Lifestyle Patterns and Weight Status in Spanish Adults: The ANIBES Study

Limited knowledge is available on lifestyle patterns in Spanish adults. We investigated dietary patterns and possible meaningful clustering of physical activity, sedentary behavior, sleep time, and smoking in Spanish adults aged 18–64 years and their association with obesity. Analysis was based on a subsample (n = 1617) of the cross-sectional ANIBES study in Spain. We performed exploratory factor analysis and subsequent cluster analysis of dietary patterns, physical activity, sedentary behaviors, sleep time, and smoking. Logistic regression analysis was used to explore the association between the cluster solutions and obesity. Factor analysis identified four dietary patterns, “Traditional DP”, “Mediterranean DP”, “Snack DP” and “Dairy-sweet DP”. Dietary patterns, physical activity behaviors, sedentary behaviors, sleep time, and smoking in Spanish adults aggregated into three different clusters of lifestyle patterns: “Mixed diet-physically active-low sedentary lifestyle pattern”, “Not poor diet-low physical activity-low sedentary lifestyle pattern” and “Poor diet-low physical activity-sedentary lifestyle pattern”. A higher proportion of people aged 18–30 years was classified into the “Poor diet-low physical activity-sedentary lifestyle pattern”. The prevalence odds ratio for obesity in men in the “Mixed diet-physically active-low sedentary lifestyle pattern” was significantly lower compared to those in the “Poor diet-low physical activity-sedentary lifestyle pattern”. Those behavior patterns are helpful to identify specific issues in population subgroups and inform intervention strategies. The findings in this study underline the importance of designing and implementing interventions that address multiple health risk practices, considering lifestyle patterns and associated determinants.


Introduction
Overweight and obesity have progressively increased during the last decades and have become a major issue in public health, both in developed and developing economies across the five continents [1,2]. In Spain, recent data report that more than half of adults aged 18-64 years are classified as overweight or obese [3,4]. This is particularly worrying for the negative impact of this condition on health and quality of life [2,5,6]. In fact, the Global Burden of Disease project highlights that high body mass index (BMI) values were among the main risk factors driving most death and disability combined in the country in 2015 [7].
Overweight and obesity result from an imbalance between energy intake and expenditure. The role of diet in obesity is complex. Most research in this area has focused on specific foods and nutrients [8,9]. Nevertheless, the analysis of food patterns is particularly interesting, since foods are usually consumed in combinations and those may have synergistic, antagonistic, or moderating effects [10].
In addition, to date, it is acknowledged that joint interactions of multiple variables acting at different levels influence weight gain, such as lifestyles [11][12][13], sleep, including rhythm, duration or quality of sleep [14,15], eating behaviors [16], socioeconomic level, education, and other factors [4,11]. Data-driven methods explore the similarities between different food options in specific population groups [8]. Specifically, cluster analyses include several techniques aimed at grouping together individuals sharing a number of features or similar lifestyles. This classification allows for a better understanding of the influences of behaviors and lifestyles, as well as the potential cumulative effects of an unhealthy combination of those factors on the development of overweight and obesity [16]. Research of dietary patterns and the potential combination of those with other lifestyles can contribute to identifying effective strategies for the prevention of overweight and obesity among adults, as well as its social and health consequences [8].
Considering the above, the objectives of this paper are (a) to identify food patterns in the Spanish adult population; (b) to investigate if energy balance-related behaviors tend to assemble into meaningful patterns in Spanish adults; (c) to describe existing relationships between socio-demographical factors and different lifestyle patterns; and, finally; (d) to analyze the potential association of those correlates with excess body weight.

Materials and Methods
The data for this analysis were obtained from the ANIBES study, an observational cross-sectional study conducted on a random sample of the Spanish population aged 9-75 years. The aims and procedures used in ANIBES have been previously reported [17,18].
Briefly, the sample design of the ANIBES study was based on the Spanish population census 2012 according to sex, age, residence, and regional population size. A stratified multistep sampling procedure was used, with random selection of addresses in the municipalities and age and sex quotas for individuals within households. Interlocked quotas were established for age in the regions and size of habitat within the region. The sample selection procedure was based on random paths; 128 sampling points were selected.
The final sample of the study consisted of 2009 individuals (1013 men, 50.4%, 996 women, 49.6%). In addition, a boost sample was recruited for the younger groups (9-12 years; 13-17 years, and 18-24 years) in order to ensure at least 200 individuals in each age group. Data for individuals aged 18-64 years were used for this analysis. That age range is meaningful from a risk prevention point of view, since it includes adults in the active working life span, thus, relevant when designing targeted preventive and health promotion strategies, considering different settings, such as workplaces, primary health care, and community-based interventions. In addition, the focus on the 18-64 year-old range was also relevant for comparison with previous studies conducted in Spain [19]. Data were collected between mid-September and mid-November 2013.
The final protocol of the study was approved by the ethical research committee of the Community of Madrid, Spain.

