Little is known about the clustering patterns of lifestyle behaviours in adult populations. We explored clusters in multiple lifestyle behaviours including physical activity (PA), smoking, alcohol use and eating habits in a sample of adult population. A cross-sectional and multi-centre study was performed with six participating groups distributed throughout Spain. Participants (n 1327) were part of the Lifestyles and Endothelial Dysfunction (EVIDENT) study and were aged between 20 and 80 years. The lifestyle and cardiovascular risk (CVR) factors were analysed using a clustering method based on the HJ-biplot coordinates to understand the variables underlying these groupings. The following three clusters were identified. Cluster 1: unhealthy, 677 subjects (51 %), with a slight majority of men (58·7 %), who were more sedentary and smokers with higher consumption of whole-fat dairy products, bigger waist circumference as well as higher TAG levels, systolic blood pressure (SBP) and CVR. Cluster 2: healthy/PA, 265 subjects (20 %), including 24·0 % of males with high PA. Cluster 3: healthy/diet, including 29 % of the participants, with a higher consumption of olive oil, fish, fruits, nuts, vegetables and lower alcohol consumption. Using the unhealthy cluster as a reference, and after adjusting for age and sex, the multiple regression analysis showed that belonging to the healthy/PA cluster was associated with a lower waist circumference, body fat percentage, SBP and CVR. In summary, the three clusters were identified according to lifestyles. The ‘unhealthy’ cluster had the least favourable clinical parameters, the ‘healthy/PA’ cluster had good HDL-cholesterol levels and low SBP and the ‘healthy/diet’ cluster had lower LDL-cholesterol levels and clinical blood pressure.

In recent years, increasing interest in the format of multi-centric studies among different populations in developing nations has evolved in the field of health surveys and epidemiology. Dietary intake data are most often part of these cross-cultural and cross-national collaborative efforts. Various questions have been raised about the appropriate endeavours for dietetics and nutritional sciences in developing societies, the instruments available for application, and the pitfalls and caveats in their use. An important consideration is that studies be hypothesis-driven and not mere ‘fishing expeditions’ of unfocused data gathering. All known dietary intake measurement tools are within the purview of developing country research, but they often must be adapted individually and differentially to suit a given population. In a multi-centric context, this is complicated. The watchword should be collecting comparable information across sites, not using identical approaches. Choice of dietary intake measurement tools must be honed to the hypotheses and assumptions, on the one hand, and the exigencies and pitfalls of working in the developing country milieu, in which linguistics, seasonality, migration, uncommonness of food systems and ethical considerations present barriers and caveats, on the other. Within the hypotheses, the assumptions regarding the penetration of the measured exposures must be borne in mind. Multi-centre studies in developing countries have relevance and importance in the context of food security, diet and disease, eating behaviour and satiety regulation, and nutritional anthropology.