDIr was computed differently from the original HDI, taking into account other components and cut-off points changed on the basis of the 2003 WHO recommendations, which do not allow us to make comparisons with other results. The only previous study applying the MAI to data at the household level (Bach et al., 2006) was done for the Spanish population (Fuentes-Bol, 2002), and shows that in 2000 their adherence to Mediterranean food pattern was lower (1.7) than in Portugal (2.1). Despite being obtained from different methods of dietary assessment, results from other Italian populations also showed that MAI increased with subject’s age and decreased over time (Alberti-Fidanza and Fidanza, 2004). Elderly people evaluated around 1990 within the

HALE (Healthy Ageing: a Longitudinal study in Europe) study presented MAI scores ranging from 1.0 to 1.5 in the Northern Europe, and scores ranging from 2.0 to 3.6 in the Southern Europe (Knoops et al., 2006). As expected, the Portuguese regions scores are closer to the Southern European subjects than to the Northern ones. Linking low-quality diets and high income and education, this study results are counter to findings from individualbased studies, where healthier diets are generally associated with higher education and income and/or to higher social class (Roos and Prattala, 1999; Moreira and Padraõ, 2004). However, our findings are consistent with other studies evaluating households diet. Adherence to the Mediterranean food pattern was also higher in Spanish households from lower social classes than in those from higher social classes (Fuentes-Bol, 2002). Within Portuguese households data from 1995 the food pattern of ‘beverages and convenience food’ were more common among those whose head had secondary or higher education, with non-manual occupation, located at the semi-urban or urban areas, and less common in households with only elderly members (Naska et al., 2006). The observation that the odds of belonging to the low dietquality group increased when education level and degree of urbanization also increases can be partially explained by use of the nutrition transition theory and its stages (Popkin, 2006). In Portugal we may be seeing the nutrition transition between stages 4 (increase in degenerative disease and dietrelated NCDs) and 5 (behavioural change and decreases in diet-related NCDs). It appears that urbanization and increased income levels are indicative of changes in food choice reflected in increased dietary intake of animal protein, processed foods and higher fat and sugar levels and lower incomes and living in rural areas indicative of healthier diets and healthier food choice. The other way of looking at this is that low-income rural populations do not have the choice, which is not necessarily a bad thing from a health perspective but from a social inclusion perspective may not be ideal. HBS-derived data has the strength of being national representative and of using standardized and regularly collected data, allowing not only the study of time trends European Journal of Clinical Nutrition


1270 Table 7 Socio-demographic characterization of low overall diet qualitya households in Portugal 2000 Household characteristics

Continued

Household characteristics

Odds ratiob and (95% confidence intervals)

b

Odds ratio and (95% confidence intervals) Univariate

Age of the household head (years) p34d 35–49 0.78 (0.64–0.96)s 50–64 0.50 (0.41–0.61)s X65 0.37 (0.31–0.45)s Education of the household head Illiterate/elementaryd Secondary 2.71 (2.31–3.17)s Higher 3.97 (2.95–5.34)s

Multivariate

0.86 (0.70–1.06) 0.65 (0.53–0.81)s 0.81 (0.60–1.08)

1.53 (1.28–1.82)s 1.87 (1.35–2.59)s

Composition type 1 Adultd 2 Adults 1 Adult þ children Adults þ children Adults þ elderly Adults þ children þ elderly 1 Elderly 2 Elderly

0.80 1.47 1.13 0.62 0.89 0.55 0.51

Locality urbanization degree Rurald Semi-urban Urban

1.24 (1.11–1.39)s 1.70 (1.53–1.89)s

Region Northd Centre Lisbon and Tejo valley Alentejo Algarve Azores Madeira

0.73 1.22 0.72 0.88 1.55 0.43

Season of survey interview Winterd Spring Summer Autumn

1.08 (0.95–1.21) 0.84 (0.75–0.94)s 0.96 (0.85–1.09)

1.10 (0.97–1.25) 0.86 (0.76–0.97) 1.00 (0.88–1.14)

Income per capita First quintiled Second quintile Third quintile Fourth quintile Fifth quintile

1.05 1.20 1.53 2.53

1.04 1.08 1.19 1.49

(0.63–1.02) (0.95–2.26) (0.90–1.40) (0.49–0.78)s (0.66–1.19) (0.44–0.70)s (0.40–0.64)s