Lifestyle Factors
Diet Dietary intake was assessed by means of a face-to-face 24-h diet recall interview assisted by a food picture atlas to estimate portion sizes. In addition participants completed a three-day food record aided by a tablet device (Samsung Galaxy Tab 2 7.0, Samsung Electronics, Suwon, South Korea), two consecutive weekdays and one weekend day, recording all foods and beverages consumed at home and away from home. Processing of food record inputs, coding and data cleaning has been extensively reported elsewhere [17,18]. Energy and nutrient intakes were calculated using new specifically-developed software, (VD-FEN 2.1 software -Dietary Evaluation Programme, Spanish Nutrition Foundation, Madrid, Spain) for the ANIBES study [20]. Consumption of all food and beverages was arranged into 16 food groups, 45 subgroups, and 754 food items, for in-depth analysis, based on the structure of the food composition database considering similarities in nutrient profile (Supplementary Materials Table S1).

Physical Activity
Physical activity data were collected by face-to-face interview using the validated International Physical Activity Questionnaire (IPAQ) [21]. Total minutes per week were computed for moderate to vigorous physical activity based on the IPAQ guidelines for data processing and analyses [22]. Data were cleaned and truncated based on IPAQ guidelines and previous research [23]. Additionally, the total minutes per week of commuting-related physical activity (walking, biking) were computed. IPAQ data was used for this analysis. Z-scores of minutes per week for each type of activity were calculated.

Sleep Duration
Sleep habits included the number of hours slept per night, on average, as reported by each individual.

Smoking
Smoking habits were assessed by different items in the protocol, including "During the past year, how many cigarettes per day did you smoke on average?"

Body Measurements
Anthropometric measurements were taken individually by trained interviewers, following international standard procedures previously tested in two pilot studies [24], reported in detail elsewhere [17,18]. Body mass index (BMI) was calculated as body weight in kilograms divided by the square of body height in meters. Overweight status was defined as BMI ≥ 25; obesity as BMI ≥ 30.

Education
The education levels were established in accordance with the Spanish educational system. After preliminary analysis of the distribution of the variable, categories were collapsed and recoded into a three-point scale, as follows: (1) low (less than seven years of education; primary school or less); (2) medium (7-12 years of education; lower to higher secondary education); and (3) high (13 years or more of education; higher vocational, college, and university studies).

Geographical Area
Geographical area in the country was collapsed into four different categories: north-northwest region; eastern region; central region and southern region.

Data Cleaning
Detailed data cleaning procedures have been previously described [17,18]. Participants were considered fully eligible after a verified quality check of the input from the tablet device of adequately completed three-day food records. After data cleaning stages, individuals remained in the database if they had successfully completed both face-to-face interviews during fieldwork and had measured weight, height, and waist circumference data. Of the initial sample of 1655 recruited individuals aged 18-64 years, 1617 individuals satisfied the inclusion criteria and had complete data for all of the variables included in this analysis.

Data Analysis
All statistical tests were performed using IBM SPSS Statistics for Windows, Version 22.0 (IBM Corp., Armonk, NY, USA). Descriptive statistics were computed for each variable.

Dietary Patterns
Exploratory factor analysis was performed to identify underlying dietary patterns, using the average food weight (g/day) consumed by each individual (three-day food record plus one-day 24-h recall) from 38 food groups as input variables. Food groups were used to further collapse dietary intake data in order to avoid missing data from non-consumers of episodically consumed foods. Z-scores for each food group were calculated to prevent the components being dominated by the foods that provide the highest amounts. Bartlett's test of sphericity and the Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy were used to verify the appropriateness of factor analysis. Factors were also orthogonally rotated (the varimax option) to enhance the difference between loadings, which allowed for easier interpretability.
Factors were retained based on the following criteria: factor eigenvalue > 1.20, identification of a break point in the scree plot, the proportion of variance explained, and factor interpretability [25].
The strength and direction of the associations between patterns and food groups were described through a rotated factor loading matrix. Food groups with factor loadings > 0.30 and communality > 0.20 were retained in the patterns identified. The factor score for each pattern was constructed by summing the observed intakes of the component food items weighted by the factor loading. A high factor score for a given pattern indicated high intake of the foods constituting that food factor, and a low score indicated low intake of those foods.

Lifestyle Patterns
To identify clusters with similar dietary patterns, physical activities, sedentary activities, sleeping habits and smoking a combination of hierarchical and non-hierarchical clustering analysis was used [26]. The variables used had different arithmetic scales; thus, Z-scores were calculated to standardize the dataset before clustering, to avoid a greater contribution to the distance of variables having larger ranges than variables with smaller ranges. Univariate and multivariate outliers (>3 SD) were removed. First, hierarchical cluster analysis was performed using Ward's method, based on squared Euclidian distances. Several possible cluster solutions were identified and compared to inform the next step, considering the coefficients and fusion level. A non-hierarchical k-means clustering procedure was used, specifying the number of clusters identified in the first step, using a random initial seed and 10 iterations in order to further refine the preliminary solution by optimizing the classification. The final cluster solution was selected based on interpretability and the percent of the study population in each cluster. Reliability and stability of the final cluster solution was tested by randomly taking a subsample (50%) of the total sample and repeating the analyses on this subsample. To check agreement, a kappa statistic was calculated between the cluster solutions of the subsample and that of the total sample.
Pearson's chi-square tests were used to investigate the differences in cluster distribution by gender, age group, education level, geographical area, and BMI status. One-way ANOVA was used to compare physical activities, sedentary behaviors and sleep time across clusters stratified by gender and age group. General linear models were used to estimate multivariate means for food consumption and dietary pattern scores across clusters adjusted for age and energy intake. Binary logistic regression analysis was used to explore the prevalence odds ratios for obesity and overweight among lifestyle patterns. The models were adjusted for energy intake, sex, age, educational level, and geographical area. Statistical tests were two-tailed with a 5% level of significance.

Sample Characteristics
After exclusion of outliers and participants with incomplete data, 1617 subjects aged 18-64 years, 781 men and 836 women, were included in the analyses (Table S2). Some 38% of the sample was classified in the BMI range 25-29.9 and 21.6% had BMI values ≥ 30. There was no significant difference between men and women in age distribution, level of education, or geographical area. However, overweight and obesity rates were significantly higher in men than women.

Dietary Patterns
Dietary patterns were computed for the entire sample. Bartlett's test of sphericity and KMO = 0.591 supported the appropriateness of factor analysis. Four major factors were extracted, which explained 33.1% of the variance in the model. The first dietary pattern (DP) was labeled "Traditional DP" which had the highest loading on olive oil and vegetables, high scores on fish, meat, and fruit, and negative scores on pasta and so-called "pre-cooked" foods, which include food items such as croquettes and other processed foods usually prepared deep-fried for consumption. A DP labeled "Mediterranean DP" had high scores on water, fruit, yoghourt, fish, vegetables, cheese, and olive oil, and negative scores on meat and sugar sweetened beverages. A DP labeled "Snack DP" had high scores on bread, processed and cold meats, alcoholic beverages, salted snacks, cheese, and juices. Finally, "Dairy-sweet DP" had high scores on milk, sugar and sweets, cakes, pastry, and juices, and negative scores on alcoholic beverages. Overall, these patterns explained 33.07% of the variance ( Figure 1).
Mean factor scores for "Traditional DP" and "Mediterranean DP" were significantly higher in the oldest age group (50-64 years), while "Dairy-sweet DP" factor scores were significantly higher in the younger ones (18-30 years). Men had significantly higher scores than women for "Snack DP" adjusted for age and energy intake; "Mediterranean DP" and "Dairy-sweet DP" had significantly higher scores among women. "Mediterranean DP" and "Traditional DP" factor scores were significantly higher in people with a higher educational level.
Factor scores for the "Traditional DP" and "Dairy-sweet DP", adjusted for energy intake, age, and gender, were significantly higher in the north-northwest region, while "Mediterranean DP" and "Snack DP" factor scores were significantly higher in the Eastern Mediterranean region. Mean factor scores for "Traditional DP" and "Mediterranean DP" were significantly higher in the oldest age group (50-64 years), while "Dairy-sweet DP" factor scores were significantly higher in the younger ones (18-30 years). Men had significantly higher scores than women for "Snack DP" adjusted for age and energy intake; "Mediterranean DP" and "Dairy-sweet DP" had significantly higher scores among women. "Mediterranean DP" and "Traditional DP" factor scores were significantly higher in people with a higher educational level.
Factor scores for the "Traditional DP" and "Dairy-sweet DP", adjusted for energy intake, age, and gender, were significantly higher in the north-northwest region, while "Mediterranean DP" and "Snack DP" factor scores were significantly higher in the Eastern Mediterranean region.