(0.62–0.85)s (1.03–1.45)s (0.62–0.84)s (0.76–1.02) (1.32–1.82)s (0.37–0.50)s

(0.93–1.18) (1.06–1.37)s (1.34–1.74)s (2.18–2.93)s

0.96 1.35 1.11 0.79 1.02 0.68 0.69

(0.74–1.25) (0.85–2.14) (0.87–1.43) (0.59–1.06) (0.73–1.42) (0.49–0.95)s (0.50–0.95)s

1.13 (1.01–1.28)s 1.21 (1.06–1.37)s

1.01 1.10 0.89 0.95 1.77 0.49

Univariate

Multivariatec

c

Occupation of the household head Manuald Non-manual 1.85 (1.61–2.12)s Retired 0.69 (0.63–0.77)s n.s. Unemployed 0.92 (0.68–1.24)

(0.86–1.19) (0.92–1.32) (0.75–1.05) (0.81–1.11) (1.49–2.11)s (0.42–0.57)s

(0.91–1.18) (0.94–1.24) (1.03–1.38)s (1.25–1.78)s

Outside home food consumption expenditures (% of total food expenditures) First quintiled Second quintile 1.49 (1.32–1.68)s 1.18 (1.03–1.34)s Third quintile 1.60 (1.41–1.80)s 1.18 (1.03–1.36)s Fourth quintile 1.79 (1.57–2.05)s 1.21 (1.04–1.41)s Fifth quintile 2.29 (1.98–2.64)s 1.37 (1.16–1.61)s


Table 7

Own food production importance money value (% of expenditures) No productiond First tercile 0.70 (0.66–0.80)s Second tercile 0.48 (0.42–0.55)s Third tercile 0.42 (0.37–0.48)s

total food 0.85 (0.73–0.98)s 0.62 (0.54–0.72)s 0.55 (0.48–0.64)s

s, significant (Po0.05); n.s., variable non-significant (PX0.05) for the final model. a Households with dietary indices below the Portuguese overall medians, considering either the Healthy Diet Indicator (HDIrp4) or the Mediterranean Adequacy Index (MAIp2). b OR were calculated for two classes: households with low diet quality versus remaining households. c Multivariate analysis made by stepwise logistic regression, OR mutually adjusted for all the other variables considered in the table. d Reference categories, corresponding to an OR of 1.0.

but also the comparison between regions or countries. Applying dietary indexes to such data proved to help on summarizing and interpreting households’ food and nutrient availability patterns. This study results are in line with others reporting the westernization of Mediterranean countries food habits (Trichopoulos and Lagiou, 2004; Garcia-Closas et al., 2006). Between 1990 and 2000, the decreasing adherence to Mediterranean food pattern and the diminishing compliance with WHO dietary goals shows that Portuguese households have reduced its diet quality. The observed influence of urbanization and region on diet quality highlights once again (Serra-Majem et al., 1997) the importance of considering regional nutrition strategies and food policy planning.

Acknowledgements This study was conducted in the context of the DAFNE IV project entitled ‘European food availability databank based on household budget surveys’ of the DG-SANCO of the European Commission. We thank the Instituto Nacional de Estatı´stica, the Portuguese institution responsible for the Household Budget Surveys, for allowing us the use of its databases. The first author is a PhD student with a scholarship sponsored by PRODEP III, European Commission.

References Alberti-Fidanza A, Fidanza F (2004). Mediterranean Adequacy Index of Italian Diets. Public Health Nutr 7, 937–941. Alberti-Fidanza A, Fidanza F, Chiuchiu MP, Verducci G, Fruttini D (1999). Dietary studies on two rural Italian population groups of