Lifestyle Patterns
Based on the four identified DPs, minutes per week of vigorous, moderate physical activity, walking, biking, sedentary time, sleep duration on weekdays, and smoking habits, the three-cluster solution was found to be adequate and meaningful regarding the different patterns. The kappa statistic (κ = 0.94) suggested good agreement.

Lifestyle Patterns
Based on the four identified DPs, minutes per week of vigorous, moderate physical activity, walking, biking, sedentary time, sleep duration on weekdays, and smoking habits, the three-cluster solution was found to be adequate and meaningful regarding the different patterns. The kappa statistic (κ = 0.94) suggested good agreement.
Characteristics of the subjects classified in the lifestyle patterns identified are described in Table 1. The "Mixed diet-physically active-low sedentary lifestyle pattern" included 13% of the sample and a significantly higher proportion of men. The "Not poor diet-low physical activity-low sedentary lifestyle pattern" included 63.3% of the sample, a significantly higher proportion of women. The "Poor diet-low physical activity-sedentary lifestyle pattern" included 23.6% of the sample. Pearson's chi-square tests were used to investigate the differences in cluster distribution by gender, age group, education level, geographical area, and BMI status.
A higher percentage of people across age groups were classified into the "Not poor diet-low physical activity-low sedentary lifestyle pattern". However, a significantly lower proportion of individuals in the older age group (50-64 years) was classified in the "Mixed diet-physically active-low sedentary lifestyle pattern" and a higher proportion of people aged 18-30 years were classified into the "Poor diet-low physical activity-sedentary lifestyle pattern". The highest proportion of people with a lower educational level was classified into the "Not poor diet-low physical activity-low sedentary lifestyle pattern". There was no difference in the clusters by geographical area.
Prevalence rates of obesity were significantly higher among people allocated in the "Not poor diet-low physical activity-low sedentary lifestyle pattern" and overweight in the "Mixed diet-physically active-low sedentary lifestyle pattern", however, those differences were not significant in the stratified analysis by age and gender. Table 2 describes physical activity behaviors, sedentary, sleep time on weekdays, smoking behavior, and dietary pattern Z-scores in the lifestyle patterns. Vigorous physical activity, moderate physical activity, walking time, as well as the Z-scores of the Mediterranean DP were significantly higher in men and women classified in the "Mixed diet-physically active-low sedentary lifestyle pattern". Z-scores for "Snack DP" were also higher in men in that lifestyle pattern. Men and women in the "Poor diet-low physical activity-sedentary lifestyle pattern" had significantly higher scores on the "Dairy-sweet DP" and sedentary time.
Consumption of selected food groups and beverages by lifestyle pattern in men and women is described in Table 3. Consumption of fruit, pasta, olive oil, water and alcoholic beverages, particularly wine and beer, was significantly higher in men and women included in the "Mixed diet-physically active-low sedentary lifestyle pattern". Consumption of milk, cakes and pastry, sugar and sweets was significantly higher in men and women classified in the "Poor diet-low physical activity-sedentary lifestyle pattern". Men in this lifestyle pattern showed significantly higher consumption of pre-cooked deep fried foods and high alcoholic content beverages. Women in this pattern had significantly higher consumption of savory snacks, juices and sugar sweetened soft drinks beverages.
Prevalence of obesity was compared between lifestyle patterns, adjusting for gender, age, educational level, geographical area and energy intake ( Table 4). The prevalence odds ratio (POR) for obesity in men, 0.52 (IC 95% 0.29-0.92) allocated in the "Mixed diet-physically active-low sedentary lifestyle pattern" was significantly lower compared to those in the "Poor diet-low physical activity-sedentary lifestyle pattern".

Discussion
In this cross-sectional study conducted in a random sample of Spanish adult population aged 18-64 years, four dietary patterns were characterized by factor analysis and subsequently used to identify three different lifestyle patterns based on the DPs, physical activities, sedentary time, and sleeping and smoking habits, using cluster analyses: a "Mixed diet-physically active-low sedentary lifestyle pattern", a "Not poor diet-low physical activity-low sedentary lifestyle pattern" which included 63.3% of the sample, and a "Poor diet-low physical activity-sedentary lifestyle pattern". A higher proportion of people aged 18-30 years was classified into the "Poor diet-low physical activity-sedentary lifestyle pattern". The prevalence odds ratio for obesity in men allocated in the "Mixed diet-physically active-low sedentary lifestyle pattern" was significantly lower compared to those in the "Poor diet-low physical activity-sedentary lifestyle pattern".
Analysis of dietary patterns has become an important approach in food consumption studies and nutritional epidemiology [11,27,28] Researchers report different methodological approaches and procedures to identify diet and lifestyle patterns, such as a priori defined patterns or data-driven, identified using principal components (PCA) [29], factor analysis [11,30], cluster analysis [31][32][33] and, more recently, hybrid methods [2]. Some pieces of research have identified lifestyle patterns associated to specific phenotypes and obesity in several population groups [34,35] and contributed to unravel the consequences that different diets may have on health [2,11,[35][36][37][38].
In this study factor analysis was used to identify dietary patterns that were later used in cluster analysis. In the ANIBES study four DPs were singled out in adults. Similar DPs have been reported [37]. In Spain, in the context of the DORICA study, a pooled analysis of population-based cross-sectional surveys conducted between 1990 and 2000, using cluster analysis of three different DPs were characterized. A "Meat rich DP" with higher consumption of meat, cereals, and potatoes, an "Unbalanced DP" with higher consumption of milk and low consumption of vegetables and cereals, and a "Mediterranean DP" with higher consumption of vegetables, fruit, fish, and olive oil. The "Unbalanced DP" was associated with higher alcohol consumption, inadequate fruit and vegetable consumption and poor physical activity. This pattern was associated to higher prevalence odds ratio for metabolic syndrome. Conversely, the "Mediterranean DP" was associated with better fat intake quality profile and fiber intake [19].
More recently other authors have identified two DPs using factor analysis in Spanish adults in the context of Food4Me Project. A pattern with higher consumption of fast and processed foods, potatoes, red and white meats, processed cereals, and low consumption of fruit and vegetables, directly associated with BMI. A second DP showed higher consumption of eggs, fish, legumes, nuts, low-calorie beverages, oil, vegetables, and white meat [8].
Despite cross-sectional studies not allowing for causal inference, food patterns with higher energy density and low fiber content have been associated with higher obesity prevalence rates [2]. In longitudinal studies, such as the Baltimore Study, dietary patterns high in reduced-fat dairy, high fiber grains and cereals, vegetables and fruits, and low in meats, sugar-sweetened soft drinks, refined grains, and high fat dairy products, were significantly associated with lower weight gain, prospectively [39]. An energy-dense, high-saturated fat and low-fiber dietary pattern with high loadings of fast foods and snacks and low loadings of fruits and vegetables was shown to increase body weight, waist circumference, blood pressure, serum insulin, and lipid profile during a 10-year follow-up in severely obese Swedish adults [40]. In line with this, in adult Canadians a strong and consistent relation between an energy-dense, high-fat, low-fiber density dietary pattern and the risk of obesity was observed, significant in different population subgroups [2].
In the ANIBES study, 45.6% of male and 27.5% of female Spanish adults did not meet the recommendation of 150 min/week of moderate PA. In addition, 56.2% of male and 74% of female adults did not meet the recommendations for vigorous PA [13]. In this study we analyzed if these physical activity behaviors combined in a meaningful way with other energy balance-related behaviors, DPs, and lifestyles. We also used this procedure to identify dietary and lifestyle patterns in Spanish children and adolescents and identified two patterns: an "Unhealthier lifestyle pattern" a combination of low physical activity and poorer diet, and a "Healthier lifestyle pattern" which combined high physical activity, low sedentary behavior, longer sleep duration, and healthier diet.
Studies analyzing individual behaviors have found healthier eating habits, such as adequate fruit and vegetable consumption, to be more prevalent among women than in men [41], but adherence to physical activity recommendations were more prevalent in men than in women [42]. In line with the above, in this study "Snack DP" mean scores adjusted for age and energy intake were significantly higher in men than in women. Conversely, those of the "Mediterranean DP" were significantly higher in women. Men and women classified in the "Mixed diet-physically active-low sedentary lifestyle pattern" showed the highest mean scores for the "Mediterranean DP". However, men in that lifestyle pattern showed the highest mean scores for the "Snack DP", as well. The highest mean scores for the "Dairy sweet DP" were observed in men and women classified into the "Poor diet-low physical activity-sedentary lifestyle pattern".
In this study a significantly higher proportion of men was classified into the "Mixed diet-physically active-low sedentary lifestyle pattern" (71.9%), while significantly more women were classified into the "Not poor diet-low physical activity-low sedentary lifestyle pattern" (58.5%). Men classified in the "Mixed diet-physically active-low sedentary lifestyle pattern" were less likely to be obese (POR 0.52; 95% IC 0.29-0.93), compared to those in the "Poor diet-low physical activity-sedentary lifestyle pattern". Cassidy et al. reported obese adults to be two to five times more likely to report an "unhealthy phenotype", a combination of low physical activity, high TV viewing, and poor sleep duration compared to normal weight adults in the UK Biobank [43].
A significantly higher proportion of women classified into the "Mixed diet-physically active-low sedentary lifestyle pattern" had a high educational level (35.6%). Women with a low educational level were more likely to be obese compared to those highly educated. Clustering and co-occurrence of multiple risk behaviors has been reported to be more likely in men and women with lower education or intermediate educational level compared with those with higher education [44].
As shown in this study, individuals may follow a variety of unhealthy lifestyle behaviors which combine to favor weight gain, or mixtures of healthy and unhealthy practices. The findings in this study underline the importance of designing and implementing interventions that address multiple health risk habits, considering lifestyle patterns, clustering of risk behaviors, and associated determinants. It is recognized that interventions which target more than one risk behavior have the potential for greater health benefits, enhance health promotion opportunities and can contribute to reduce health care costs. Such an approach allows for more adequate tailoring of interventions to participants' profiles and helps to address health inequalities, reinforcing policies to ensure interventions accessible to socially-disadvantaged groups.
The ANIBES study has several strengths which include the careful design, protocol, and methodology used for dietary data collection and validated questionnaires used to collect information on physical activity which have shown good reliability and reproducibility. The study was conducted on a random population sample of the Spanish population aged 9-75 years. A limitation of this study is its cross-sectional design, which provides evidence for associations but not causal relationships. Measures of physical activity relied on self-reports and could be biased, although a careful multistep quality control procedure was implemented to minimize bias. In addition, we used factor analysis to identify DPs based on food consumption information collected by three-day food records and one 24-h recall, considering food groups to further collapse dietary intake data and avoid missing data from non-consumers of episodically consumed foods. Finally, both factor analysis and cluster analysis are procedures commonly used to identify DPs and analyze clustering of lifestyles. However, those procedures rely on several subjective decisions that may influence outcomes regarding the number and type of patterns and clusters identified.