1271 the Seven Countries Study 3. Trend of food and nutrient intake from 1960 to 1991. Eur J Clin Nutr 53, 854–860. Assman G, Sacks F, Awad AB, Bonanome A, Booyse FM, Carmena R et al. (2000). International consensus statement: dietary fat, the Mediterranean diet, and lifelong good health. 2000 International Conference on Mediterranean Diet.(Accessed 18 December 2005: http://www.chd-taskforce.de/2000consensusstatement/index_e.htm). Bach A, Serra-Majem L, Carrasco JL, Roman B, Ngo J, Bertomeu I et al. (2006). The use of indexes evaluating the adherence to the Mediterranean diet in epidemiological studies: a review. Public Health Nutr 9, 132–146. CEC (1993). Nutrient and energy intakes for the European Community. Reports of the Scientific Committee for Food: 31st Series. Directorate-General Industry, Commission of the European Communities: Luxembourg. CEC (2005). Green paper, promoting healthy diets and physical activity: a European dimension for the prevention of overweight, obesity and chronic diseases. Commission of the European Communities. Deaton A, Paxson C (1998). Economies of scale, household size, and the demand for food. J Political Econ 106, 897–930. EuroDiet (2001). EuroDiet core report: nutrition and diet for healthy lifestyles in Europe: science and policy implications.(Accessed: 15 July 2006) (Available at:http://ec.europa.eu/health/ph_determinants/ life_style/nutrition/report01_en.pdf). European Commission (2005). The DAFNE food classification system. Operationalisation in 16 European countries.DG-SANCO, Directorate-General for Health and Consumer Protection. Services of the European Commission: Luxembourg. Fidanza F, Alberti-Fidanza A, Lanti M, Menotti A (2004). Mediterranean Adequacy Index: correlation with 25-year mortality from coronary heart disease in the Seven Countries Study. Nutr Metab Cardiovasc Dis 14, 254–258. Fieldhouse P (1995). Food and Nutrition, Customs and Culture,2nd edn. Chapman & Hall: London. Fuentes-Bol M (2002). La dieta mediterrańea: su evolucioń.In: SerraMajem L, Ngo J (eds). Que´ es la Dieta Mediterrańea? Fundacioń para el desarrollo de la Dieta Mediterrańea. Nexus Editions: Barcelona, pp 60–73. Garcia-Closas R, Berenguer A, Gonza´lez CA (2006). Changes in food supply in Mediterranean countries from 1961 to 2001. Public Health Nutr 9, 53–60. Gracia A, Albisu LM (2001). Food consumption in the European Union: main determinants and country differences. Agribusiness 17, 469–488. Holland B, Welch AA, Unwin ID, Buss DH, Paul AA, Southgate DAT (1991). McCance and Widdowson’s The Composition of Foods,5th edn, Royal Society of Chemistry: Cambridge. Huijbregts P, Feskens E, Rasanen L, Fidanza F, Nissinem A, Menotti A et al. (1997). Dietary pattern and 20 year mortality in elderly men in Finland, Italy, and the Netherlands: longitudinal cohort study. BMJ 315, 13–17. INE (1990). Inque´rito aos orçamentos familiares 1989/90. Metodologia. Instituto Nacional de Estatı´stica: Portugal. INE (1997). Inque´rito aos orçamentos familiares 1994/95. Metodologia. Instituto Nacional de Estatı´stica: Portugal. INE (1999). Balança alimentar Portuguesa 1990/97. Instituto Nacional de Estatı´stica: Portugal. INE (2002). Inque´rito aos orçamentos familiares 2000. Instituto Nacional de Estatı´stica: Portugal. Kant A (1996). Indexes of overall diet quality: a review. J Am Diet Assoc 96, 785–791. Kant A (2004). Dietary patterns and health outcomes. J Am Diet Assoc 104, 615–635. Knoops KTB, Groot de LC, Fidanza F, Alberti-Fidanza A, Kromhout D, Van Staveren WA (2006). Comparison of three different dietary scores in relation to 10-year mortality in elderly European subjects: the Hale project. Eur J Clin Nutr 60, 746–755.

Lagiou P, Trichopoulou A, DAFNE contributors (2001). DAta Food Networking. The DAFNE initiative: the methodology for assessing dietary patterns across Europe using household budget survey data. Public Health Nutr 4, 1135–1141. Lagiou P, Trichopoulou A, Henderickx HK, Kelleher C, Leonhauser IU, Moreiras O et al. (1999). Household budget survey nutritional data in relation to mortality from coronary heart disease, colorectal cancer and female breast cancer in European countries. DAta Food Networking. Eur J Clin Nutr 53, 328–332. Marques-Vidal P, Dias CM (2005). Trends and determinants of alcohol consumption in Portugal: results from the National Health Surveys 1995 to 1996 and 1998 to 1999. Alcohol Clin Exp Res 29, 89–97. Marques-Vidal P, Ravasco P, Dias CM, Camilo ME (2006). Trends of food intake in Portugal, 1987–1999: results from the National Health Surveys. Eur J Clin Nutr 60, 1414–1422. Mennel S, Murcott A, Otterloo AH (1992). The Sociology of Food, Eating, Diet and Culture. SAGE Publications, London. Moreira P, Padraõ P (2004). Educational and economic determinants of food intake in Portuguese adults: a cross-sectional survey. BMC Public Health 4. Naska A, Fouskakis D, Oikonomou E, de Almeida MDV, Berg MA, Gedrich K et al. (2006). Dietary patterns and their sociodemographic determinants in 10 European countries: data from the DAFNE databank. Eur J Clin Nutr 60, 181–190. Nestle M (1995). Mediterranean Diets—historical and research overview. Am J Clin Nutr 61, 1313S–1320S. Popkin B (2006). Global nutrition dynamics: the world is shifting rapidly toward a diet linked with noncommunicable diseases. Am J Clin Nutr 84, 289–298. Ribeiro O (1998). Portugal, o Mediterraˆneo e o Atlaˆntico.(7a Ediçaõ) Lisboa, Livraria Sa´ da Costa Editora. Ricciuto L, Tarasuk V, Yatchew A (2006). Socio-demographic influences on food purchasing among Canadian households. Eur J Clin Nutr 60, 778–790. Robertson A, Tirado C, Lobstein T, Jermini M, Knai C, Jensen JH et al. (2004). Food and Health in Europe: A New Basis for Action. WHO Regional Publications: Copenhagen,European Series no 96. Rodrigues SSP, de Almeida MDV (2001). Portuguese household food availability in 1990 and 1995. Public Health Nutr 4, 1167–1171. Rodrigues SSP, Lopes C, Naska A, Trichopoulou A, de Almeida MDV (2007). Comparison of national food supply, household food availability and individual food consumption data in Portugal. J Public Health,First online publication (DOI 10.1007/s10389-0070102-8). Roos G, Prattala R (1999). Disparities in Food Habits, Review of Research in 15 European Countries (Disparities part of the FAIR-97-3096 project).Publications of the National Public Health Institute: Helsinki. Serra-Majem L, Ferro-Luzzi A, Bellizzi M, Salleras L (1997). Nutrition policies in Mediterranean Europe. Nutr Rev 55, S42–S57. Serra-Majem L, Roman B, Estruch R (2006). Scientific evidence of interventions using the Mediterranean diet: a systematic review. Nutr Rev 64 (2 Part 2), S27–S47. Trichopoulos D, Lagiou P (2004). Mediterranean diet and overall mortality differences in the European Union. Public Health Nutr 7, 949–951. Trichopoulou A (2001). Mediterranean diet: the past and the present. Nutr Metab Cardiovasc Dis 11, 1–4. Trichopoulou A, Kouris-Blazos A, Wahlqvist ML, Gnardellis C, Lagiou P, Polychronopoulos E et al. (1995). Diet and overall survival in elderly people. BMJ 311, 1457–1460. Trichopoulou A, Naska A, Costacou T, DAFNE III Group (2002). Disparities in food habits across Europe. Proc Nutr Soc 61, 553–558. Trichopoulou A, Orfanos P, Norat T, Bueno-de-Mesquita B, Ooke MC, Peeters PH et al. (2005). Modified Mediterranean diet and survival: the EPIC-elderly prospective cohort study. BMJ 330, 991–997.



1272 WHA (2004). Global strategy on diet, physical activity and health.Fifty-seventh World Health Assembly, 57.17, 22 May. WHO (1990). Diet, Nutrition, and the Prevention of Chronic Diseases, Report of a Study Group. Technical Report Series No. 797. World Health Organization: Geneva. WHO/FAO (2003). Diet, Nutrition and the Prevention of Chronic Diseases: Report of a Joint WHO/FAO Expert Consultation.WHO


Technical Report Series, Technical Report 916. WHO: Geneva; FAO: Rome. Willett W (2006). The Mediterranean diet: science and practice. Public Health Nutr 9, 105–110. Willett WC, Sacks F, Trichopoulou A, Drescher G, Ferro-Luzzi A, Helsing E et al. (1995). Mediterranean diet pyramid: a cultural model for healthy eating. Am J Clin Nutr 61, 1402S–1406S.