Conclusions
Four dietary patterns were characterized in Spanish adults aged 18-64 years, one of them closer to the Mediterranean DP. These patterns were subsequently used to identify three different lifestyle patterns: a "Mixed diet-physically active-low sedentary lifestyle pattern", a "Not poor diet-low physical activity-low sedentary lifestyle pattern" which included 63.3% of the sample, and a "Poor diet-low physical activity-sedentary lifestyle pattern". A higher proportion of people aged 18-30 years was classified into the "Poor diet-low physical activity-sedentary lifestyle pattern". The prevalence odds ratio for obesity in men allocated in the "Mixed diet-physically active-low sedentary lifestyle pattern" was significantly lower compared to those in the "Poor diet-low physical activity-sedentary lifestyle pattern". Although prospective research on large population samples would be desirable to further analyze and track over time the lifestyle patterns and how they influence the development of overweight and obesity, these behavior patterns are helpful to identify specific issues in population subgroups and inform intervention strategies. Further research is needed on factors associated with lifestyle patterns to gain insight in the population subgroups at higher risk. The findings in this study underline the importance of designing and implementing interventions that address multiple health risk practices, considering lifestyle patterns and associated determinants.
Supplementary Materials: The following are available online at www.mdpi.com/2072-6643/9/6/606/s1, Table S1: Food groups and subgroups in the ANIBES Study, Table S2: Characteristics of the sample analyzed, Figure S1: Final cluster center scores for lifestyle patterns identified in Spanish adults aged 18-64 years in the ANIBES study.